System

Abstract

A system is provided.SOLUTION: A system, comprising: means for receiving a question from a user; means for analyzing the received question; means for generating an answer based on the analyzed question; and means for providing the generated answer to the user, wherein the means for providing the generated answer to the user is included in the system, wherein the means for providing the generated answer to the user is included in the system, and wherein the means for providing the generated answer to the user is included in the system, wherein the means for providing the generated answer to the user is included in the system, and wherein the means for providing the generated answer to the user is included in the system. AI AI.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] The present invention aims to solve the problems that arise in large-scale companies, such as unclear business processes and operational flows, and the large number of people involved, which makes it easy for problems to arise. Another problem is that the information available for review is not always up-to-date, and its reliability is not guaranteed. This can hinder business efficiency and delay decision-making. [Means for solving the problem]

[0005] The present invention is a system that includes a means for collecting data from a data collection system, a means for filtering and cleaning the collected data, a means for training a generative AI model using the filtered and cleaned data, a means for storing the data trained by the generative AI model in a knowledge base, a means for accepting questions from users, a means for analyzing the accepted questions, a means for generating answers based on the analyzed questions, and a means for providing the generated answers to users, thereby solving the above-mentioned problems.

[0006] A "data collection system" is a system for automatically collecting data from various data sources (CRM, ERP, mail servers, chat systems, etc.).

[0007] "Means for collecting data" refers to the function of obtaining data from a data collection system and incorporating it into the system.

[0008] "Filtering" refers to the process of removing unnecessary information from collected data and extracting and organizing only the necessary information.

[0009] "Cleaning" refers to the process of formatting data and correcting noise and errors.

[0010] A "generative AI model" refers to an artificial intelligence model that uses natural language processing and machine learning techniques to analyze data and learn patterns.

[0011] "Means of training" refers to the ability to use filtered and cleaned data to teach a generative AI model new knowledge.

[0012] A "knowledge base" refers to a database that stores the data and information learned by a generative AI model and can refer to it as needed.

[0013] "User" refers to a party who utilizes the system to input information and obtain responses.

[0014] "Means for accepting questions" refers to the function of inputting questions from users into the system and capturing those questions for analysis.

[0015] "Means of analysis" refers to the function of analyzing received questions using natural language processing technology and identifying relevant data and information.

[0016] "Means for generating answers" refers to the function of using a generative AI model to create appropriate answers based on the analyzed questions.

[0017] "Means for providing" refers to the function of displaying or notifying the user of the generated answer. [Brief explanation of the drawings]

[0018] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0019] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0020] First, the terms used in the following description will be explained.

[0021] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0022] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0023] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0024] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0025] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0026] [First embodiment]

[0027] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0028] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0029] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0030] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0031] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0032] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0033] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0034] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0035] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0036] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0037] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0038] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0039] This invention is an "Internal Knowledge GPT" system that visualizes internal business processes and operational flows and improves the reliability of the latest information. In this system, servers, terminals, and users cooperate to collect, organize, learn, and answer questions about information.

[0040] Data collection

[0041] The server collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.) via APIs and data files. For example, the server retrieves sales transaction data from a customer relationship management system (CRM) and organizes the information.

[0042] Data Filtering and Cleaning

[0043] The server filters and cleans the collected data, which includes removing duplicate data and unnecessary information, and normalizing the data format. For example, the server standardizes date formats and corrects typos in text data.

[0044] Generative AI model training

[0045] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process enables the model to understand the latest business information and operational flows, and generate optimal answers based on that information.

[0046] Knowledge Base Updates

[0047] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[0048] Enter a question

[0049] A user inputs a question into the system using a terminal. For example, the user inputs, "Please tell me about the application flow for a new project."

[0050] question analysis

[0051] The server receives a question from a user, analyzes the question using natural language processing technology, and retrieves relevant information from a knowledge base based on the analysis result.

[0052] Answer generation

[0053] The server uses a generative AI model to generate the best answer to the question, based on the most up-to-date information in the knowledge base.

[0054] Provide answers

[0055] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, it explains the detailed steps in the form of "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0056] Specific examples

[0057] For example, if a user asks, "What is the latest format for sales reports?"

[0058] 1. The server first collects the latest data on sales reports from the CRM system.

[0059] 2. This data is then filtered and cleaned to remove duplicates and extract the information you need.

[0060] 3. The server uses this cleaned data to train a generative AI model.

[0061] 4. When a user enters a question, the server analyzes the question using natural language processing technology and searches the knowledge base for relevant information.

[0062] 5. The generative AI model generates the optimal answer, and the server sends the answer to the device.

[0063] 6. The terminal displays to the user, "The latest sales report format is as follows," and indicates the specific items.

[0064] In this way, the in-house knowledge GPT system provides quick and accurate answers to user questions, improving work efficiency and the reliability of information.

[0065] The processing flow will be explained below.

[0066] Step 1:

[0067] The server automatically collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.). The server uses APIs to obtain data in real time and stores the collected data in temporary storage. For example, the server obtains transaction data and customer information from a CRM system.

[0068] Step 2:

[0069] The server filters the collected data. In this step, the server removes data duplication and noise and extracts only the necessary information. For example, it removes duplicate transaction information from past data logs.

[0070] Step 3:

[0071] The server cleans the filtered data, standardizing the data format and correcting typos and errors (e.g., converting date data in different formats into a standard format).

[0072] Step 4:

[0073] The server feeds the cleaned data to a generative AI model (GPT) to train the model, which then understands the latest business information and operational flows.

[0074] Step 5:

[0075] The server stores the learned knowledge of the generative AI model in a knowledge base, which is periodically updated to reflect the latest information.

[0076] Step 6:

[0077] A user uses a terminal to input a question into the system. For example, the user inputs, "Please tell me about the application flow for a new project."

[0078] Step 7:

[0079] The terminal sends the user's question to the server in text format.

[0080] Step 8:

[0081] The server receives the question, analyzes it using natural language processing technology, extracts keywords and the content of the request, and searches for relevant information from a knowledge base.

[0082] Step 9:

[0083] The server uses the generative AI model to generate the best answer to the question, and creates an answer that includes specific steps and information based on the analysis results.

[0084] Step 10:

[0085] The server sends the generated answer to the terminal, which then communicates the answer to the user in text format.

[0086] Step 11:

[0087] The device will display a response to the user, providing detailed instructions such as, "The flow for applying for a new project is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0088] Example 1

[0089] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0090] In many companies today, information about business processes and operational flows is scattered across various departments and systems, making it difficult to quickly and accurately collect and organize that information and utilize it in a timely manner. This results in the time and effort required to obtain the necessary information, which reduces the efficiency of internal operations. Furthermore, information may not be up-to-date or reliable, which can hinder decision-making and business execution.

[0091] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0092] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for storing knowledge learned by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions using natural language processing technology, means for acquiring information from the knowledge base based on the analyzed questions and generating answers, and means for providing the generated answers to users. This makes it possible to centrally manage data from each business system and train a generative AI model based on the latest filtered and cleaned data, thereby providing quick and accurate answers to user questions.

[0093] A "data collection system" is a system for collecting data from various business systems (CRM, ERP, mail servers, chat systems, etc.).

[0094] "Filtering" is the process of removing duplicate and unnecessary information from collected data.

[0095] "Cleaning" is the process of standardizing the format of collected data and correcting typos and formatting.

[0096] A "generative AI model" is an artificial intelligence model that learns from filtered and cleaned data and is used to generate answers to user questions.

[0097] A "knowledge base" is a database in which the knowledge of a trained generative AI model is stored and referenced as needed.

[0098] "Natural language processing technology" is a technology for analyzing questions from users and understanding context and keywords.

[0099] "Analysis" refers to the process of analyzing a user's question using natural language processing technology to understand the intent of the question and related information.

[0100] "Answer generation" is the process of creating optimal answers using a generative AI model based on the analyzed question.

[0101] "Providing" refers to the act of displaying or transmitting the generated answer to the user.

[0102] This invention is an "Internal Knowledge GPT" system that visualizes internal business processes and operational flows and improves the reliability of the latest information. In this system, servers, terminals, and users cooperate to collect, organize, learn, and answer questions about information.

[0103] Data collection

[0104] The server collects the necessary data from various business systems, such as CRM, ERP, mail servers, and chat systems, via APIs and data files. Specifically, it uses REST API clients and database connectors. For example, the server retrieves sales transaction data from a customer relationship management system (CRM) and organizes the information.

[0105] Data Filtering and Cleaning

[0106] The server filters and cleans the collected data. This includes processing the data using Python's pandas and SQL queries. It also removes duplicate data and standardizes data formats. For example, it removes duplicate data for the same customer and standardizes date formats.

[0107] Generative AI model training

[0108] The server uses the filtered and cleaned data to train a generative AI model (GPT). This process uses machine learning frameworks such as PyTorch and TensorFlow. By training the model with the latest business information and operational flows, the server is able to generate accurate answers that are adapted to user questions.

[0109] Knowledge Base Updates

[0110] The knowledge of the generative AI model trained by the server is stored in a knowledge base. The knowledge base is constructed as an SQL database or a NoSQL database (e.g., MongoDB). The stored knowledge can be referenced by the server as needed.

[0111] Enter a question

[0112] Users use a terminal to input questions into the system, either through a browser-based interface or a dedicated application. For example, a question might be, "Please tell me about the application process for a new project."

[0113] question analysis

[0114] The server receives questions from users and analyzes them using natural language processing technology. Specifically, it uses natural language processing libraries such as spaCy and NLTK. The server extracts keywords from the questions and understands their context, then retrieves the necessary information from a knowledge base.

[0115] Answer generation

[0116] The server uses a generative AI model to generate the best answer to the question, and references the latest business information based on a knowledge base to provide an accurate answer to the user's question.

[0117] Provide answers

[0118] The server sends the generated answer to the terminal, which then displays it to the user, such as "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0119] Examples and prompts

[0120] For example, if a user asks, "What is the latest format for sales reports?"

[0121] 1. The server collects the latest data on sales reports from the CRM system.

[0122] 2. Filter and clean this data to remove duplicates and extract the information you need.

[0123] 3. The server uses this cleaned data to train a generative AI model.

[0124] 4. When a user enters a question into the terminal, the server analyzes the question using natural language processing technology and searches for relevant information from a knowledge base.

[0125] 5. The generative AI model generates the optimal answer, and the server sends the answer to the device.

[0126] 6. The terminal displays to the user, "The latest sales report format is as follows," and indicates the specific items.

[0127] In this way, the in-house knowledge GPT system provides quick and accurate answers to user questions, improving work efficiency and the reliability of information.

[0128] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0129] Step 1:

[0130] The server collects the necessary data from the data collection system. Inputs are APIs and data files from each business system (CRM, ERP, mail server, chat system, etc.). Based on these inputs, the server collects data using a REST API client or database connector. The output is saved as raw data.

[0131] Step 2:

[0132] The server filters and cleans the collected data. The input is the collected raw data. The server processes the data using Python pandas and SQL queries to remove duplicate data and standardize data formats. Specifically, it removes duplicate data for the same customer and standardizes date formats. The output is the filtered and cleaned data.

[0133] Step 3:

[0134] The server trains a generative AI model using the filtered and cleaned data. The input is the cleaned data, and the server trains a generative AI model (GPT) using a machine learning framework such as PyTorch or TensorFlow. The output is a trained generative AI model.

[0135] Step 4:

[0136] The server stores the knowledge of the trained generative AI model in a knowledge base. The input is the trained generative AI model, which the server stores in an SQL database or a NoSQL database (e.g., MongoDB). The output is the data stored as a knowledge base.

[0137] Step 5:

[0138] A user uses a terminal to input a question into the system. The input is the user's question, and the terminal uses a browser-based interface or a dedicated application. A specific example is "Please tell me about the application flow for a new project." The output is the question.

[0139] Step 6:

[0140] The server analyzes questions from users using natural language processing technology. The input is the user's question, and the server analyzes the question using natural language processing libraries such as spaCy or NLTK. The analysis extracts keywords from the question and understands the context. The output is the analysis results.

[0141] Step 7:

[0142] The server retrieves information from the knowledge base based on the analyzed question and generates the optimal answer using a generative AI model. The input is the analysis result and information from the knowledge base. The server creates the optimal answer based on the data it has learned in advance. The output is the generated answer.

[0143] Step 8:

[0144] The server sends the generated answer to the terminal, which then displays it to the user. The input is the generated answer, and the terminal displays the answer using a browser-based interface or a dedicated application. The output is the display of the answer to the user. The specific operation is displayed as detailed steps: "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0145] (Application example 1)

[0146] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0147] In conventional factory management systems, information related to manufacturing processes and maintenance procedures is managed in a decentralized manner, making it difficult for workers and managers to quickly obtain the information they need.It is also difficult to provide accurate information based on the latest business and operational flows, creating challenges in improving business efficiency and reliability.

[0148] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0149] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for saving the data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions, means for generating answers based on the analyzed questions, means for collecting manufacturing process data and maintenance history from a factory management system and generating answers to the questions based thereon, and means for providing the generated answers to users. This enables quick and accurate provision of information on work flows, equipment operating procedures, and maintenance procedures within the factory.

[0150] A "data collection system" is an infrastructure for acquiring data from various business systems, sensors, etc.

[0151] "Filtering" is the process of removing unnecessary information and duplicate data from collected data.

[0152] "Cleaning" is the process of improving data quality by correcting errors and standardizing the format of the data.

[0153] A "generative AI model" is an artificial intelligence model that learns from collected and organized data and generates appropriate answers to user questions.

[0154] A "knowledge base" is a database where information learned by a generative AI model is stored and can be referenced at any time.

[0155] "Accepting a question" is the process by which the system obtains a query from a user as input.

[0156] "Question analysis" is the process of using natural language processing techniques to understand the intent of a user's question and extract relevant information.

[0157] "Generating an answer" is the process of using a generative AI model to create the optimal answer to a user's question.

[0158] A "factory management system" is a system for managing operational information on manufacturing processes and equipment.

[0159] "Manufacturing process data" refers to a series of information related to the manufacturing process of a product, including process progress and quality control data.

[0160] "Maintenance history" refers to records relating to the maintenance and repair of equipment and facilities, including procedures for dealing with problems when they occur and repair history.

[0161] This invention relates to the "Factory Knowledge GPT" system, which enables workers and managers engaged in factory management and maintenance work to quickly and accurately obtain the information they need. This system involves the collaboration of servers, terminals, and users to collect, organize, learn, and answer questions about information.

[0162] Data collection

[0163] The server collects information such as manufacturing process data and equipment maintenance history from the factory management system via APIs and data files. For example, the server obtains daily production line progress data and past maintenance records for each piece of equipment from the factory management system and organizes the information.

[0164] Data Filtering and Cleaning

[0165] The server then filters and cleans the collected data, removing unnecessary information and duplicate data, standardizing data formats, etc. For example, the server may standardize date formats and correct typos in text data to improve data quality.

[0166] Generative AI model training

[0167] The server uses the filtered and cleaned data to train a generative AI model (GPT). This model understands the latest business information and operational flows, and can generate optimal answers to user questions. For example, it learns troubleshooting procedures and equipment maintenance procedures on a production line.

[0168] Knowledge Base Updates

[0169] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[0170] Question input and analysis

[0171] A user inputs a question into the system using a terminal, such as a smartphone or robot interface, asking, "What are the troubleshooting steps for assembly line B?" The server then analyzes the question using natural language processing technology and retrieves relevant information from the knowledge base.

[0172] Answer generation and provision

[0173] The server uses a generative AI model to generate the best answer to the question. The generated answer is based on the latest information in the knowledge base. For example, it provides detailed instructions in the form of "Troubleshooting steps for assembly line B are as follows: 1. Check the power supply of the equipment. 2. Check the location of the sensors. 3. Stop and reset the production line. 4. Check each error log. 5. If you are unsure, contact the administrator."

[0174] Hardware and software used

[0175] The system uses the following hardware and software:

[0176] Hardware: Servers (cloud services recommended, e.g., AWS EC2), smartphones (e.g., iPhone, Android device), factory robots (e.g., robotic arms from ABB or FANUC)

[0177] Software: GPT model (Hugging Face Transformers library), API requests (requests library), data management (JSON format)

[0178] Specific examples

[0179] For example, a factory worker types the following question:

[0180] Question: "What is the maintenance procedure for device A?"

[0181] In response to this question, the server collects and filters the necessary data and uses a generative AI model to generate an answer like this:

[0182] "The maintenance procedure for Equipment A is as follows: 1. Power off the equipment. 2. Put on protective equipment. 3. Inspect each part. 4. Apply lubricant. 5. Report any problems immediately."

[0183] This allows users to quickly obtain accurate and reliable information.

[0184] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0185] Step 1: Data collection

[0186] The server collects manufacturing process data and equipment maintenance history from the factory management system. Specifically, it obtains data through APIs and data files. The input to this step is data from the factory management system, and the output is the collected raw data. The server sends a request to the API endpoint and receives the data in JSON format.

[0187] Step 2: Data filtering

[0188] The server filters the collected data, specifically removing unnecessary information and duplicate data based on manually set filter conditions. The input to this step is the raw data collected in step 1, and the output is the filtered data. The server processes the data by removing unnecessary fields and removing duplicate data.

[0189] Step 3: Data cleaning

[0190] The server cleans the filtered data. Specifically, it standardizes the data format and corrects typos. The input to this step is the data filtered in step 2, and the output is the cleaned data. Standardizing the data format includes standardizing the date format and text encoding. Typos are automatically corrected using a preset dictionary.

[0191] Step 4: Generative AI model training

[0192] The server uses the cleaned data to train a generative AI model (GPT). Specifically, it inputs the cleaned data into the model and updates the generative AI model. The input of this step is the cleaned data, and the output is a trained generative AI model. The server tokenizes the data and trains it to optimize the model parameters.

[0193] Step 5: Update your knowledge base

[0194] The server saves the knowledge of the trained generative AI model in a knowledge base. Specifically, it stores the information learned by the model in a database. The input of this step is the trained generative AI model, and the output is an updated knowledge base. The server inserts the information into the database for the knowledge base and creates the necessary indexes.

[0195] Step 6: Enter your question

[0196] The user uses a terminal to input a question into the system. Specifically, the question is uttered via text or voice through a smartphone or robot interface. The input in this step is the user's question, and the output is digital data of the question. The terminal receives the user's input, converts it into text format, and sends it to the server.

[0197] Step 7: Question Analysis

[0198] The server analyzes the user's question. Specifically, it uses natural language processing technology to understand the intent of the question and searches for relevant information in a knowledge base. The input to this step is the digital data of the user's question, and the output is the analysis result. The server tokenizes the question and analyzes it using natural language processing (NLP) algorithms to identify relevant information.

[0199] Step 8: Answer Generation

[0200] The server uses a generative AI model to generate the optimal answer to the question. Specifically, the model generates an answer based on the analysis results. The input to this step is the analysis results, and the output is the generated answer. The generative AI model generates answer text based on the tokenized analysis results and converts it into a text format.

[0201] Step 9: Provide your answers

[0202] The server sends the generated answer to the terminal, which then displays the answer to the user. Specifically, the answer is displayed on the interface of a smartphone or robot. The input of this step is the generated answer, and the output is the answer displayed to the user. The terminal displays the text data received from the server on the user interface.

[0203] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0204] This invention is an "Internal Knowledge GPT" system that combines an emotion engine to visualize internal business processes and operational flows and improve the reliability of the latest information. In this system, the server, terminals, and users cooperate to collect, organize, learn, answer questions, and recognize emotions.

[0205] Data collection

[0206] The server collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.) via APIs and data files. For example, the server retrieves sales transaction data and customer information from a CRM system and organizes the information.

[0207] Data Filtering and Cleaning

[0208] The server filters the collected data, removing unnecessary information, and then performs cleaning and formatting on the data. This includes removing duplicate data and standardizing formats. For example, the server standardizes date formats and corrects typos and errors.

[0209] Generative AI model training

[0210] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process enables the AI ​​model to understand the latest business information and operational flow, and generate optimal answers based on that information.

[0211] Knowledge Base Updates

[0212] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[0213] Enter a question

[0214] The user inputs a question into the system using a terminal. For example, the user might input, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[0215] question analysis

[0216] The server receives questions from users and analyzes them using natural language processing technology along with emotional data detected by the emotion engine, extracting keywords and requests from the questions and retrieving relevant information from the knowledge base.

[0217] Answer generation

[0218] The server takes into account the user's emotional data when generating the optimal answer to the question using a generative AI model. Based on the analysis results, it creates an answer that includes specific steps and information, and adjusts the tone of the answer. For example, if the user is feeling frustrated, it will soften the tone of the answer.

[0219] Provide answers

[0220] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, it shows detailed steps such as, "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department." The answer is also provided in a tone that corresponds to the user's emotions.

[0221] Specific examples

[0222] For example, if a user asks, "Tell me about your recent sales report?"

[0223] 1. The server first collects the latest data on sales reports from the CRM system.

[0224] 2. This data is then filtered and cleaned to remove duplicates and extract the information you need.

[0225] 3. The server uses this cleaned data to train a generative AI model.

[0226] 4. When the user enters a question, the emotion engine recognizes the user's emotion at the time of entering the question and sends the emotion data to the server.

[0227] 5. The server receives the question, analyzes keywords such as "sales report" and "recent" using natural language processing technology, and searches for related information in the knowledge base.

[0228] 6. The server uses a generative AI model to generate the optimal answer, taking into account the user's emotional data. For example, if the user is dissatisfied, the answer will be delivered in a more friendly tone.

[0229] 7. The server sends the generated answer to the terminal, which displays it to the user.

[0230] In this way, the in-house knowledge GPT system combined with the emotion engine provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[0231] The processing flow will be explained below.

[0232] Step 1:

[0233] The server automatically collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.). The server uses APIs to obtain data in real time and stores the collected data in temporary storage. For example, the server obtains transaction data and customer information from a CRM system.

[0234] Step 2:

[0235] The server filters the collected data. In this step, the server removes data duplication and noise and extracts only the necessary information. For example, it removes duplicate transaction information from past data logs.

[0236] Step 3:

[0237] The server cleans the filtered data, standardizing the data format and correcting typos and errors (e.g., converting date data in different formats into a standard format).

[0238] Step 4:

[0239] The server feeds the cleaned data to a generative AI model (GPT) to train the model, which then understands the latest business information and operational flows.

[0240] Step 5:

[0241] The server stores the learned knowledge of the generative AI model in a knowledge base, which is periodically updated to reflect the latest information.

[0242] Step 6:

[0243] A user inputs a question into the system using a terminal. For example, the user might input, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[0244] Step 7:

[0245] The device sends the user's question and emotional data to the server. The question is sent in text format, and the emotional data is sent as a numerical value or category.

[0246] Step 8:

[0247] The server receives the question and analyzes it using natural language processing technology, extracting keywords and the content of the request, and collating emotional data to search for relevant information from a knowledge base.

[0248] Step 9:

[0249] The server uses a generative AI model to generate the optimal answer to the question, taking into account the emotional data and creating an answer in a tone that corresponds to the user's emotions. For example, if the data indicates anger, the tone of the answer will be calmer.

[0250] Step 10:

[0251] The server sends the generated answer to the terminal, which then communicates the answer to the user in text format.

[0252] Step 11:

[0253] The device will then display answers and sentiment analysis information to the user, providing detailed instructions in a tone that reflects their emotions, such as "The new project application process is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0254] Specific examples

[0255] Example 1: When a user asks, "Tell me about your recent sales report."

[0256] Step 1:

[0257] The server collects the latest sales report data from the CRM system.

[0258] Step 2:

[0259] The server filters the collected data and removes duplicate data.

[0260] Step 3:

[0261] The server standardizes the data format and corrects typos and errors.

[0262] Step 4:

[0263] The server feeds the cleaned data into a generative AI model (GPT) to train the model.

[0264] Step 5:

[0265] The server stores the knowledge of the generated AI model in a knowledge base.

[0266] Step 6:

[0267] The user inputs a question into the terminal, such as "Please tell me about the recent sales report." At this time, the emotion engine recognizes the user's emotion.

[0268] Step 7:

[0269] The device sends the question and emotion data to the server.

[0270] Step 8:

[0271] The server analyzes the question and searches a knowledge base for relevant information.

[0272] Step 9:

[0273] The server uses generative AI models to generate optimal answers, crafting them in a tone that reflects the user's emotions.

[0274] Step 10:

[0275] The server sends the generated response to the terminal.

[0276] Step 11:

[0277] The device will then display the answer and sentiment analysis information to the user, such as "Regarding the recent sales report, please pay attention to the following points: 1. Overview of sales activities 2. Successes and challenges 3. Next action plan," in a tone that reflects the user's sentiment.

[0278] In this way, the in-house knowledge GPT system combined with the emotion engine provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[0279] Example 2

[0280] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0281] Conventional data collection and question answering systems generate answers without considering the user's feelings, which leads to low user satisfaction. Furthermore, it is difficult to provide accurate answers when inaccurate or duplicate data exists.

[0282] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0283] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for storing the data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing emotions, means for analyzing the accepted questions using natural language processing technology, means for generating answers based on the analyzed questions and emotion data, and means for providing the generated answers to the users. This makes it possible to quickly provide accurate and appropriate answers that take the user's emotions into consideration.

[0284] A "data collection system" is a system for efficiently collecting necessary data from various sources.

[0285] "Filtering" is the process of removing unnecessary or redundant information from collected data.

[0286] "Cleaning" is a process that involves correcting data and standardizing its format in order to ensure its accuracy.

[0287] A "generative AI model" is an artificial intelligence model that can learn large amounts of data and generate sentences in natural language like a human.

[0288] A "knowledge base" is a database that systematically stores collected and learned information and can be referenced and used as needed.

[0289] "User" refers to a person or entity that uses the system to search for information or ask a question.

[0290] The "means for accepting questions" is an interface for inputting questions from users into the system.

[0291] "Natural language processing technology" is a technology that enables computers to understand and process human language.

[0292] An "emotion engine" is a system that detects emotions from user input and reactions and analyzes the data.

[0293] "Answer generation means" refers to a method or technology for creating an optimal answer based on the user's question and emotion data.

[0294] "Means for adjusting tone" refers to methods or techniques for appropriately adjusting the expression or tone of the generated response to match the user's emotions.

[0295] This invention is an "Internal Knowledge GPT" system that combines an emotion engine to visualize internal business processes and operational flows and improve the reliability of the latest information. In this system, the server, terminals, and users cooperate to collect, organize, learn, answer questions, and recognize emotions.

[0296] Data collection

[0297] The server collects data from each business system (e.g., customer relationship management system (CRM), enterprise resource planning system (ERP), mail server, chat system, etc.) via APIs and data files. For example, the server obtains sales transaction data and customer information from the customer relationship management system and stores it in storage.

[0298] Data Filtering and Cleaning

[0299] The server filters the collected data, removing unnecessary and duplicate information, and then performs cleaning and formatting. This includes standardizing date formats and correcting typos and errors. For example, it standardizes date data stored in different formats to the "YYYY-MM-DD" format.

[0300] Generative AI model training

[0301] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process uses machine learning libraries such as TensorFlow and PyTorch, allowing the AI ​​model to understand the latest business information and operational flows and generate optimal answers.

[0302] Knowledge Base Updates

[0303] The server stores the learned knowledge of the generative AI model in a knowledge base, which functions as a database that can be referenced whenever necessary, allowing users to access the latest business information.

[0304] Enter a question

[0305] The user uses a terminal to input a question into the system. For example, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and simultaneously records that emotion data.

[0306] question analysis

[0307] The server receives questions from users and analyzes them using natural language processing technology along with emotional data detected by the emotion engine. Keywords and requests from the questions are extracted and analyzed. For example, keywords such as "new project" and "application flow" are extracted.

[0308] Answer generation

[0309] The server uses a generative AI model to generate the best answer to the question, taking into account the user's emotional data. It provides specific instructions and information in a toned manner. For example, if the user is frustrated, it softens the tone of the answer.

[0310] Provide answers

[0311] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, detailed steps are shown in the form of "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0312] Specific examples

[0313] For example, if a user asks, "Tell me about your recent sales report?":

[0314] 1. The server first collects the latest data on sales reports from the customer relationship management system.

[0315] 2. The server filters and cleans the data, removing duplicates and extracting the information you need.

[0316] 3. The server trains the generative AI model using the cleaned data, for example, incorporating the latest sales results and customer feedback.

[0317] 4. When the user enters a question, the device recognizes the emotion data using the emotion engine and sends it to the server. For example, if the user is feeling impatient, that emotion will be recorded.

[0318] 5. The server receives the question, analyzes keywords such as "sales report" and "recent" using natural language processing technology, and searches for related information in the knowledge base.

[0319] 6. The server uses a generative AI model to generate the optimal answer, taking into account the user's emotional data (impatience), and provides a friendly response such as, "Our recent sales report shows that after the mid-October version upgrade, feedback from major customers has been positive, and sales have increased by 20%."

[0320] 7. The server sends the generated answer to the terminal, which displays it to the user.

[0321] In this way, the "Internal Knowledge GPT" system, which combines an emotion engine, provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[0322] Prompt Sentence Examples

[0323] To ask "Tell me about recent sales reports," a user would type the following prompt into the terminal:

[0324] Please tell me about your recent sales report.

[0325] The system analyzes this prompt, collects, organizes, and analyzes the necessary information, and provides a satisfactory answer to the user.

[0326] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0327] Step 1:

[0328] The server collects data from each business system (e.g. CRM, ERP, mail server, chat system, etc.). The input is the API or data file from each business system, and the output is data saved in temporary storage. Specific operations include sending requests to the API endpoints of each system and receiving data as a response.

[0329] Step 2:

[0330] The server filters the collected data to remove unnecessary and duplicated information. The input is the raw data collected in step 1, and the output is the filtered data. Specifically, it identifies duplicate records and keeps only the most recent data. It also includes the removal of unnecessary temporary notes and duplicate information.

[0331] Step 3:

[0332] The server cleans and formats the filtered data. The input is the data filtered in step 2, and the output is the cleaned data. Specific operations include standardizing date formats, correcting typos, and imputing missing values. For example, standardizing date data in different formats to the "YYYY-MM-DD" format.

[0333] Step 4:

[0334] The server trains a generative AI model (GPT) using the cleaned data. The input is the cleaned data from step 3, and the output is a trained generative AI model. Specifically, the data is tokenized and the model is trained using a machine learning library such as TensorFlow or PyTorch. The model accuracy is improved through 100 epochs of training.

[0335] Step 5:

[0336] The server stores the learned knowledge of the generative AI model in a knowledge base. The input is the generative AI model trained in step 4, and the output is the updated knowledge base. Specifically, the server saves the model parameters in a database and updates the knowledge base accordingly.

[0337] Step 6:

[0338] The user uses a terminal to input a question into the system. The input is the user's question, and the output is the inquiry displayed on the terminal. For example, the user might input, "Please tell me about the application flow for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[0339] Step 7:

[0340] The device passes the input question to the emotion engine and records the emotion data. The input is the question entered by the user, and the output is the recognized emotion data. Specifically, if the user is feeling impatient, that emotion is recorded as "impatience."

[0341] Step 8:

[0342] The server receives a question from the user and analyzes it using natural language processing technology along with the emotional data detected by the emotion engine. The input is the question received in step 6 and the emotional data obtained in step 7, and the output is keywords and request details as the analysis results. For example, the keywords "new project" and "application flow" are extracted.

[0343] Step 9:

[0344] The server uses the generative AI model to generate the optimal answer to the question, taking the user's emotional data into consideration. The input is the analysis results and emotional data obtained in step 8, and the output is the generated answer. Specific operations include generating an answer that includes specific steps and information based on the analysis results. For example, if the user is feeling impatient, the tone of the answer will be softened.

[0345] Step 10:

[0346] The server sends the generated answer to the terminal. The input is the answer generated in step 9, and the output is the answer sent to the terminal. Specific operations include the server sending the answer to the terminal and the user confirming it.

[0347] Step 11:

[0348] The terminal displays the received response to the user. The input is the response sent in step 10, and the output is the response displayed on the user's screen. For example, detailed steps might be displayed, such as, "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0349] (Application example 2)

[0350] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0351] In conventional factory production lines, it was difficult to detect malfunctions or abnormalities early on, and it was also difficult to immediately provide workers with appropriate countermeasures. Furthermore, there was a lack of means to reduce the stress workers felt when malfunctions occurred. This prevented efficient production management and maintenance, leading to problems that led to production line shutdowns and a decline in quality.

[0352] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for saving data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions, means for generating answers based on the analyzed questions, means for recognizing user emotion data, means for adjusting the tone of the answer taking the recognized emotion data into consideration, and means for providing the generated answer to the user. This makes it possible to detect abnormalities in the production line early and quickly provide appropriate countermeasures in a tone that corresponds to the emotion.

[0353] A "data collection system" is a system that automatically collects necessary data from various sensor devices and other information systems within a factory.

[0354] "Filtering" is the process of removing unnecessary information and noise from collected data, making it suitable for analysis and processing.

[0355] "Cleaning" is the process of further refining the filtered data, correcting inconsistencies and inappropriate formats, and compiling it into a unified format.

[0356] A "generative AI model" is an AI model that can perform natural language processing and other tasks based on large amounts of data, specifically referring to generative artificial intelligence models such as GPT.

[0357] A "knowledge base" is a database that stores the knowledge and information of trained generative AI models and makes them easily accessible.

[0358] "Emotion data" is information about emotions inferred from the user's input and actions, and is data that indicates an emotional state such as happiness, dissatisfaction, or impatience.

[0359] "Tone" refers to the style of writing and expression of the generated response, which is appropriately adjusted depending on the user's emotional state.

[0360] The system embodying this invention monitors and controls production lines in a factory, and detects and responds to malfunctions and abnormalities at an early stage. The detailed configuration and operation of this system will be described below.

[0361] System configuration

[0362] Hardware

[0363] Sensors: Attached to various equipment in the factory, they collect data in real time.

[0364] Server: A central location responsible for collecting, filtering, and cleaning data, training generative AI models, and managing the knowledge base.

[0365] User device: A tablet or PC used by workers to enter questions and receive answers.

[0366] software

[0367] CRM, ERP system: Manages data such as production plans and inventory status.

[0368] Natural Language Processing library (NLP): Parse user questions.

[0369] Generative AI model (GPT-3.5 or GPT-4): Generates appropriate answers to user questions.

[0370] Emotion Recognition Engine (Affectiva): Recognizes the user's emotional data and adjusts the tone of the response based on the analysis results.

[0371] Data collection

[0372] The server automatically collects necessary data from sensors in the factory and ERP systems, such as temperature sensors, vibration sensors, and operating hours, in real time.

[0373] Data Filtering and Cleaning

[0374] The collected data is filtered by the server to remove unnecessary information and noise, and then data cleaning is performed to correct inconsistencies and incorrect formats and standardize the data into a consistent format.

[0375] Training generative AI models

[0376] The filtered and cleaned data is used to train a generative AI model (such as GPT-3.5 or GPT-4) that understands the production line's operational information and management procedures to provide optimal answers.

[0377] Knowledge Base Updates

[0378] The knowledge of the learned generative AI model is stored in a knowledge base by the server, which is always updated to the latest version and can be referenced by users as needed.

[0379] Emotional Data Recognition

[0380] When a user uses a terminal to input a question into the system, the emotion recognition engine collects and analyzes emotion data from the user's facial expressions and voice, which is then sent to the server along with the question.

[0381] Generate and provide answers

[0382] The server analyzes the user's question using natural language processing technology and retrieves relevant information from a knowledge base. The generative AI model takes into account the user's emotional data to generate the optimal answer. The generated answer is then adjusted in tone according to the emotional data and provided to the user's device.

[0383] Specific examples

[0384] When a worker at a factory asks, "There is an error with XX device. Please tell me how to fix it," the server processes the following:

[0385] 1. Collect real-time anomaly data from sensors.

[0386] 2. Filter and clean the data.

[0387] 3. Use a trained AI model to identify the cause of the anomaly and how to address it.

[0388] 4. Detect user impatience with an emotion recognition engine.

[0389] 5. Start by saying "Please stay calm" in a gentle tone, then provide detailed instructions on how to proceed.

[0390] Example prompt sentence:

[0391] "Please tell me the maintenance procedures for the production line."

[0392] "XX device has an error. Please tell me how to deal with it."

[0393] "Please tell me the production schedule for next week."

[0394] In this way, the present invention makes it possible to detect abnormalities on the production line early and quickly provide appropriate countermeasures in a tone that reflects the worker's emotions, thereby improving the operating efficiency of the factory and reducing the stress of workers.

[0395] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0396] Step 1:

[0397] The server collects the necessary data from various sensors in the factory and from the ERP system. Specifically, it collects real-time information from temperature sensors, vibration sensors, operating time data, etc. This data collection makes it possible to grasp the latest production status. The input is real-time data from each sensor, and the output is raw data stored on the server.

[0398] Step 2:

[0399] The server filters the collected data to remove unnecessary information and noise, for example, removing sensor error data and redundant data. This process generates a data set with fewer errors. The input is the collected raw data, and the output is the filtered, clean data.

[0400] Step 3:

[0401] The server then cleans the filtered data, correcting any inconsistencies or incorrect formats and standardizing the data (for example, changing all date and time data to the same format). The input is filtered data, and the output is clean, uniform data.

[0402] Step 4:

[0403] The server uses the cleaned data to train a generative AI model. In this process, a generative AI model (such as GPT-3.5 or GPT-4) is retrained using a large amount of data to understand the operational information and management procedures of the production line. The input is clean, unified data, and the output is an updated generative AI model.

[0404] Step 5:

[0405] Using the learned knowledge of the generative AI model, the server updates the knowledge base. This knowledge base stores new knowledge acquired by the generative AI model and allows it to be referenced as needed. The input is the updated knowledge of the generative AI model, and the output is the latest knowledge base.

[0406] Step 6:

[0407] A user uses a terminal to input a question into the system. For example, "An error has occurred with XX device. Please tell me how to deal with it." The input is the user's question, and the output is the question data sent from the terminal to the server.

[0408] Step 7:

[0409] The server uses an emotion recognition engine to recognize the user's emotional data and analyzes it along with the question. This allows the server to understand the user's emotional state (e.g., impatience or dissatisfaction). The input is the user's question and emotional data, and the output is the question data with the emotional data added.

[0410] Step 8:

[0411] The server uses natural language processing technology to analyze the received questions. It extracts the keywords and intent of the questions and searches for related information from a knowledge base. For example, it extracts keywords such as "XX device," "abnormality," and "how to deal with it." The input is question data with emotional data added, and the output is the analyzed question data.

[0412] Step 9:

[0413] The server takes into account the user's emotional data when generating the optimal answer to a question using a generative AI model. For example, if the user is impatient, the tone of the answer can be adjusted to be calmer. The input is the analyzed question data and emotional data, and the output is an answer with the tone adjusted according to the emotion.

[0414] Step 10:

[0415] The server sends the generated answer to the terminal, which then displays it to the user. Specifically, instructions such as "An abnormality has been detected in XX equipment. Please follow the steps below: 1. Check the abnormality. 2. Pause the system. 3. Perform maintenance on the affected area. 4. Perform a restart test" are provided in a friendly tone. The input is an answer with a tone adjusted according to the emotion, and the output is specific instructions displayed on the user's terminal.

[0416] In this way, a series of processing steps can be used to quickly and appropriately respond to abnormalities in the factory production line. In addition, the use of emotion data can reduce worker stress.

[0417] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0418] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0419] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0420] [Second embodiment]

[0421] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0422] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0423] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0424] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0425] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0426] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0427] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0428] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0429] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0430] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0431] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0432] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0433] This invention is an "Internal Knowledge GPT" system that visualizes internal business processes and operational flows and improves the reliability of the latest information. In this system, servers, terminals, and users cooperate to collect, organize, learn, and answer questions about information.

[0434] Data collection

[0435] The server collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.) via APIs and data files. For example, the server retrieves sales transaction data from a customer relationship management system (CRM) and organizes the information.

[0436] Data Filtering and Cleaning

[0437] The server filters and cleans the collected data, which includes removing duplicate data and unnecessary information, and normalizing the data format. For example, the server standardizes date formats and corrects typos in text data.

[0438] Generative AI model training

[0439] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process enables the model to understand the latest business information and operational flows, and generate optimal answers based on that information.

[0440] Knowledge Base Updates

[0441] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[0442] Enter a question

[0443] A user inputs a question into the system using a terminal. For example, the user inputs, "Please tell me about the application flow for a new project."

[0444] question analysis

[0445] The server receives a question from a user, analyzes the question using natural language processing technology, and retrieves relevant information from a knowledge base based on the analysis result.

[0446] Answer generation

[0447] The server uses a generative AI model to generate the best answer to the question, based on the most up-to-date information in the knowledge base.

[0448] Provide answers

[0449] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, it explains the detailed steps in the form of "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0450] Specific examples

[0451] For example, if a user asks, "What is the latest format for sales reports?"

[0452] 1. The server first collects the latest data on sales reports from the CRM system.

[0453] 2. This data is then filtered and cleaned to remove duplicates and extract the information you need.

[0454] 3. The server uses this cleaned data to train a generative AI model.

[0455] 4. When a user enters a question, the server analyzes the question using natural language processing technology and searches the knowledge base for relevant information.

[0456] 5. The generative AI model generates the optimal answer, and the server sends the answer to the device.

[0457] 6. The terminal displays to the user, "The latest sales report format is as follows," and indicates the specific items.

[0458] In this way, the in-house knowledge GPT system provides quick and accurate answers to user questions, improving work efficiency and the reliability of information.

[0459] The processing flow will be explained below.

[0460] Step 1:

[0461] The server automatically collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.). The server uses APIs to obtain data in real time and stores the collected data in temporary storage. For example, the server obtains transaction data and customer information from a CRM system.

[0462] Step 2:

[0463] The server filters the collected data. In this step, the server removes data duplication and noise and extracts only the necessary information. For example, it removes duplicate transaction information from past data logs.

[0464] Step 3:

[0465] The server cleans the filtered data, standardizing the data format and correcting typos and errors (e.g., converting date data in different formats into a standard format).

[0466] Step 4:

[0467] The server feeds the cleaned data to a generative AI model (GPT) to train the model, which then understands the latest business information and operational flows.

[0468] Step 5:

[0469] The server stores the learned knowledge of the generative AI model in a knowledge base, which is periodically updated to reflect the latest information.

[0470] Step 6:

[0471] A user uses a terminal to input a question into the system. For example, the user inputs, "Please tell me about the application flow for a new project."

[0472] Step 7:

[0473] The terminal sends the user's question to the server in text format.

[0474] Step 8:

[0475] The server receives the question, analyzes it using natural language processing technology, extracts keywords and the content of the request, and searches for relevant information from a knowledge base.

[0476] Step 9:

[0477] The server uses the generative AI model to generate the best answer to the question, and creates an answer that includes specific steps and information based on the analysis results.

[0478] Step 10:

[0479] The server sends the generated answer to the terminal, which then communicates the answer to the user in text format.

[0480] Step 11:

[0481] The device will display a response to the user, providing detailed instructions such as, "The flow for applying for a new project is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0482] Example 1

[0483] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0484] In many companies today, information about business processes and operational flows is scattered across various departments and systems, making it difficult to quickly and accurately collect and organize that information and utilize it in a timely manner. This results in the time and effort required to obtain the necessary information, which reduces the efficiency of internal operations. Furthermore, information may not be up-to-date or reliable, which can hinder decision-making and business execution.

[0485] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0486] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for storing knowledge learned by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions using natural language processing technology, means for acquiring information from the knowledge base based on the analyzed questions and generating answers, and means for providing the generated answers to users. This makes it possible to centrally manage data from each business system and train a generative AI model based on the latest filtered and cleaned data, thereby providing quick and accurate answers to user questions.

[0487] A "data collection system" is a system for collecting data from various business systems (CRM, ERP, mail servers, chat systems, etc.).

[0488] "Filtering" is the process of removing duplicate and unnecessary information from collected data.

[0489] "Cleaning" is the process of standardizing the format of collected data and correcting typos and formatting.

[0490] A "generative AI model" is an artificial intelligence model that learns from filtered and cleaned data and is used to generate answers to user questions.

[0491] A "knowledge base" is a database in which the knowledge of a trained generative AI model is stored and referenced as needed.

[0492] "Natural language processing technology" is a technology for analyzing questions from users and understanding context and keywords.

[0493] "Analysis" refers to the process of analyzing a user's question using natural language processing technology to understand the intent of the question and related information.

[0494] "Answer generation" is the process of creating optimal answers using a generative AI model based on the analyzed question.

[0495] "Providing" refers to the act of displaying or transmitting the generated answer to the user.

[0496] This invention is an "Internal Knowledge GPT" system that visualizes internal business processes and operational flows and improves the reliability of the latest information. In this system, servers, terminals, and users cooperate to collect, organize, learn, and answer questions about information.

[0497] Data collection

[0498] The server collects the necessary data from various business systems, such as CRM, ERP, mail servers, and chat systems, via APIs and data files. Specifically, it uses REST API clients and database connectors. For example, the server retrieves sales transaction data from a customer relationship management system (CRM) and organizes the information.

[0499] Data Filtering and Cleaning

[0500] The server filters and cleans the collected data. This includes processing the data using Python's pandas and SQL queries. It also removes duplicate data and standardizes data formats. For example, it removes duplicate data for the same customer and standardizes date formats.

[0501] Generative AI model training

[0502] The server uses the filtered and cleaned data to train a generative AI model (GPT). This process uses machine learning frameworks such as PyTorch and TensorFlow. By training the model with the latest business information and operational flows, the server is able to generate accurate answers that are adapted to user questions.

[0503] Knowledge Base Updates

[0504] The knowledge of the generative AI model trained by the server is stored in a knowledge base. The knowledge base is constructed as an SQL database or a NoSQL database (e.g., MongoDB). The stored knowledge can be referenced by the server as needed.

[0505] Enter a question

[0506] Users use a terminal to input questions into the system, either through a browser-based interface or a dedicated application. For example, a question might be, "Please tell me about the application process for a new project."

[0507] question analysis

[0508] The server receives questions from users and analyzes them using natural language processing technology. Specifically, it uses natural language processing libraries such as spaCy and NLTK. The server extracts keywords from the questions and understands their context, then retrieves the necessary information from a knowledge base.

[0509] Answer generation

[0510] The server uses a generative AI model to generate the best answer to the question, and references the latest business information based on a knowledge base to provide an accurate answer to the user's question.

[0511] Provide answers

[0512] The server sends the generated answer to the terminal, which then displays it to the user, such as "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0513] Examples and prompts

[0514] For example, if a user asks, "What is the latest format for sales reports?"

[0515] 1. The server collects the latest data on sales reports from the CRM system.

[0516] 2. Filter and clean this data to remove duplicates and extract the information you need.

[0517] 3. The server uses this cleaned data to train a generative AI model.

[0518] 4. When a user enters a question into the terminal, the server analyzes the question using natural language processing technology and searches for relevant information from a knowledge base.

[0519] 5. The generative AI model generates the optimal answer, and the server sends the answer to the device.

[0520] 6. The terminal displays to the user, "The latest sales report format is as follows," and indicates the specific items.

[0521] In this way, the in-house knowledge GPT system provides quick and accurate answers to user questions, improving work efficiency and the reliability of information.

[0522] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0523] Step 1:

[0524] The server collects the necessary data from the data collection system. Inputs are APIs and data files from each business system (CRM, ERP, mail server, chat system, etc.). Based on these inputs, the server collects data using a REST API client or database connector. The output is saved as raw data.

[0525] Step 2:

[0526] The server filters and cleans the collected data. The input is the collected raw data. The server processes the data using Python pandas and SQL queries to remove duplicate data and standardize data formats. Specifically, it removes duplicate data for the same customer and standardizes date formats. The output is the filtered and cleaned data.

[0527] Step 3:

[0528] The server trains a generative AI model using the filtered and cleaned data. The input is the cleaned data, and the server trains a generative AI model (GPT) using a machine learning framework such as PyTorch or TensorFlow. The output is a trained generative AI model.

[0529] Step 4:

[0530] The server stores the knowledge of the trained generative AI model in a knowledge base. The input is the trained generative AI model, which the server stores in an SQL database or a NoSQL database (e.g., MongoDB). The output is the data stored as a knowledge base.

[0531] Step 5:

[0532] A user uses a terminal to input a question into the system. The input is the user's question, and the terminal uses a browser-based interface or a dedicated application. A specific example is "Please tell me about the application flow for a new project." The output is the question.

[0533] Step 6:

[0534] The server analyzes questions from users using natural language processing technology. The input is the user's question, and the server analyzes the question using natural language processing libraries such as spaCy or NLTK. The analysis extracts keywords from the question and understands the context. The output is the analysis results.

[0535] Step 7:

[0536] The server retrieves information from the knowledge base based on the analyzed question and generates the optimal answer using a generative AI model. The input is the analysis result and information from the knowledge base. The server creates the optimal answer based on the data it has learned in advance. The output is the generated answer.

[0537] Step 8:

[0538] The server sends the generated answer to the terminal, which then displays it to the user. The input is the generated answer, and the terminal displays the answer using a browser-based interface or a dedicated application. The output is the display of the answer to the user. The specific operation is displayed as detailed steps: "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0539] (Application example 1)

[0540] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0541] In conventional factory management systems, information related to manufacturing processes and maintenance procedures is managed in a decentralized manner, making it difficult for workers and managers to quickly obtain the information they need.It is also difficult to provide accurate information based on the latest business and operational flows, creating challenges in improving business efficiency and reliability.

[0542] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0543] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for saving the data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions, means for generating answers based on the analyzed questions, means for collecting manufacturing process data and maintenance history from a factory management system and generating answers to the questions based thereon, and means for providing the generated answers to users. This enables quick and accurate provision of information on work flows, equipment operating procedures, and maintenance procedures within the factory.

[0544] A "data collection system" is an infrastructure for acquiring data from various business systems, sensors, etc.

[0545] "Filtering" is the process of removing unnecessary information and duplicate data from collected data.

[0546] "Cleaning" is the process of improving data quality by correcting errors and standardizing the format of the data.

[0547] A "generative AI model" is an artificial intelligence model that learns from collected and organized data and generates appropriate answers to user questions.

[0548] A "knowledge base" is a database where information learned by a generative AI model is stored and can be referenced at any time.

[0549] "Accepting a question" is the process by which the system obtains a query from a user as input.

[0550] "Question analysis" is the process of using natural language processing techniques to understand the intent of a user's question and extract relevant information.

[0551] "Generating an answer" is the process of using a generative AI model to create the optimal answer to a user's question.

[0552] A "factory management system" is a system for managing operational information on manufacturing processes and equipment.

[0553] "Manufacturing process data" refers to a series of information related to the manufacturing process of a product, including process progress and quality control data.

[0554] "Maintenance history" refers to records relating to the maintenance and repair of equipment and facilities, including procedures for dealing with problems when they occur and repair history.

[0555] This invention relates to the "Factory Knowledge GPT" system, which enables workers and managers engaged in factory management and maintenance work to quickly and accurately obtain the information they need. This system involves the collaboration of servers, terminals, and users to collect, organize, learn, and answer questions about information.

[0556] Data collection

[0557] The server collects information such as manufacturing process data and equipment maintenance history from the factory management system via APIs and data files. For example, the server obtains daily production line progress data and past maintenance records for each piece of equipment from the factory management system and organizes the information.

[0558] Data Filtering and Cleaning

[0559] The server then filters and cleans the collected data, removing unnecessary information and duplicate data, standardizing data formats, etc. For example, the server may standardize date formats and correct typos in text data to improve data quality.

[0560] Generative AI model training

[0561] The server uses the filtered and cleaned data to train a generative AI model (GPT). This model understands the latest business information and operational flows, and can generate optimal answers to user questions. For example, it learns troubleshooting procedures and equipment maintenance procedures on a production line.

[0562] Knowledge Base Updates

[0563] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[0564] Question input and analysis

[0565] A user inputs a question into the system using a terminal, such as a smartphone or robot interface, asking, "What are the troubleshooting steps for assembly line B?" The server then analyzes the question using natural language processing technology and retrieves relevant information from the knowledge base.

[0566] Answer generation and provision

[0567] The server uses a generative AI model to generate the best answer to the question. The generated answer is based on the latest information in the knowledge base. For example, it provides detailed instructions in the form of "Troubleshooting steps for assembly line B are as follows: 1. Check the power supply of the equipment. 2. Check the location of the sensors. 3. Stop and reset the production line. 4. Check each error log. 5. If you are unsure, contact the administrator."

[0568] Hardware and software used

[0569] The system uses the following hardware and software:

[0570] Hardware: Servers (cloud services recommended, e.g., AWS EC2), smartphones (e.g., iPhone, Android device), factory robots (e.g., robotic arms from ABB or FANUC)

[0571] Software: GPT model (Hugging Face Transformers library), API requests (requests library), data management (JSON format)

[0572] Specific examples

[0573] For example, a factory worker types the following question:

[0574] Question: "What is the maintenance procedure for device A?"

[0575] In response to this question, the server collects and filters the necessary data and uses a generative AI model to generate an answer like this:

[0576] "The maintenance procedure for Equipment A is as follows: 1. Power off the equipment. 2. Put on protective equipment. 3. Inspect each part. 4. Apply lubricant. 5. Report any problems immediately."

[0577] This allows users to quickly obtain accurate and reliable information.

[0578] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0579] Step 1: Data collection

[0580] The server collects manufacturing process data and equipment maintenance history from the factory management system. Specifically, it obtains data through APIs and data files. The input to this step is data from the factory management system, and the output is the collected raw data. The server sends a request to the API endpoint and receives the data in JSON format.

[0581] Step 2: Data filtering

[0582] The server filters the collected data, specifically removing unnecessary information and duplicate data based on manually set filter conditions. The input to this step is the raw data collected in step 1, and the output is the filtered data. The server processes the data by removing unnecessary fields and removing duplicate data.

[0583] Step 3: Data cleaning

[0584] The server cleans the filtered data. Specifically, it standardizes the data format and corrects typos. The input to this step is the data filtered in step 2, and the output is the cleaned data. Standardizing the data format includes standardizing the date format and text encoding. Typos are automatically corrected using a preset dictionary.

[0585] Step 4: Generative AI model training

[0586] The server uses the cleaned data to train a generative AI model (GPT). Specifically, it inputs the cleaned data into the model and updates the generative AI model. The input of this step is the cleaned data, and the output is a trained generative AI model. The server tokenizes the data and trains it to optimize the model parameters.

[0587] Step 5: Update your knowledge base

[0588] The server saves the knowledge of the trained generative AI model in a knowledge base. Specifically, it stores the information learned by the model in a database. The input of this step is the trained generative AI model, and the output is an updated knowledge base. The server inserts the information into the database for the knowledge base and creates the necessary indexes.

[0589] Step 6: Enter your question

[0590] The user uses a terminal to input a question into the system. Specifically, the question is uttered via text or voice through a smartphone or robot interface. The input in this step is the user's question, and the output is digital data of the question. The terminal receives the user's input, converts it into text format, and sends it to the server.

[0591] Step 7: Question Analysis

[0592] The server analyzes the user's question. Specifically, it uses natural language processing technology to understand the intent of the question and searches for relevant information in a knowledge base. The input to this step is the digital data of the user's question, and the output is the analysis result. The server tokenizes the question and analyzes it using natural language processing (NLP) algorithms to identify relevant information.

[0593] Step 8: Answer Generation

[0594] The server uses a generative AI model to generate the optimal answer to the question. Specifically, the model generates an answer based on the analysis results. The input to this step is the analysis results, and the output is the generated answer. The generative AI model generates answer text based on the tokenized analysis results and converts it into a text format.

[0595] Step 9: Provide your answers

[0596] The server sends the generated answer to the terminal, which then displays the answer to the user. Specifically, the answer is displayed on the interface of a smartphone or robot. The input of this step is the generated answer, and the output is the answer displayed to the user. The terminal displays the text data received from the server on the user interface.

[0597] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0598] This invention is an "Internal Knowledge GPT" system that combines an emotion engine to visualize internal business processes and operational flows and improve the reliability of the latest information. In this system, the server, terminals, and users cooperate to collect, organize, learn, answer questions, and recognize emotions.

[0599] Data collection

[0600] The server collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.) via APIs and data files. For example, the server retrieves sales transaction data and customer information from a CRM system and organizes the information.

[0601] Data Filtering and Cleaning

[0602] The server filters the collected data, removing unnecessary information, and then performs cleaning and formatting on the data. This includes removing duplicate data and standardizing formats. For example, the server standardizes date formats and corrects typos and errors.

[0603] Generative AI model training

[0604] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process enables the AI ​​model to understand the latest business information and operational flow, and generate optimal answers based on that information.

[0605] Knowledge Base Updates

[0606] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[0607] Enter a question

[0608] The user inputs a question into the system using a terminal. For example, the user might input, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[0609] question analysis

[0610] The server receives questions from users and analyzes them using natural language processing technology along with emotional data detected by the emotion engine, extracting keywords and requests from the questions and retrieving relevant information from the knowledge base.

[0611] Answer generation

[0612] The server takes into account the user's emotional data when generating the optimal answer to the question using a generative AI model. Based on the analysis results, it creates an answer that includes specific steps and information, and adjusts the tone of the answer. For example, if the user is feeling frustrated, it will soften the tone of the answer.

[0613] Provide answers

[0614] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, it shows detailed steps such as, "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department." The answer is also provided in a tone that corresponds to the user's emotions.

[0615] Specific examples

[0616] For example, if a user asks, "Tell me about your recent sales report?"

[0617] 1. The server first collects the latest data on sales reports from the CRM system.

[0618] 2. This data is then filtered and cleaned to remove duplicates and extract the information you need.

[0619] 3. The server uses this cleaned data to train a generative AI model.

[0620] 4. When the user enters a question, the emotion engine recognizes the user's emotion at the time of entering the question and sends the emotion data to the server.

[0621] 5. The server receives the question, analyzes keywords such as "sales report" and "recent" using natural language processing technology, and searches for related information in the knowledge base.

[0622] 6. The server uses a generative AI model to generate the optimal answer, taking into account the user's emotional data. For example, if the user is dissatisfied, the answer will be delivered in a more friendly tone.

[0623] 7. The server sends the generated answer to the terminal, which displays it to the user.

[0624] In this way, the in-house knowledge GPT system combined with the emotion engine provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[0625] The processing flow will be explained below.

[0626] Step 1:

[0627] The server automatically collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.). The server uses APIs to obtain data in real time and stores the collected data in temporary storage. For example, the server obtains transaction data and customer information from a CRM system.

[0628] Step 2:

[0629] The server filters the collected data. In this step, the server removes data duplication and noise and extracts only the necessary information. For example, it removes duplicate transaction information from past data logs.

[0630] Step 3:

[0631] The server cleans the filtered data, standardizing the data format and correcting typos and errors (e.g., converting date data in different formats into a standard format).

[0632] Step 4:

[0633] The server feeds the cleaned data to a generative AI model (GPT) to train the model, which then understands the latest business information and operational flows.

[0634] Step 5:

[0635] The server stores the learned knowledge of the generative AI model in a knowledge base, which is periodically updated to reflect the latest information.

[0636] Step 6:

[0637] A user inputs a question into the system using a terminal. For example, the user might input, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[0638] Step 7:

[0639] The device sends the user's question and emotional data to the server. The question is sent in text format, and the emotional data is sent as a numerical value or category.

[0640] Step 8:

[0641] The server receives the question and analyzes it using natural language processing technology, extracting keywords and the content of the request, and collating emotional data to search for relevant information from a knowledge base.

[0642] Step 9:

[0643] The server uses a generative AI model to generate the optimal answer to the question, taking into account the emotional data and creating an answer in a tone that corresponds to the user's emotions. For example, if the data indicates anger, the tone of the answer will be calmer.

[0644] Step 10:

[0645] The server sends the generated answer to the terminal, which then communicates the answer to the user in text format.

[0646] Step 11:

[0647] The device will then display answers and sentiment analysis information to the user, providing detailed instructions in a tone that reflects their emotions, such as "The new project application process is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0648] Specific examples

[0649] Example 1: When a user asks, "Tell me about your recent sales report."

[0650] Step 1:

[0651] The server collects the latest sales report data from the CRM system.

[0652] Step 2:

[0653] The server filters the collected data and removes duplicate data.

[0654] Step 3:

[0655] The server standardizes the data format and corrects typos and errors.

[0656] Step 4:

[0657] The server feeds the cleaned data into a generative AI model (GPT) to train the model.

[0658] Step 5:

[0659] The server stores the knowledge of the generated AI model in a knowledge base.

[0660] Step 6:

[0661] The user inputs a question into the terminal, such as "Please tell me about the recent sales report." At this time, the emotion engine recognizes the user's emotion.

[0662] Step 7:

[0663] The device sends the question and emotion data to the server.

[0664] Step 8:

[0665] The server analyzes the question and searches a knowledge base for relevant information.

[0666] Step 9:

[0667] The server uses generative AI models to generate optimal answers, crafting them in a tone that reflects the user's emotions.

[0668] Step 10:

[0669] The server sends the generated response to the terminal.

[0670] Step 11:

[0671] The device will then display the answer and sentiment analysis information to the user, such as "Regarding the recent sales report, please pay attention to the following points: 1. Overview of sales activities 2. Successes and challenges 3. Next action plan," in a tone that reflects the user's sentiment.

[0672] In this way, the in-house knowledge GPT system combined with the emotion engine provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[0673] Example 2

[0674] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0675] Conventional data collection and question answering systems generate answers without considering the user's feelings, which leads to low user satisfaction. Furthermore, it is difficult to provide accurate answers when inaccurate or duplicate data exists.

[0676] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0677] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for storing the data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing emotions, means for analyzing the accepted questions using natural language processing technology, means for generating answers based on the analyzed questions and emotion data, and means for providing the generated answers to the users. This makes it possible to quickly provide accurate and appropriate answers that take the user's emotions into consideration.

[0678] A "data collection system" is a system for efficiently collecting necessary data from various sources.

[0679] "Filtering" is the process of removing unnecessary or redundant information from collected data.

[0680] "Cleaning" is a process that involves correcting data and standardizing its format in order to ensure its accuracy.

[0681] A "generative AI model" is an artificial intelligence model that can learn large amounts of data and generate sentences in natural language like a human.

[0682] A "knowledge base" is a database that systematically stores collected and learned information and can be referenced and used as needed.

[0683] "User" refers to a person or entity that uses the system to search for information or ask a question.

[0684] The "means for accepting questions" is an interface for inputting questions from users into the system.

[0685] "Natural language processing technology" is a technology that enables computers to understand and process human language.

[0686] An "emotion engine" is a system that detects emotions from user input and reactions and analyzes the data.

[0687] "Answer generation means" refers to a method or technology for creating an optimal answer based on the user's question and emotion data.

[0688] "Means for adjusting tone" refers to methods or techniques for appropriately adjusting the expression or tone of the generated response to match the user's emotions.

[0689] This invention is an "Internal Knowledge GPT" system that combines an emotion engine to visualize internal business processes and operational flows and improve the reliability of the latest information. In this system, the server, terminals, and users cooperate to collect, organize, learn, answer questions, and recognize emotions.

[0690] Data collection

[0691] The server collects data from each business system (e.g., customer relationship management system (CRM), enterprise resource planning system (ERP), mail server, chat system, etc.) via APIs and data files. For example, the server obtains sales transaction data and customer information from the customer relationship management system and stores it in storage.

[0692] Data Filtering and Cleaning

[0693] The server filters the collected data, removing unnecessary and duplicate information, and then performs cleaning and formatting. This includes standardizing date formats and correcting typos and errors. For example, it standardizes date data stored in different formats to the "YYYY-MM-DD" format.

[0694] Generative AI model training

[0695] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process uses machine learning libraries such as TensorFlow and PyTorch, allowing the AI ​​model to understand the latest business information and operational flows and generate optimal answers.

[0696] Knowledge Base Updates

[0697] The server stores the learned knowledge of the generative AI model in a knowledge base, which functions as a database that can be referenced whenever necessary, allowing users to access the latest business information.

[0698] Enter a question

[0699] The user uses a terminal to input a question into the system. For example, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and simultaneously records that emotion data.

[0700] question analysis

[0701] The server receives questions from users and analyzes them using natural language processing technology along with emotional data detected by the emotion engine. Keywords and requests from the questions are extracted and analyzed. For example, keywords such as "new project" and "application flow" are extracted.

[0702] Answer generation

[0703] The server uses a generative AI model to generate the best answer to the question, taking into account the user's emotional data. It provides specific instructions and information in a toned manner. For example, if the user is frustrated, it softens the tone of the answer.

[0704] Provide answers

[0705] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, detailed steps are shown in the form of "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0706] Specific examples

[0707] For example, if a user asks, "Tell me about your recent sales report?":

[0708] 1. The server first collects the latest data on sales reports from the customer relationship management system.

[0709] 2. The server filters and cleans the data, removing duplicates and extracting the information you need.

[0710] 3. The server trains the generative AI model using the cleaned data, for example, incorporating the latest sales results and customer feedback.

[0711] 4. When the user enters a question, the device recognizes the emotion data using the emotion engine and sends it to the server. For example, if the user is feeling impatient, that emotion will be recorded.

[0712] 5. The server receives the question, analyzes keywords such as "sales report" and "recent" using natural language processing technology, and searches for related information in the knowledge base.

[0713] 6. The server uses a generative AI model to generate the optimal answer, taking into account the user's emotional data (impatience), and provides a friendly response such as, "Our recent sales report shows that after the mid-October version upgrade, feedback from major customers has been positive, and sales have increased by 20%."

[0714] 7. The server sends the generated answer to the terminal, which displays it to the user.

[0715] In this way, the "Internal Knowledge GPT" system, which combines an emotion engine, provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[0716] Prompt Sentence Examples

[0717] To ask "Tell me about recent sales reports," a user would type the following prompt into the terminal:

[0718] Please tell me about your recent sales report.

[0719] The system analyzes this prompt, collects, organizes, and analyzes the necessary information, and provides a satisfactory answer to the user.

[0720] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0721] Step 1:

[0722] The server collects data from each business system (e.g. CRM, ERP, mail server, chat system, etc.). The input is the API or data file from each business system, and the output is data saved in temporary storage. Specific operations include sending requests to the API endpoints of each system and receiving data as a response.

[0723] Step 2:

[0724] The server filters the collected data to remove unnecessary and duplicated information. The input is the raw data collected in step 1, and the output is the filtered data. Specifically, it identifies duplicate records and keeps only the most recent data. It also includes the removal of unnecessary temporary notes and duplicate information.

[0725] Step 3:

[0726] The server cleans and formats the filtered data. The input is the data filtered in step 2, and the output is the cleaned data. Specific operations include standardizing date formats, correcting typos, and imputing missing values. For example, standardizing date data in different formats to the "YYYY-MM-DD" format.

[0727] Step 4:

[0728] The server trains a generative AI model (GPT) using the cleaned data. The input is the cleaned data from step 3, and the output is a trained generative AI model. Specifically, the data is tokenized and the model is trained using a machine learning library such as TensorFlow or PyTorch. The model accuracy is improved through 100 epochs of training.

[0729] Step 5:

[0730] The server stores the learned knowledge of the generative AI model in a knowledge base. The input is the generative AI model trained in step 4, and the output is the updated knowledge base. Specifically, the server saves the model parameters in a database and updates the knowledge base accordingly.

[0731] Step 6:

[0732] The user uses a terminal to input a question into the system. The input is the user's question, and the output is the inquiry displayed on the terminal. For example, the user might input, "Please tell me about the application flow for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[0733] Step 7:

[0734] The device passes the input question to the emotion engine and records the emotion data. The input is the question entered by the user, and the output is the recognized emotion data. Specifically, if the user is feeling impatient, that emotion is recorded as "impatience."

[0735] Step 8:

[0736] The server receives a question from the user and analyzes it using natural language processing technology along with the emotional data detected by the emotion engine. The input is the question received in step 6 and the emotional data obtained in step 7, and the output is keywords and request details as the analysis results. For example, the keywords "new project" and "application flow" are extracted.

[0737] Step 9:

[0738] The server uses the generative AI model to generate the optimal answer to the question, taking the user's emotional data into consideration. The input is the analysis results and emotional data obtained in step 8, and the output is the generated answer. Specific operations include generating an answer that includes specific steps and information based on the analysis results. For example, if the user is feeling impatient, the tone of the answer will be softened.

[0739] Step 10:

[0740] The server sends the generated answer to the terminal. The input is the answer generated in step 9, and the output is the answer sent to the terminal. Specific operations include the server sending the answer to the terminal and the user confirming it.

[0741] Step 11:

[0742] The terminal displays the received response to the user. The input is the response sent in step 10, and the output is the response displayed on the user's screen. For example, detailed steps might be displayed, such as, "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0743] (Application example 2)

[0744] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0745] In conventional factory production lines, it was difficult to detect malfunctions or abnormalities early on, and it was also difficult to immediately provide workers with appropriate countermeasures. Furthermore, there was a lack of means to reduce the stress workers felt when malfunctions occurred. This prevented efficient production management and maintenance, leading to problems that led to production line shutdowns and a decline in quality.

[0746] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for saving data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions, means for generating answers based on the analyzed questions, means for recognizing user emotion data, means for adjusting the tone of the answer taking the recognized emotion data into consideration, and means for providing the generated answer to the user. This makes it possible to detect abnormalities in the production line early and quickly provide appropriate countermeasures in a tone that corresponds to the emotion.

[0747] A "data collection system" is a system that automatically collects necessary data from various sensor devices and other information systems within a factory.

[0748] "Filtering" is the process of removing unnecessary information and noise from collected data, making it suitable for analysis and processing.

[0749] "Cleaning" is the process of further refining the filtered data, correcting inconsistencies and inappropriate formats, and compiling it into a unified format.

[0750] A "generative AI model" is an AI model that can perform natural language processing and other tasks based on large amounts of data, specifically referring to generative artificial intelligence models such as GPT.

[0751] A "knowledge base" is a database that stores the knowledge and information of trained generative AI models and makes them easily accessible.

[0752] "Emotion data" is information about emotions inferred from the user's input and actions, and is data that indicates an emotional state such as happiness, dissatisfaction, or impatience.

[0753] "Tone" refers to the style of writing and expression of the generated response, which is appropriately adjusted depending on the user's emotional state.

[0754] The system embodying this invention monitors and controls production lines in a factory, and detects and responds to malfunctions and abnormalities at an early stage. The detailed configuration and operation of this system will be described below.

[0755] System configuration

[0756] Hardware

[0757] Sensors: Attached to various equipment in the factory, they collect data in real time.

[0758] Server: A central location responsible for collecting, filtering, and cleaning data, training generative AI models, and managing the knowledge base.

[0759] User device: A tablet or PC used by workers to enter questions and receive answers.

[0760] software

[0761] CRM, ERP system: Manages data such as production plans and inventory status.

[0762] Natural Language Processing library (NLP): Parse user questions.

[0763] Generative AI model (GPT-3.5 or GPT-4): Generates appropriate answers to user questions.

[0764] Emotion Recognition Engine (Affectiva): Recognizes the user's emotional data and adjusts the tone of the response based on the analysis results.

[0765] Data collection

[0766] The server automatically collects necessary data from sensors in the factory and ERP systems, such as temperature sensors, vibration sensors, and operating hours, in real time.

[0767] Data Filtering and Cleaning

[0768] The collected data is filtered by the server to remove unnecessary information and noise, and then data cleaning is performed to correct inconsistencies and incorrect formats and standardize the data into a consistent format.

[0769] Training generative AI models

[0770] The filtered and cleaned data is used to train a generative AI model (such as GPT-3.5 or GPT-4) that understands the production line's operational information and management procedures to provide optimal answers.

[0771] Knowledge Base Updates

[0772] The knowledge of the learned generative AI model is stored in a knowledge base by the server, which is always updated to the latest version and can be referenced by users as needed.

[0773] Emotional Data Recognition

[0774] When a user uses a terminal to input a question into the system, the emotion recognition engine collects and analyzes emotion data from the user's facial expressions and voice, which is then sent to the server along with the question.

[0775] Generate and provide answers

[0776] The server analyzes the user's question using natural language processing technology and retrieves relevant information from a knowledge base. The generative AI model takes into account the user's emotional data to generate the optimal answer. The generated answer is then adjusted in tone according to the emotional data and provided to the user's device.

[0777] Specific examples

[0778] When a worker at a factory asks, "There is an error with XX device. Please tell me how to fix it," the server processes the following:

[0779] 1. Collect real-time anomaly data from sensors.

[0780] 2. Filter and clean the data.

[0781] 3. Use a trained AI model to identify the cause of the anomaly and how to address it.

[0782] 4. Detect user impatience with an emotion recognition engine.

[0783] 5. Start by saying "Please stay calm" in a gentle tone, then provide detailed instructions on how to proceed.

[0784] Example prompt sentence:

[0785] "Please tell me the maintenance procedures for the production line."

[0786] "XX device has an error. Please tell me how to deal with it."

[0787] "Please tell me the production schedule for next week."

[0788] In this way, the present invention makes it possible to detect abnormalities on the production line early and quickly provide appropriate countermeasures in a tone that reflects the worker's emotions, thereby improving the operating efficiency of the factory and reducing the stress of workers.

[0789] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0790] Step 1:

[0791] The server collects the necessary data from various sensors in the factory and from the ERP system. Specifically, it collects real-time information from temperature sensors, vibration sensors, operating time data, etc. This data collection makes it possible to grasp the latest production status. The input is real-time data from each sensor, and the output is raw data stored on the server.

[0792] Step 2:

[0793] The server filters the collected data to remove unnecessary information and noise, for example, removing sensor error data and redundant data. This process generates a data set with fewer errors. The input is the collected raw data, and the output is the filtered, clean data.

[0794] Step 3:

[0795] The server then cleans the filtered data, correcting any inconsistencies or incorrect formats and standardizing the data (for example, changing all date and time data to the same format). The input is filtered data, and the output is clean, uniform data.

[0796] Step 4:

[0797] The server uses the cleaned data to train a generative AI model. In this process, a generative AI model (such as GPT-3.5 or GPT-4) is retrained using a large amount of data to understand the operational information and management procedures of the production line. The input is clean, unified data, and the output is an updated generative AI model.

[0798] Step 5:

[0799] Using the learned knowledge of the generative AI model, the server updates the knowledge base. This knowledge base stores new knowledge acquired by the generative AI model and allows it to be referenced as needed. The input is the updated knowledge of the generative AI model, and the output is the latest knowledge base.

[0800] Step 6:

[0801] A user uses a terminal to input a question into the system. For example, "An error has occurred with XX device. Please tell me how to deal with it." The input is the user's question, and the output is the question data sent from the terminal to the server.

[0802] Step 7:

[0803] The server uses an emotion recognition engine to recognize the user's emotional data and analyzes it along with the question. This allows the server to understand the user's emotional state (e.g., impatience or dissatisfaction). The input is the user's question and emotional data, and the output is the question data with the emotional data added.

[0804] Step 8:

[0805] The server uses natural language processing technology to analyze the received questions. It extracts the keywords and intent of the questions and searches for related information from a knowledge base. For example, it extracts keywords such as "XX device," "abnormality," and "how to deal with it." The input is question data with emotional data added, and the output is the analyzed question data.

[0806] Step 9:

[0807] The server takes into account the user's emotional data when generating the optimal answer to a question using a generative AI model. For example, if the user is impatient, the tone of the answer can be adjusted to be calmer. The input is the analyzed question data and emotional data, and the output is an answer with the tone adjusted according to the emotion.

[0808] Step 10:

[0809] The server sends the generated answer to the terminal, which then displays it to the user. Specifically, instructions such as "An abnormality has been detected in XX equipment. Please follow the steps below: 1. Check the abnormality. 2. Pause the system. 3. Perform maintenance on the affected area. 4. Perform a restart test" are provided in a friendly tone. The input is an answer with a tone adjusted according to the emotion, and the output is specific instructions displayed on the user's terminal.

[0810] In this way, a series of processing steps can be used to quickly and appropriately respond to abnormalities in the factory production line. In addition, the use of emotion data can reduce worker stress.

[0811] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0812] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0813] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0814] [Third embodiment]

[0815] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0816] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0817] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0818] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0819] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0820] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0821] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0822] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0823] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0824] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0825] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0826] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0827] This invention is an "Internal Knowledge GPT" system that visualizes internal business processes and operational flows and improves the reliability of the latest information. In this system, servers, terminals, and users cooperate to collect, organize, learn, and answer questions about information.

[0828] Data collection

[0829] The server collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.) via APIs and data files. For example, the server retrieves sales transaction data from a customer relationship management system (CRM) and organizes the information.

[0830] Data Filtering and Cleaning

[0831] The server filters and cleans the collected data, which includes removing duplicate data and unnecessary information, and normalizing the data format. For example, the server standardizes date formats and corrects typos in text data.

[0832] Generative AI model training

[0833] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process enables the model to understand the latest business information and operational flows, and generate optimal answers based on that information.

[0834] Knowledge Base Updates

[0835] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[0836] Enter a question

[0837] A user inputs a question into the system using a terminal. For example, the user inputs, "Please tell me about the application flow for a new project."

[0838] question analysis

[0839] The server receives a question from a user, analyzes the question using natural language processing technology, and retrieves relevant information from a knowledge base based on the analysis result.

[0840] Answer generation

[0841] The server uses a generative AI model to generate the best answer to the question, based on the most up-to-date information in the knowledge base.

[0842] Provide answers

[0843] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, it explains the detailed steps in the form of "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0844] Specific examples

[0845] For example, if a user asks, "What is the latest format for sales reports?"

[0846] 1. The server first collects the latest data on sales reports from the CRM system.

[0847] 2. This data is then filtered and cleaned to remove duplicates and extract the information you need.

[0848] 3. The server uses this cleaned data to train a generative AI model.

[0849] 4. When a user enters a question, the server analyzes the question using natural language processing technology and searches the knowledge base for relevant information.

[0850] 5. The generative AI model generates the optimal answer, and the server sends the answer to the device.

[0851] 6. The terminal displays to the user, "The latest sales report format is as follows," and indicates the specific items.

[0852] In this way, the in-house knowledge GPT system provides quick and accurate answers to user questions, improving work efficiency and the reliability of information.

[0853] The processing flow will be explained below.

[0854] Step 1:

[0855] The server automatically collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.). The server uses APIs to obtain data in real time and stores the collected data in temporary storage. For example, the server obtains transaction data and customer information from a CRM system.

[0856] Step 2:

[0857] The server filters the collected data. In this step, the server removes data duplication and noise and extracts only the necessary information. For example, it removes duplicate transaction information from past data logs.

[0858] Step 3:

[0859] The server cleans the filtered data, standardizing the data format and correcting typos and errors (e.g., converting date data in different formats into a standard format).

[0860] Step 4:

[0861] The server feeds the cleaned data to a generative AI model (GPT) to train the model, which then understands the latest business information and operational flows.

[0862] Step 5:

[0863] The server stores the learned knowledge of the generative AI model in a knowledge base, which is periodically updated to reflect the latest information.

[0864] Step 6:

[0865] A user uses a terminal to input a question into the system. For example, the user inputs, "Please tell me about the application flow for a new project."

[0866] Step 7:

[0867] The terminal sends the user's question to the server in text format.

[0868] Step 8:

[0869] The server receives the question, analyzes it using natural language processing technology, extracts keywords and the content of the request, and searches for relevant information from a knowledge base.

[0870] Step 9:

[0871] The server uses the generative AI model to generate the best answer to the question, and creates an answer that includes specific steps and information based on the analysis results.

[0872] Step 10:

[0873] The server sends the generated answer to the terminal, which then communicates the answer to the user in text format.

[0874] Step 11:

[0875] The device will display a response to the user, providing detailed instructions such as, "The flow for applying for a new project is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0876] Example 1

[0877] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0878] In many companies today, information about business processes and operational flows is scattered across various departments and systems, making it difficult to quickly and accurately collect and organize that information and utilize it in a timely manner. This results in the time and effort required to obtain the necessary information, which reduces the efficiency of internal operations. Furthermore, information may not be up-to-date or reliable, which can hinder decision-making and business execution.

[0879] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0880] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for storing knowledge learned by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions using natural language processing technology, means for acquiring information from the knowledge base based on the analyzed questions and generating answers, and means for providing the generated answers to users. This makes it possible to centrally manage data from each business system and train a generative AI model based on the latest filtered and cleaned data, thereby providing quick and accurate answers to user questions.

[0881] A "data collection system" is a system for collecting data from various business systems (CRM, ERP, mail servers, chat systems, etc.).

[0882] "Filtering" is the process of removing duplicate and unnecessary information from collected data.

[0883] "Cleaning" is the process of standardizing the format of collected data and correcting typos and formatting.

[0884] A "generative AI model" is an artificial intelligence model that learns from filtered and cleaned data and is used to generate answers to user questions.

[0885] A "knowledge base" is a database in which the knowledge of a trained generative AI model is stored and referenced as needed.

[0886] "Natural language processing technology" is a technology for analyzing questions from users and understanding context and keywords.

[0887] "Analysis" refers to the process of analyzing a user's question using natural language processing technology to understand the intent of the question and related information.

[0888] "Answer generation" is the process of creating optimal answers using a generative AI model based on the analyzed question.

[0889] "Providing" refers to the act of displaying or transmitting the generated answer to the user.

[0890] This invention is an "Internal Knowledge GPT" system that visualizes internal business processes and operational flows and improves the reliability of the latest information. In this system, servers, terminals, and users cooperate to collect, organize, learn, and answer questions about information.

[0891] Data collection

[0892] The server collects the necessary data from various business systems, such as CRM, ERP, mail servers, and chat systems, via APIs and data files. Specifically, it uses REST API clients and database connectors. For example, the server retrieves sales transaction data from a customer relationship management system (CRM) and organizes the information.

[0893] Data Filtering and Cleaning

[0894] The server filters and cleans the collected data. This includes processing the data using Python's pandas and SQL queries. It also removes duplicate data and standardizes data formats. For example, it removes duplicate data for the same customer and standardizes date formats.

[0895] Generative AI model training

[0896] The server uses the filtered and cleaned data to train a generative AI model (GPT). This process uses machine learning frameworks such as PyTorch and TensorFlow. By training the model with the latest business information and operational flows, the server is able to generate accurate answers that are adapted to user questions.

[0897] Knowledge Base Updates

[0898] The knowledge of the generative AI model trained by the server is stored in a knowledge base. The knowledge base is constructed as an SQL database or a NoSQL database (e.g., MongoDB). The stored knowledge can be referenced by the server as needed.

[0899] Enter a question

[0900] Users use a terminal to input questions into the system, either through a browser-based interface or a dedicated application. For example, a question might be, "Please tell me about the application process for a new project."

[0901] question analysis

[0902] The server receives questions from users and analyzes them using natural language processing technology. Specifically, it uses natural language processing libraries such as spaCy and NLTK. The server extracts keywords from the questions and understands their context, then retrieves the necessary information from a knowledge base.

[0903] Answer generation

[0904] The server uses a generative AI model to generate the best answer to the question, and references the latest business information based on a knowledge base to provide an accurate answer to the user's question.

[0905] Provide answers

[0906] The server sends the generated answer to the terminal, which then displays it to the user, such as "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0907] Examples and prompts

[0908] For example, if a user asks, "What is the latest format for sales reports?"

[0909] 1. The server collects the latest data on sales reports from the CRM system.

[0910] 2. Filter and clean this data to remove duplicates and extract the information you need.

[0911] 3. The server uses this cleaned data to train a generative AI model.

[0912] 4. When a user enters a question into the terminal, the server analyzes the question using natural language processing technology and searches for relevant information from a knowledge base.

[0913] 5. The generative AI model generates the optimal answer, and the server sends the answer to the device.

[0914] 6. The terminal displays to the user, "The latest sales report format is as follows," and indicates the specific items.

[0915] In this way, the in-house knowledge GPT system provides quick and accurate answers to user questions, improving work efficiency and the reliability of information.

[0916] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0917] Step 1:

[0918] The server collects the necessary data from the data collection system. Inputs are APIs and data files from each business system (CRM, ERP, mail server, chat system, etc.). Based on these inputs, the server collects data using a REST API client or database connector. The output is saved as raw data.

[0919] Step 2:

[0920] The server filters and cleans the collected data. The input is the collected raw data. The server processes the data using Python pandas and SQL queries to remove duplicate data and standardize data formats. Specifically, it removes duplicate data for the same customer and standardizes date formats. The output is the filtered and cleaned data.

[0921] Step 3:

[0922] The server trains a generative AI model using the filtered and cleaned data. The input is the cleaned data, and the server trains a generative AI model (GPT) using a machine learning framework such as PyTorch or TensorFlow. The output is a trained generative AI model.

[0923] Step 4:

[0924] The server stores the knowledge of the trained generative AI model in a knowledge base. The input is the trained generative AI model, which the server stores in an SQL database or a NoSQL database (e.g., MongoDB). The output is the data stored as a knowledge base.

[0925] Step 5:

[0926] A user uses a terminal to input a question into the system. The input is the user's question, and the terminal uses a browser-based interface or a dedicated application. A specific example is "Please tell me about the application flow for a new project." The output is the question.

[0927] Step 6:

[0928] The server analyzes questions from users using natural language processing technology. The input is the user's question, and the server analyzes the question using natural language processing libraries such as spaCy or NLTK. The analysis extracts keywords from the question and understands the context. The output is the analysis results.

[0929] Step 7:

[0930] The server retrieves information from the knowledge base based on the analyzed question and generates the optimal answer using a generative AI model. The input is the analysis result and information from the knowledge base. The server creates the optimal answer based on the data it has learned in advance. The output is the generated answer.

[0931] Step 8:

[0932] The server sends the generated answer to the terminal, which then displays it to the user. The input is the generated answer, and the terminal displays the answer using a browser-based interface or a dedicated application. The output is the display of the answer to the user. The specific operation is displayed as detailed steps: "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[0933] (Application example 1)

[0934] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0935] In conventional factory management systems, information related to manufacturing processes and maintenance procedures is managed in a decentralized manner, making it difficult for workers and managers to quickly obtain the information they need.It is also difficult to provide accurate information based on the latest business and operational flows, creating challenges in improving business efficiency and reliability.

[0936] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0937] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for saving the data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions, means for generating answers based on the analyzed questions, means for collecting manufacturing process data and maintenance history from a factory management system and generating answers to the questions based thereon, and means for providing the generated answers to users. This enables quick and accurate provision of information on work flows, equipment operating procedures, and maintenance procedures within the factory.

[0938] A "data collection system" is an infrastructure for acquiring data from various business systems, sensors, etc.

[0939] "Filtering" is the process of removing unnecessary information and duplicate data from collected data.

[0940] "Cleaning" is the process of improving data quality by correcting errors and standardizing the format of the data.

[0941] A "generative AI model" is an artificial intelligence model that learns from collected and organized data and generates appropriate answers to user questions.

[0942] A "knowledge base" is a database where information learned by a generative AI model is stored and can be referenced at any time.

[0943] "Accepting a question" is the process by which the system obtains a query from a user as input.

[0944] "Question analysis" is the process of using natural language processing techniques to understand the intent of a user's question and extract relevant information.

[0945] "Generating an answer" is the process of using a generative AI model to create the optimal answer to a user's question.

[0946] A "factory management system" is a system for managing operational information on manufacturing processes and equipment.

[0947] "Manufacturing process data" refers to a series of information related to the manufacturing process of a product, including process progress and quality control data.

[0948] "Maintenance history" refers to records relating to the maintenance and repair of equipment and facilities, including procedures for dealing with problems when they occur and repair history.

[0949] This invention relates to the "Factory Knowledge GPT" system, which enables workers and managers engaged in factory management and maintenance work to quickly and accurately obtain the information they need. This system involves the collaboration of servers, terminals, and users to collect, organize, learn, and answer questions about information.

[0950] Data collection

[0951] The server collects information such as manufacturing process data and equipment maintenance history from the factory management system via APIs and data files. For example, the server obtains daily production line progress data and past maintenance records for each piece of equipment from the factory management system and organizes the information.

[0952] Data Filtering and Cleaning

[0953] The server then filters and cleans the collected data, removing unnecessary information and duplicate data, standardizing data formats, etc. For example, the server may standardize date formats and correct typos in text data to improve data quality.

[0954] Generative AI model training

[0955] The server uses the filtered and cleaned data to train a generative AI model (GPT). This model understands the latest business information and operational flows, and can generate optimal answers to user questions. For example, it learns troubleshooting procedures and equipment maintenance procedures on a production line.

[0956] Knowledge Base Updates

[0957] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[0958] Question input and analysis

[0959] A user inputs a question into the system using a terminal, such as a smartphone or robot interface, asking, "What are the troubleshooting steps for assembly line B?" The server then analyzes the question using natural language processing technology and retrieves relevant information from the knowledge base.

[0960] Answer generation and provision

[0961] The server uses a generative AI model to generate the best answer to the question. The generated answer is based on the latest information in the knowledge base. For example, it provides detailed instructions in the form of "Troubleshooting steps for assembly line B are as follows: 1. Check the power supply of the equipment. 2. Check the location of the sensors. 3. Stop and reset the production line. 4. Check each error log. 5. If you are unsure, contact the administrator."

[0962] Hardware and software used

[0963] The system uses the following hardware and software:

[0964] Hardware: Servers (cloud services recommended, e.g., AWS EC2), smartphones (e.g., iPhone, Android device), factory robots (e.g., robotic arms from ABB or FANUC)

[0965] Software: GPT model (Hugging Face Transformers library), API requests (requests library), data management (JSON format)

[0966] Specific examples

[0967] For example, a factory worker types the following question:

[0968] Question: "What is the maintenance procedure for device A?"

[0969] In response to this question, the server collects and filters the necessary data and uses a generative AI model to generate an answer like this:

[0970] "The maintenance procedure for Equipment A is as follows: 1. Power off the equipment. 2. Put on protective equipment. 3. Inspect each part. 4. Apply lubricant. 5. Report any problems immediately."

[0971] This allows users to quickly obtain accurate and reliable information.

[0972] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0973] Step 1: Data collection

[0974] The server collects manufacturing process data and equipment maintenance history from the factory management system. Specifically, it obtains data through APIs and data files. The input to this step is data from the factory management system, and the output is the collected raw data. The server sends a request to the API endpoint and receives the data in JSON format.

[0975] Step 2: Data filtering

[0976] The server filters the collected data, specifically removing unnecessary information and duplicate data based on manually set filter conditions. The input to this step is the raw data collected in step 1, and the output is the filtered data. The server processes the data by removing unnecessary fields and removing duplicate data.

[0977] Step 3: Data cleaning

[0978] The server cleans the filtered data. Specifically, it standardizes the data format and corrects typos. The input to this step is the data filtered in step 2, and the output is the cleaned data. Standardizing the data format includes standardizing the date format and text encoding. Typos are automatically corrected using a preset dictionary.

[0979] Step 4: Generative AI model training

[0980] The server uses the cleaned data to train a generative AI model (GPT). Specifically, it inputs the cleaned data into the model and updates the generative AI model. The input of this step is the cleaned data, and the output is a trained generative AI model. The server tokenizes the data and trains it to optimize the model parameters.

[0981] Step 5: Update your knowledge base

[0982] The server saves the knowledge of the trained generative AI model in a knowledge base. Specifically, it stores the information learned by the model in a database. The input of this step is the trained generative AI model, and the output is an updated knowledge base. The server inserts the information into the database for the knowledge base and creates the necessary indexes.

[0983] Step 6: Enter your question

[0984] The user uses a terminal to input a question into the system. Specifically, the question is uttered via text or voice through a smartphone or robot interface. The input in this step is the user's question, and the output is digital data of the question. The terminal receives the user's input, converts it into text format, and sends it to the server.

[0985] Step 7: Question Analysis

[0986] The server analyzes the user's question. Specifically, it uses natural language processing technology to understand the intent of the question and searches for relevant information in a knowledge base. The input to this step is the digital data of the user's question, and the output is the analysis result. The server tokenizes the question and analyzes it using natural language processing (NLP) algorithms to identify relevant information.

[0987] Step 8: Answer Generation

[0988] The server uses a generative AI model to generate the optimal answer to the question. Specifically, the model generates an answer based on the analysis results. The input to this step is the analysis results, and the output is the generated answer. The generative AI model generates answer text based on the tokenized analysis results and converts it into a text format.

[0989] Step 9: Provide your answers

[0990] The server sends the generated answer to the terminal, which then displays the answer to the user. Specifically, the answer is displayed on the interface of a smartphone or robot. The input of this step is the generated answer, and the output is the answer displayed to the user. The terminal displays the text data received from the server on the user interface.

[0991] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0992] This invention is an "Internal Knowledge GPT" system that combines an emotion engine to visualize internal business processes and operational flows and improve the reliability of the latest information. In this system, the server, terminals, and users cooperate to collect, organize, learn, answer questions, and recognize emotions.

[0993] Data collection

[0994] The server collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.) via APIs and data files. For example, the server retrieves sales transaction data and customer information from a CRM system and organizes the information.

[0995] Data Filtering and Cleaning

[0996] The server filters the collected data, removing unnecessary information, and then performs cleaning and formatting on the data. This includes removing duplicate data and standardizing formats. For example, the server standardizes date formats and corrects typos and errors.

[0997] Generative AI model training

[0998] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process enables the AI ​​model to understand the latest business information and operational flow, and generate optimal answers based on that information.

[0999] Knowledge Base Updates

[1000] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[1001] Enter a question

[1002] The user inputs a question into the system using a terminal. For example, the user might input, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[1003] question analysis

[1004] The server receives questions from users and analyzes them using natural language processing technology along with emotional data detected by the emotion engine, extracting keywords and requests from the questions and retrieving relevant information from the knowledge base.

[1005] Answer generation

[1006] The server takes into account the user's emotional data when generating the optimal answer to the question using a generative AI model. Based on the analysis results, it creates an answer that includes specific steps and information, and adjusts the tone of the answer. For example, if the user is feeling frustrated, it will soften the tone of the answer.

[1007] Provide answers

[1008] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, it shows detailed steps such as, "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department." The answer is also provided in a tone that corresponds to the user's emotions.

[1009] Specific examples

[1010] For example, if a user asks, "Tell me about your recent sales report?"

[1011] 1. The server first collects the latest data on sales reports from the CRM system.

[1012] 2. This data is then filtered and cleaned to remove duplicates and extract the information you need.

[1013] 3. The server uses this cleaned data to train a generative AI model.

[1014] 4. When the user enters a question, the emotion engine recognizes the user's emotion at the time of entering the question and sends the emotion data to the server.

[1015] 5. The server receives the question, analyzes keywords such as "sales report" and "recent" using natural language processing technology, and searches for related information in the knowledge base.

[1016] 6. The server uses a generative AI model to generate the optimal answer, taking into account the user's emotional data. For example, if the user is dissatisfied, the answer will be delivered in a more friendly tone.

[1017] 7. The server sends the generated answer to the terminal, which displays it to the user.

[1018] In this way, the in-house knowledge GPT system combined with the emotion engine provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[1019] The processing flow will be explained below.

[1020] Step 1:

[1021] The server automatically collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.). The server uses APIs to obtain data in real time and stores the collected data in temporary storage. For example, the server obtains transaction data and customer information from a CRM system.

[1022] Step 2:

[1023] The server filters the collected data. In this step, the server removes data duplication and noise and extracts only the necessary information. For example, it removes duplicate transaction information from past data logs.

[1024] Step 3:

[1025] The server cleans the filtered data, standardizing the data format and correcting typos and errors (e.g., converting date data in different formats into a standard format).

[1026] Step 4:

[1027] The server feeds the cleaned data to a generative AI model (GPT) to train the model, which then understands the latest business information and operational flows.

[1028] Step 5:

[1029] The server stores the learned knowledge of the generative AI model in a knowledge base, which is periodically updated to reflect the latest information.

[1030] Step 6:

[1031] A user inputs a question into the system using a terminal. For example, the user might input, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[1032] Step 7:

[1033] The device sends the user's question and emotional data to the server. The question is sent in text format, and the emotional data is sent as a numerical value or category.

[1034] Step 8:

[1035] The server receives the question and analyzes it using natural language processing technology, extracting keywords and the content of the request, and collating emotional data to search for relevant information from a knowledge base.

[1036] Step 9:

[1037] The server uses a generative AI model to generate the optimal answer to the question, taking into account the emotional data and creating an answer in a tone that corresponds to the user's emotions. For example, if the data indicates anger, the tone of the answer will be calmer.

[1038] Step 10:

[1039] The server sends the generated answer to the terminal, which then communicates the answer to the user in text format.

[1040] Step 11:

[1041] The device will then display answers and sentiment analysis information to the user, providing detailed instructions in a tone that reflects their emotions, such as "The new project application process is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[1042] Specific examples

[1043] Example 1: When a user asks, "Tell me about your recent sales report."

[1044] Step 1:

[1045] The server collects the latest sales report data from the CRM system.

[1046] Step 2:

[1047] The server filters the collected data and removes duplicate data.

[1048] Step 3:

[1049] The server standardizes the data format and corrects typos and errors.

[1050] Step 4:

[1051] The server feeds the cleaned data into a generative AI model (GPT) to train the model.

[1052] Step 5:

[1053] The server stores the knowledge of the generated AI model in a knowledge base.

[1054] Step 6:

[1055] The user inputs a question into the terminal, such as "Please tell me about the recent sales report." At this time, the emotion engine recognizes the user's emotion.

[1056] Step 7:

[1057] The device sends the question and emotion data to the server.

[1058] Step 8:

[1059] The server analyzes the question and searches a knowledge base for relevant information.

[1060] Step 9:

[1061] The server uses generative AI models to generate optimal answers, crafting them in a tone that reflects the user's emotions.

[1062] Step 10:

[1063] The server sends the generated response to the terminal.

[1064] Step 11:

[1065] The device will then display the answer and sentiment analysis information to the user, such as "Regarding the recent sales report, please pay attention to the following points: 1. Overview of sales activities 2. Successes and challenges 3. Next action plan," in a tone that reflects the user's sentiment.

[1066] In this way, the in-house knowledge GPT system combined with the emotion engine provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[1067] Example 2

[1068] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1069] Conventional data collection and question answering systems generate answers without considering the user's feelings, which leads to low user satisfaction. Furthermore, it is difficult to provide accurate answers when inaccurate or duplicate data exists.

[1070] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1071] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for storing the data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing emotions, means for analyzing the accepted questions using natural language processing technology, means for generating answers based on the analyzed questions and emotion data, and means for providing the generated answers to the users. This makes it possible to quickly provide accurate and appropriate answers that take the user's emotions into consideration.

[1072] A "data collection system" is a system for efficiently collecting necessary data from various sources.

[1073] "Filtering" is the process of removing unnecessary or redundant information from collected data.

[1074] "Cleaning" is a process that involves correcting data and standardizing its format in order to ensure its accuracy.

[1075] A "generative AI model" is an artificial intelligence model that can learn large amounts of data and generate sentences in natural language like a human.

[1076] A "knowledge base" is a database that systematically stores collected and learned information and can be referenced and used as needed.

[1077] "User" refers to a person or entity that uses the system to search for information or ask a question.

[1078] The "means for accepting questions" is an interface for inputting questions from users into the system.

[1079] "Natural language processing technology" is a technology that enables computers to understand and process human language.

[1080] An "emotion engine" is a system that detects emotions from user input and reactions and analyzes the data.

[1081] "Answer generation means" refers to a method or technology for creating an optimal answer based on the user's question and emotion data.

[1082] "Means for adjusting tone" refers to methods or techniques for appropriately adjusting the expression or tone of the generated response to match the user's emotions.

[1083] This invention is an "Internal Knowledge GPT" system that combines an emotion engine to visualize internal business processes and operational flows and improve the reliability of the latest information. In this system, the server, terminals, and users cooperate to collect, organize, learn, answer questions, and recognize emotions.

[1084] Data collection

[1085] The server collects data from each business system (e.g., customer relationship management system (CRM), enterprise resource planning system (ERP), mail server, chat system, etc.) via APIs and data files. For example, the server obtains sales transaction data and customer information from the customer relationship management system and stores it in storage.

[1086] Data Filtering and Cleaning

[1087] The server filters the collected data, removing unnecessary and duplicate information, and then performs cleaning and formatting. This includes standardizing date formats and correcting typos and errors. For example, it standardizes date data stored in different formats to the "YYYY-MM-DD" format.

[1088] Generative AI model training

[1089] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process uses machine learning libraries such as TensorFlow and PyTorch, allowing the AI ​​model to understand the latest business information and operational flows and generate optimal answers.

[1090] Knowledge Base Updates

[1091] The server stores the learned knowledge of the generative AI model in a knowledge base, which functions as a database that can be referenced whenever necessary, allowing users to access the latest business information.

[1092] Enter a question

[1093] The user uses a terminal to input a question into the system. For example, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and simultaneously records that emotion data.

[1094] question analysis

[1095] The server receives questions from users and analyzes them using natural language processing technology along with emotional data detected by the emotion engine. Keywords and requests from the questions are extracted and analyzed. For example, keywords such as "new project" and "application flow" are extracted.

[1096] Answer generation

[1097] The server uses a generative AI model to generate the best answer to the question, taking into account the user's emotional data. It provides specific instructions and information in a toned manner. For example, if the user is frustrated, it softens the tone of the answer.

[1098] Provide answers

[1099] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, detailed steps are shown in the form of "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[1100] Specific examples

[1101] For example, if a user asks, "Tell me about your recent sales report?":

[1102] 1. The server first collects the latest data on sales reports from the customer relationship management system.

[1103] 2. The server filters and cleans the data, removing duplicates and extracting the information you need.

[1104] 3. The server trains the generative AI model using the cleaned data, for example, incorporating the latest sales results and customer feedback.

[1105] 4. When the user enters a question, the device recognizes the emotion data using the emotion engine and sends it to the server. For example, if the user is feeling impatient, that emotion will be recorded.

[1106] 5. The server receives the question, analyzes keywords such as "sales report" and "recent" using natural language processing technology, and searches for related information in the knowledge base.

[1107] 6. The server uses a generative AI model to generate the optimal answer, taking into account the user's emotional data (impatience), and provides a friendly response such as, "Our recent sales report shows that after the mid-October version upgrade, feedback from major customers has been positive, and sales have increased by 20%."

[1108] 7. The server sends the generated answer to the terminal, which displays it to the user.

[1109] In this way, the "Internal Knowledge GPT" system, which combines an emotion engine, provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[1110] Prompt Sentence Examples

[1111] To ask "Tell me about recent sales reports," a user would type the following prompt into the terminal:

[1112] Please tell me about your recent sales report.

[1113] The system analyzes this prompt, collects, organizes, and analyzes the necessary information, and provides a satisfactory answer to the user.

[1114] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1115] Step 1:

[1116] The server collects data from each business system (e.g. CRM, ERP, mail server, chat system, etc.). The input is the API or data file from each business system, and the output is data saved in temporary storage. Specific operations include sending requests to the API endpoints of each system and receiving data as a response.

[1117] Step 2:

[1118] The server filters the collected data to remove unnecessary and duplicated information. The input is the raw data collected in step 1, and the output is the filtered data. Specifically, it identifies duplicate records and keeps only the most recent data. It also includes the removal of unnecessary temporary notes and duplicate information.

[1119] Step 3:

[1120] The server cleans and formats the filtered data. The input is the data filtered in step 2, and the output is the cleaned data. Specific operations include standardizing date formats, correcting typos, and imputing missing values. For example, standardizing date data in different formats to the "YYYY-MM-DD" format.

[1121] Step 4:

[1122] The server trains a generative AI model (GPT) using the cleaned data. The input is the cleaned data from step 3, and the output is a trained generative AI model. Specifically, the data is tokenized and the model is trained using a machine learning library such as TensorFlow or PyTorch. The model accuracy is improved through 100 epochs of training.

[1123] Step 5:

[1124] The server stores the learned knowledge of the generative AI model in a knowledge base. The input is the generative AI model trained in step 4, and the output is the updated knowledge base. Specifically, the server saves the model parameters in a database and updates the knowledge base accordingly.

[1125] Step 6:

[1126] The user uses a terminal to input a question into the system. The input is the user's question, and the output is the inquiry displayed on the terminal. For example, the user might input, "Please tell me about the application flow for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[1127] Step 7:

[1128] The device passes the input question to the emotion engine and records the emotion data. The input is the question entered by the user, and the output is the recognized emotion data. Specifically, if the user is feeling impatient, that emotion is recorded as "impatience."

[1129] Step 8:

[1130] The server receives a question from the user and analyzes it using natural language processing technology along with the emotional data detected by the emotion engine. The input is the question received in step 6 and the emotional data obtained in step 7, and the output is keywords and request details as the analysis results. For example, the keywords "new project" and "application flow" are extracted.

[1131] Step 9:

[1132] The server uses the generative AI model to generate the optimal answer to the question, taking the user's emotional data into consideration. The input is the analysis results and emotional data obtained in step 8, and the output is the generated answer. Specific operations include generating an answer that includes specific steps and information based on the analysis results. For example, if the user is feeling impatient, the tone of the answer will be softened.

[1133] Step 10:

[1134] The server sends the generated answer to the terminal. The input is the answer generated in step 9, and the output is the answer sent to the terminal. Specific operations include the server sending the answer to the terminal and the user confirming it.

[1135] Step 11:

[1136] The terminal displays the received response to the user. The input is the response sent in step 10, and the output is the response displayed on the user's screen. For example, detailed steps might be displayed, such as, "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[1137] (Application example 2)

[1138] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1139] In conventional factory production lines, it was difficult to detect malfunctions or abnormalities early on, and it was also difficult to immediately provide workers with appropriate countermeasures. Furthermore, there was a lack of means to reduce the stress workers felt when malfunctions occurred. This prevented efficient production management and maintenance, leading to problems that led to production line shutdowns and a decline in quality.

[1140] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for saving data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions, means for generating answers based on the analyzed questions, means for recognizing user emotion data, means for adjusting the tone of the answer taking the recognized emotion data into consideration, and means for providing the generated answer to the user. This makes it possible to detect abnormalities in the production line early and quickly provide appropriate countermeasures in a tone that corresponds to the emotion.

[1141] A "data collection system" is a system that automatically collects necessary data from various sensor devices and other information systems within a factory.

[1142] "Filtering" is the process of removing unnecessary information and noise from collected data, making it suitable for analysis and processing.

[1143] "Cleaning" is the process of further refining the filtered data, correcting inconsistencies and inappropriate formats, and compiling it into a unified format.

[1144] A "generative AI model" is an AI model that can perform natural language processing and other tasks based on large amounts of data, specifically referring to generative artificial intelligence models such as GPT.

[1145] A "knowledge base" is a database that stores the knowledge and information of trained generative AI models and makes them easily accessible.

[1146] "Emotion data" is information about emotions inferred from the user's input and actions, and is data that indicates an emotional state such as happiness, dissatisfaction, or impatience.

[1147] "Tone" refers to the style of writing and expression of the generated response, which is appropriately adjusted depending on the user's emotional state.

[1148] The system embodying this invention monitors and controls production lines in a factory, and detects and responds to malfunctions and abnormalities at an early stage. The detailed configuration and operation of this system will be described below.

[1149] System configuration

[1150] Hardware

[1151] Sensors: Attached to various equipment in the factory, they collect data in real time.

[1152] Server: A central location responsible for collecting, filtering, and cleaning data, training generative AI models, and managing the knowledge base.

[1153] User device: A tablet or PC used by workers to enter questions and receive answers.

[1154] software

[1155] CRM, ERP system: Manages data such as production plans and inventory status.

[1156] Natural Language Processing library (NLP): Parse user questions.

[1157] Generative AI model (GPT-3.5 or GPT-4): Generates appropriate answers to user questions.

[1158] Emotion Recognition Engine (Affectiva): Recognizes the user's emotional data and adjusts the tone of the response based on the analysis results.

[1159] Data collection

[1160] The server automatically collects necessary data from sensors in the factory and ERP systems, such as temperature sensors, vibration sensors, and operating hours, in real time.

[1161] Data Filtering and Cleaning

[1162] The collected data is filtered by the server to remove unnecessary information and noise, and then data cleaning is performed to correct inconsistencies and incorrect formats and standardize the data into a consistent format.

[1163] Training generative AI models

[1164] The filtered and cleaned data is used to train a generative AI model (such as GPT-3.5 or GPT-4) that understands the production line's operational information and management procedures to provide optimal answers.

[1165] Knowledge Base Updates

[1166] The knowledge of the learned generative AI model is stored in a knowledge base by the server, which is always updated to the latest version and can be referenced by users as needed.

[1167] Emotional Data Recognition

[1168] When a user uses a terminal to input a question into the system, the emotion recognition engine collects and analyzes emotion data from the user's facial expressions and voice, which is then sent to the server along with the question.

[1169] Generate and provide answers

[1170] The server analyzes the user's question using natural language processing technology and retrieves relevant information from a knowledge base. The generative AI model takes into account the user's emotional data to generate the optimal answer. The generated answer is then adjusted in tone according to the emotional data and provided to the user's device.

[1171] Specific examples

[1172] When a worker at a factory asks, "There is an error with XX device. Please tell me how to fix it," the server processes the following:

[1173] 1. Collect real-time anomaly data from sensors.

[1174] 2. Filter and clean the data.

[1175] 3. Use a trained AI model to identify the cause of the anomaly and how to address it.

[1176] 4. Detect user impatience with an emotion recognition engine.

[1177] 5. Start by saying "Please stay calm" in a gentle tone, then provide detailed instructions on how to proceed.

[1178] Example prompt sentence:

[1179] "Please tell me the maintenance procedures for the production line."

[1180] "XX device has an error. Please tell me how to deal with it."

[1181] "Please tell me the production schedule for next week."

[1182] In this way, the present invention makes it possible to detect abnormalities on the production line early and quickly provide appropriate countermeasures in a tone that reflects the worker's emotions, thereby improving the operating efficiency of the factory and reducing the stress of workers.

[1183] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1184] Step 1:

[1185] The server collects the necessary data from various sensors in the factory and from the ERP system. Specifically, it collects real-time information from temperature sensors, vibration sensors, operating time data, etc. This data collection makes it possible to grasp the latest production status. The input is real-time data from each sensor, and the output is raw data stored on the server.

[1186] Step 2:

[1187] The server filters the collected data to remove unnecessary information and noise, for example, removing sensor error data and redundant data. This process generates a data set with fewer errors. The input is the collected raw data, and the output is the filtered, clean data.

[1188] Step 3:

[1189] The server then cleans the filtered data, correcting any inconsistencies or incorrect formats and standardizing the data (for example, changing all date and time data to the same format). The input is filtered data, and the output is clean, uniform data.

[1190] Step 4:

[1191] The server uses the cleaned data to train a generative AI model. In this process, a generative AI model (such as GPT-3.5 or GPT-4) is retrained using a large amount of data to understand the operational information and management procedures of the production line. The input is clean, unified data, and the output is an updated generative AI model.

[1192] Step 5:

[1193] Using the learned knowledge of the generative AI model, the server updates the knowledge base. This knowledge base stores new knowledge acquired by the generative AI model and allows it to be referenced as needed. The input is the updated knowledge of the generative AI model, and the output is the latest knowledge base.

[1194] Step 6:

[1195] A user uses a terminal to input a question into the system. For example, "An error has occurred with XX device. Please tell me how to deal with it." The input is the user's question, and the output is the question data sent from the terminal to the server.

[1196] Step 7:

[1197] The server uses an emotion recognition engine to recognize the user's emotional data and analyzes it along with the question. This allows the server to understand the user's emotional state (e.g., impatience or dissatisfaction). The input is the user's question and emotional data, and the output is the question data with the emotional data added.

[1198] Step 8:

[1199] The server uses natural language processing technology to analyze the received questions. It extracts the keywords and intent of the questions and searches for related information from a knowledge base. For example, it extracts keywords such as "XX device," "abnormality," and "how to deal with it." The input is question data with emotional data added, and the output is the analyzed question data.

[1200] Step 9:

[1201] The server takes into account the user's emotional data when generating the optimal answer to a question using a generative AI model. For example, if the user is impatient, the tone of the answer can be adjusted to be calmer. The input is the analyzed question data and emotional data, and the output is an answer with the tone adjusted according to the emotion.

[1202] Step 10:

[1203] The server sends the generated answer to the terminal, which then displays it to the user. Specifically, instructions such as "An abnormality has been detected in XX equipment. Please follow the steps below: 1. Check the abnormality. 2. Pause the system. 3. Perform maintenance on the affected area. 4. Perform a restart test" are provided in a friendly tone. The input is an answer with a tone adjusted according to the emotion, and the output is specific instructions displayed on the user's terminal.

[1204] In this way, a series of processing steps can be used to quickly and appropriately respond to abnormalities in the factory production line. In addition, the use of emotion data can reduce worker stress.

[1205] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1206] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1207] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1208] [Fourth embodiment]

[1209] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1210] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1211] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1212] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1213] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1214] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1215] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1216] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1217] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1218] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1219] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1220] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1221] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1222] This invention is an "Internal Knowledge GPT" system that visualizes internal business processes and operational flows and improves the reliability of the latest information. In this system, servers, terminals, and users cooperate to collect, organize, learn, and answer questions about information.

[1223] Data collection

[1224] The server collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.) via APIs and data files. For example, the server retrieves sales transaction data from a customer relationship management system (CRM) and organizes the information.

[1225] Data Filtering and Cleaning

[1226] The server filters and cleans the collected data, which includes removing duplicate data and unnecessary information, and normalizing the data format. For example, the server standardizes date formats and corrects typos in text data.

[1227] Generative AI model training

[1228] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process enables the model to understand the latest business information and operational flows, and generate optimal answers based on that information.

[1229] Knowledge Base Updates

[1230] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[1231] Enter a question

[1232] A user inputs a question into the system using a terminal. For example, the user inputs, "Please tell me about the application flow for a new project."

[1233] question analysis

[1234] The server receives a question from a user, analyzes the question using natural language processing technology, and retrieves relevant information from a knowledge base based on the analysis result.

[1235] Answer generation

[1236] The server uses a generative AI model to generate the best answer to the question, based on the most up-to-date information in the knowledge base.

[1237] Provide answers

[1238] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, it explains the detailed steps in the form of "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[1239] Specific examples

[1240] For example, if a user asks, "What is the latest format for sales reports?"

[1241] 1. The server first collects the latest data on sales reports from the CRM system.

[1242] 2. This data is then filtered and cleaned to remove duplicates and extract the information you need.

[1243] 3. The server uses this cleaned data to train a generative AI model.

[1244] 4. When a user enters a question, the server analyzes the question using natural language processing technology and searches the knowledge base for relevant information.

[1245] 5. The generative AI model generates the optimal answer, and the server sends the answer to the device.

[1246] 6. The terminal displays to the user, "The latest sales report format is as follows," and indicates the specific items.

[1247] In this way, the in-house knowledge GPT system provides quick and accurate answers to user questions, improving work efficiency and the reliability of information.

[1248] The processing flow will be explained below.

[1249] Step 1:

[1250] The server automatically collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.). The server uses APIs to obtain data in real time and stores the collected data in temporary storage. For example, the server obtains transaction data and customer information from a CRM system.

[1251] Step 2:

[1252] The server filters the collected data. In this step, the server removes data duplication and noise and extracts only the necessary information. For example, it removes duplicate transaction information from past data logs.

[1253] Step 3:

[1254] The server cleans the filtered data, standardizing the data format and correcting typos and errors (e.g., converting date data in different formats into a standard format).

[1255] Step 4:

[1256] The server feeds the cleaned data to a generative AI model (GPT) to train the model, which then understands the latest business information and operational flows.

[1257] Step 5:

[1258] The server stores the learned knowledge of the generative AI model in a knowledge base, which is periodically updated to reflect the latest information.

[1259] Step 6:

[1260] A user uses a terminal to input a question into the system. For example, the user inputs, "Please tell me about the application flow for a new project."

[1261] Step 7:

[1262] The terminal sends the user's question to the server in text format.

[1263] Step 8:

[1264] The server receives the question, analyzes it using natural language processing technology, extracts keywords and the content of the request, and searches for relevant information from a knowledge base.

[1265] Step 9:

[1266] The server uses the generative AI model to generate the best answer to the question, and creates an answer that includes specific steps and information based on the analysis results.

[1267] Step 10:

[1268] The server sends the generated answer to the terminal, which then communicates the answer to the user in text format.

[1269] Step 11:

[1270] The device will display a response to the user, providing detailed instructions such as, "The flow for applying for a new project is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[1271] Example 1

[1272] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1273] In many companies today, information about business processes and operational flows is scattered across various departments and systems, making it difficult to quickly and accurately collect and organize that information and utilize it in a timely manner. This results in the time and effort required to obtain the necessary information, which reduces the efficiency of internal operations. Furthermore, information may not be up-to-date or reliable, which can hinder decision-making and business execution.

[1274] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1275] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for storing knowledge learned by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions using natural language processing technology, means for acquiring information from the knowledge base based on the analyzed questions and generating answers, and means for providing the generated answers to users. This makes it possible to centrally manage data from each business system and train a generative AI model based on the latest filtered and cleaned data, thereby providing quick and accurate answers to user questions.

[1276] A "data collection system" is a system for collecting data from various business systems (CRM, ERP, mail servers, chat systems, etc.).

[1277] "Filtering" is the process of removing duplicate and unnecessary information from collected data.

[1278] "Cleaning" is the process of standardizing the format of collected data and correcting typos and formatting.

[1279] A "generative AI model" is an artificial intelligence model that learns from filtered and cleaned data and is used to generate answers to user questions.

[1280] A "knowledge base" is a database in which the knowledge of a trained generative AI model is stored and referenced as needed.

[1281] "Natural language processing technology" is a technology for analyzing questions from users and understanding context and keywords.

[1282] "Analysis" refers to the process of analyzing a user's question using natural language processing technology to understand the intent of the question and related information.

[1283] "Answer generation" is the process of creating optimal answers using a generative AI model based on the analyzed question.

[1284] "Providing" refers to the act of displaying or transmitting the generated answer to the user.

[1285] This invention is an "Internal Knowledge GPT" system that visualizes internal business processes and operational flows and improves the reliability of the latest information. In this system, servers, terminals, and users cooperate to collect, organize, learn, and answer questions about information.

[1286] Data collection

[1287] The server collects the necessary data from various business systems, such as CRM, ERP, mail servers, and chat systems, via APIs and data files. Specifically, it uses REST API clients and database connectors. For example, the server retrieves sales transaction data from a customer relationship management system (CRM) and organizes the information.

[1288] Data Filtering and Cleaning

[1289] The server filters and cleans the collected data. This includes processing the data using Python's pandas and SQL queries. It also removes duplicate data and standardizes data formats. For example, it removes duplicate data for the same customer and standardizes date formats.

[1290] Generative AI model training

[1291] The server uses the filtered and cleaned data to train a generative AI model (GPT). This process uses machine learning frameworks such as PyTorch and TensorFlow. By training the model with the latest business information and operational flows, the server is able to generate accurate answers that are adapted to user questions.

[1292] Knowledge Base Updates

[1293] The knowledge of the generative AI model trained by the server is stored in a knowledge base. The knowledge base is constructed as an SQL database or a NoSQL database (e.g., MongoDB). The stored knowledge can be referenced by the server as needed.

[1294] Enter a question

[1295] Users use a terminal to input questions into the system, either through a browser-based interface or a dedicated application. For example, a question might be, "Please tell me about the application process for a new project."

[1296] question analysis

[1297] The server receives questions from users and analyzes them using natural language processing technology. Specifically, it uses natural language processing libraries such as spaCy and NLTK. The server extracts keywords from the questions and understands their context, then retrieves the necessary information from a knowledge base.

[1298] Answer generation

[1299] The server uses a generative AI model to generate the best answer to the question, and references the latest business information based on a knowledge base to provide an accurate answer to the user's question.

[1300] Provide answers

[1301] The server sends the generated answer to the terminal, which then displays it to the user, such as "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[1302] Examples and prompts

[1303] For example, if a user asks, "What is the latest format for sales reports?"

[1304] 1. The server collects the latest data on sales reports from the CRM system.

[1305] 2. Filter and clean this data to remove duplicates and extract the information you need.

[1306] 3. The server uses this cleaned data to train a generative AI model.

[1307] 4. When a user enters a question into the terminal, the server analyzes the question using natural language processing technology and searches for relevant information from a knowledge base.

[1308] 5. The generative AI model generates the optimal answer, and the server sends the answer to the device.

[1309] 6. The terminal displays to the user, "The latest sales report format is as follows," and indicates the specific items.

[1310] In this way, the in-house knowledge GPT system provides quick and accurate answers to user questions, improving work efficiency and the reliability of information.

[1311] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1312] Step 1:

[1313] The server collects the necessary data from the data collection system. Inputs are APIs and data files from each business system (CRM, ERP, mail server, chat system, etc.). Based on these inputs, the server collects data using a REST API client or database connector. The output is saved as raw data.

[1314] Step 2:

[1315] The server filters and cleans the collected data. The input is the collected raw data. The server processes the data using Python pandas and SQL queries to remove duplicate data and standardize data formats. Specifically, it removes duplicate data for the same customer and standardizes date formats. The output is the filtered and cleaned data.

[1316] Step 3:

[1317] The server trains a generative AI model using the filtered and cleaned data. The input is the cleaned data, and the server trains a generative AI model (GPT) using a machine learning framework such as PyTorch or TensorFlow. The output is a trained generative AI model.

[1318] Step 4:

[1319] The server stores the knowledge of the trained generative AI model in a knowledge base. The input is the trained generative AI model, which the server stores in an SQL database or a NoSQL database (e.g., MongoDB). The output is the data stored as a knowledge base.

[1320] Step 5:

[1321] A user uses a terminal to input a question into the system. The input is the user's question, and the terminal uses a browser-based interface or a dedicated application. A specific example is "Please tell me about the application flow for a new project." The output is the question.

[1322] Step 6:

[1323] The server analyzes questions from users using natural language processing technology. The input is the user's question, and the server analyzes the question using natural language processing libraries such as spaCy or NLTK. The analysis extracts keywords from the question and understands the context. The output is the analysis results.

[1324] Step 7:

[1325] The server retrieves information from the knowledge base based on the analyzed question and generates the optimal answer using a generative AI model. The input is the analysis result and information from the knowledge base. The server creates the optimal answer based on the data it has learned in advance. The output is the generated answer.

[1326] Step 8:

[1327] The server sends the generated answer to the terminal, which then displays it to the user. The input is the generated answer, and the terminal displays the answer using a browser-based interface or a dedicated application. The output is the display of the answer to the user. The specific operation is displayed as detailed steps: "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[1328] (Application example 1)

[1329] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1330] In conventional factory management systems, information related to manufacturing processes and maintenance procedures is managed in a decentralized manner, making it difficult for workers and managers to quickly obtain the information they need.It is also difficult to provide accurate information based on the latest business and operational flows, creating challenges in improving business efficiency and reliability.

[1331] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1332] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for saving the data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions, means for generating answers based on the analyzed questions, means for collecting manufacturing process data and maintenance history from a factory management system and generating answers to the questions based thereon, and means for providing the generated answers to users. This enables quick and accurate provision of information on work flows, equipment operating procedures, and maintenance procedures within the factory.

[1333] A "data collection system" is an infrastructure for acquiring data from various business systems, sensors, etc.

[1334] "Filtering" is the process of removing unnecessary information and duplicate data from collected data.

[1335] "Cleaning" is the process of improving data quality by correcting errors and standardizing the format of the data.

[1336] A "generative AI model" is an artificial intelligence model that learns from collected and organized data and generates appropriate answers to user questions.

[1337] A "knowledge base" is a database where information learned by a generative AI model is stored and can be referenced at any time.

[1338] "Accepting a question" is the process by which the system obtains a query from a user as input.

[1339] "Question analysis" is the process of using natural language processing techniques to understand the intent of a user's question and extract relevant information.

[1340] "Generating an answer" is the process of using a generative AI model to create the optimal answer to a user's question.

[1341] A "factory management system" is a system for managing operational information on manufacturing processes and equipment.

[1342] "Manufacturing process data" refers to a series of information related to the manufacturing process of a product, including process progress and quality control data.

[1343] "Maintenance history" refers to records relating to the maintenance and repair of equipment and facilities, including procedures for dealing with problems when they occur and repair history.

[1344] This invention relates to the "Factory Knowledge GPT" system, which enables workers and managers engaged in factory management and maintenance work to quickly and accurately obtain the information they need. This system involves the collaboration of servers, terminals, and users to collect, organize, learn, and answer questions about information.

[1345] Data collection

[1346] The server collects information such as manufacturing process data and equipment maintenance history from the factory management system via APIs and data files. For example, the server obtains daily production line progress data and past maintenance records for each piece of equipment from the factory management system and organizes the information.

[1347] Data Filtering and Cleaning

[1348] The server then filters and cleans the collected data, removing unnecessary information and duplicate data, standardizing data formats, etc. For example, the server may standardize date formats and correct typos in text data to improve data quality.

[1349] Generative AI model training

[1350] The server uses the filtered and cleaned data to train a generative AI model (GPT). This model understands the latest business information and operational flows, and can generate optimal answers to user questions. For example, it learns troubleshooting procedures and equipment maintenance procedures on a production line.

[1351] Knowledge Base Updates

[1352] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[1353] Question input and analysis

[1354] A user inputs a question into the system using a terminal, such as a smartphone or robot interface, asking, "What are the troubleshooting steps for assembly line B?" The server then analyzes the question using natural language processing technology and retrieves relevant information from the knowledge base.

[1355] Answer generation and provision

[1356] The server uses a generative AI model to generate the best answer to the question. The generated answer is based on the latest information in the knowledge base. For example, it provides detailed instructions in the form of "Troubleshooting steps for assembly line B are as follows: 1. Check the power supply of the equipment. 2. Check the location of the sensors. 3. Stop and reset the production line. 4. Check each error log. 5. If you are unsure, contact the administrator."

[1357] Hardware and software used

[1358] The system uses the following hardware and software:

[1359] Hardware: Servers (cloud services recommended, e.g., AWS EC2), smartphones (e.g., iPhone, Android device), factory robots (e.g., robotic arms from ABB or FANUC)

[1360] Software: GPT model (Hugging Face Transformers library), API requests (requests library), data management (JSON format)

[1361] Specific examples

[1362] For example, a factory worker types the following question:

[1363] Question: "What is the maintenance procedure for device A?"

[1364] In response to this question, the server collects and filters the necessary data and uses a generative AI model to generate an answer like this:

[1365] "The maintenance procedure for Equipment A is as follows: 1. Power off the equipment. 2. Put on protective equipment. 3. Inspect each part. 4. Apply lubricant. 5. Report any problems immediately."

[1366] This allows users to quickly obtain accurate and reliable information.

[1367] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1368] Step 1: Data collection

[1369] The server collects manufacturing process data and equipment maintenance history from the factory management system. Specifically, it obtains data through APIs and data files. The input to this step is data from the factory management system, and the output is the collected raw data. The server sends a request to the API endpoint and receives the data in JSON format.

[1370] Step 2: Data filtering

[1371] The server filters the collected data, specifically removing unnecessary information and duplicate data based on manually set filter conditions. The input to this step is the raw data collected in step 1, and the output is the filtered data. The server processes the data by removing unnecessary fields and removing duplicate data.

[1372] Step 3: Data cleaning

[1373] The server cleans the filtered data. Specifically, it standardizes the data format and corrects typos. The input to this step is the data filtered in step 2, and the output is the cleaned data. Standardizing the data format includes standardizing the date format and text encoding. Typos are automatically corrected using a preset dictionary.

[1374] Step 4: Generative AI model training

[1375] The server uses the cleaned data to train a generative AI model (GPT). Specifically, it inputs the cleaned data into the model and updates the generative AI model. The input of this step is the cleaned data, and the output is a trained generative AI model. The server tokenizes the data and trains it to optimize the model parameters.

[1376] Step 5: Update your knowledge base

[1377] The server saves the knowledge of the trained generative AI model in a knowledge base. Specifically, it stores the information learned by the model in a database. The input of this step is the trained generative AI model, and the output is an updated knowledge base. The server inserts the information into the database for the knowledge base and creates the necessary indexes.

[1378] Step 6: Enter your question

[1379] The user uses a terminal to input a question into the system. Specifically, the question is uttered via text or voice through a smartphone or robot interface. The input in this step is the user's question, and the output is digital data of the question. The terminal receives the user's input, converts it into text format, and sends it to the server.

[1380] Step 7: Question Analysis

[1381] The server analyzes the user's question. Specifically, it uses natural language processing technology to understand the intent of the question and searches for relevant information in a knowledge base. The input to this step is the digital data of the user's question, and the output is the analysis result. The server tokenizes the question and analyzes it using natural language processing (NLP) algorithms to identify relevant information.

[1382] Step 8: Answer Generation

[1383] The server uses a generative AI model to generate the optimal answer to the question. Specifically, the model generates an answer based on the analysis results. The input to this step is the analysis results, and the output is the generated answer. The generative AI model generates answer text based on the tokenized analysis results and converts it into a text format.

[1384] Step 9: Provide your answers

[1385] The server sends the generated answer to the terminal, which then displays the answer to the user. Specifically, the answer is displayed on the interface of a smartphone or robot. The input of this step is the generated answer, and the output is the answer displayed to the user. The terminal displays the text data received from the server on the user interface.

[1386] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1387] This invention is an "Internal Knowledge GPT" system that combines an emotion engine to visualize internal business processes and operational flows and improve the reliability of the latest information. In this system, the server, terminals, and users cooperate to collect, organize, learn, answer questions, and recognize emotions.

[1388] Data collection

[1389] The server collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.) via APIs and data files. For example, the server retrieves sales transaction data and customer information from a CRM system and organizes the information.

[1390] Data Filtering and Cleaning

[1391] The server filters the collected data, removing unnecessary information, and then performs cleaning and formatting on the data. This includes removing duplicate data and standardizing formats. For example, the server standardizes date formats and corrects typos and errors.

[1392] Generative AI model training

[1393] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process enables the AI ​​model to understand the latest business information and operational flow, and generate optimal answers based on that information.

[1394] Knowledge Base Updates

[1395] The server stores the learned knowledge of the generative AI model in a knowledge base, which acts as a database that can be referenced whenever necessary.

[1396] Enter a question

[1397] The user inputs a question into the system using a terminal. For example, the user might input, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[1398] question analysis

[1399] The server receives questions from users and analyzes them using natural language processing technology along with emotional data detected by the emotion engine, extracting keywords and requests from the questions and retrieving relevant information from the knowledge base.

[1400] Answer generation

[1401] The server takes into account the user's emotional data when generating the optimal answer to the question using a generative AI model. Based on the analysis results, it creates an answer that includes specific steps and information, and adjusts the tone of the answer. For example, if the user is feeling frustrated, it will soften the tone of the answer.

[1402] Provide answers

[1403] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, it shows detailed steps such as, "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department." The answer is also provided in a tone that corresponds to the user's emotions.

[1404] Specific examples

[1405] For example, if a user asks, "Tell me about your recent sales report?"

[1406] 1. The server first collects the latest data on sales reports from the CRM system.

[1407] 2. This data is then filtered and cleaned to remove duplicates and extract the information you need.

[1408] 3. The server uses this cleaned data to train a generative AI model.

[1409] 4. When the user enters a question, the emotion engine recognizes the user's emotion at the time of entering the question and sends the emotion data to the server.

[1410] 5. The server receives the question, analyzes keywords such as "sales report" and "recent" using natural language processing technology, and searches for related information in the knowledge base.

[1411] 6. The server uses a generative AI model to generate the optimal answer, taking into account the user's emotional data. For example, if the user is dissatisfied, the answer will be delivered in a more friendly tone.

[1412] 7. The server sends the generated answer to the terminal, which displays it to the user.

[1413] In this way, the in-house knowledge GPT system combined with the emotion engine provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[1414] The processing flow will be explained below.

[1415] Step 1:

[1416] The server automatically collects data from each business system (e.g., CRM, ERP, mail server, chat system, etc.). The server uses APIs to obtain data in real time and stores the collected data in temporary storage. For example, the server obtains transaction data and customer information from a CRM system.

[1417] Step 2:

[1418] The server filters the collected data. In this step, the server removes data duplication and noise and extracts only the necessary information. For example, it removes duplicate transaction information from past data logs.

[1419] Step 3:

[1420] The server cleans the filtered data, standardizing the data format and correcting typos and errors (e.g., converting date data in different formats into a standard format).

[1421] Step 4:

[1422] The server feeds the cleaned data to a generative AI model (GPT) to train the model, which then understands the latest business information and operational flows.

[1423] Step 5:

[1424] The server stores the learned knowledge of the generative AI model in a knowledge base, which is periodically updated to reflect the latest information.

[1425] Step 6:

[1426] A user inputs a question into the system using a terminal. For example, the user might input, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[1427] Step 7:

[1428] The device sends the user's question and emotional data to the server. The question is sent in text format, and the emotional data is sent as a numerical value or category.

[1429] Step 8:

[1430] The server receives the question and analyzes it using natural language processing technology, extracting keywords and the content of the request, and collating emotional data to search for relevant information from a knowledge base.

[1431] Step 9:

[1432] The server uses a generative AI model to generate the optimal answer to the question, taking into account the emotional data and creating an answer in a tone that corresponds to the user's emotions. For example, if the data indicates anger, the tone of the answer will be calmer.

[1433] Step 10:

[1434] The server sends the generated answer to the terminal, which then communicates the answer to the user in text format.

[1435] Step 11:

[1436] The device will then display answers and sentiment analysis information to the user, providing detailed instructions in a tone that reflects their emotions, such as "The new project application process is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[1437] Specific examples

[1438] Example 1: When a user asks, "Tell me about your recent sales report."

[1439] Step 1:

[1440] The server collects the latest sales report data from the CRM system.

[1441] Step 2:

[1442] The server filters the collected data and removes duplicate data.

[1443] Step 3:

[1444] The server standardizes the data format and corrects typos and errors.

[1445] Step 4:

[1446] The server feeds the cleaned data into a generative AI model (GPT) to train the model.

[1447] Step 5:

[1448] The server stores the knowledge of the generated AI model in a knowledge base.

[1449] Step 6:

[1450] The user inputs a question into the terminal, such as "Please tell me about the recent sales report." At this time, the emotion engine recognizes the user's emotion.

[1451] Step 7:

[1452] The device sends the question and emotion data to the server.

[1453] Step 8:

[1454] The server analyzes the question and searches a knowledge base for relevant information.

[1455] Step 9:

[1456] The server uses generative AI models to generate optimal answers, crafting them in a tone that reflects the user's emotions.

[1457] Step 10:

[1458] The server sends the generated response to the terminal.

[1459] Step 11:

[1460] The device will then display the answer and sentiment analysis information to the user, such as "Regarding the recent sales report, please pay attention to the following points: 1. Overview of sales activities 2. Successes and challenges 3. Next action plan," in a tone that reflects the user's sentiment.

[1461] In this way, the in-house knowledge GPT system combined with the emotion engine provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[1462] Example 2

[1463] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1464] Conventional data collection and question answering systems generate answers without considering the user's feelings, which leads to low user satisfaction. Furthermore, it is difficult to provide accurate answers when inaccurate or duplicate data exists.

[1465] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1466] In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for storing the data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing emotions, means for analyzing the accepted questions using natural language processing technology, means for generating answers based on the analyzed questions and emotion data, and means for providing the generated answers to the users. This makes it possible to quickly provide accurate and appropriate answers that take the user's emotions into consideration.

[1467] A "data collection system" is a system for efficiently collecting necessary data from various sources.

[1468] "Filtering" is the process of removing unnecessary or redundant information from collected data.

[1469] "Cleaning" is a process that involves correcting data and standardizing its format in order to ensure its accuracy.

[1470] A "generative AI model" is an artificial intelligence model that can learn large amounts of data and generate sentences in natural language like a human.

[1471] A "knowledge base" is a database that systematically stores collected and learned information and can be referenced and used as needed.

[1472] "User" refers to a person or entity that uses the system to search for information or ask a question.

[1473] The "means for accepting questions" is an interface for inputting questions from users into the system.

[1474] "Natural language processing technology" is a technology that enables computers to understand and process human language.

[1475] An "emotion engine" is a system that detects emotions from user input and reactions and analyzes the data.

[1476] "Answer generation means" refers to a method or technology for creating an optimal answer based on the user's question and emotion data.

[1477] "Means for adjusting tone" refers to methods or techniques for appropriately adjusting the expression or tone of the generated response to match the user's emotions.

[1478] This invention is an "Internal Knowledge GPT" system that combines an emotion engine to visualize internal business processes and operational flows and improve the reliability of the latest information. In this system, the server, terminals, and users cooperate to collect, organize, learn, answer questions, and recognize emotions.

[1479] Data collection

[1480] The server collects data from each business system (e.g., customer relationship management system (CRM), enterprise resource planning system (ERP), mail server, chat system, etc.) via APIs and data files. For example, the server obtains sales transaction data and customer information from the customer relationship management system and stores it in storage.

[1481] Data Filtering and Cleaning

[1482] The server filters the collected data, removing unnecessary and duplicate information, and then performs cleaning and formatting. This includes standardizing date formats and correcting typos and errors. For example, it standardizes date data stored in different formats to the "YYYY-MM-DD" format.

[1483] Generative AI model training

[1484] The server uses the filtered and cleaned data to train a generative AI model (GPT). This training process uses machine learning libraries such as TensorFlow and PyTorch, allowing the AI ​​model to understand the latest business information and operational flows and generate optimal answers.

[1485] Knowledge Base Updates

[1486] The server stores the learned knowledge of the generative AI model in a knowledge base, which functions as a database that can be referenced whenever necessary, allowing users to access the latest business information.

[1487] Enter a question

[1488] The user uses a terminal to input a question into the system. For example, "Please tell me about the application process for a new project." At this time, the emotion engine recognizes the user's emotion and simultaneously records that emotion data.

[1489] question analysis

[1490] The server receives questions from users and analyzes them using natural language processing technology along with emotional data detected by the emotion engine. Keywords and requests from the questions are extracted and analyzed. For example, keywords such as "new project" and "application flow" are extracted.

[1491] Answer generation

[1492] The server uses a generative AI model to generate the best answer to the question, taking into account the user's emotional data. It provides specific instructions and information in a toned manner. For example, if the user is frustrated, it softens the tone of the answer.

[1493] Provide answers

[1494] The server sends the generated answer to the terminal, and the terminal displays the answer to the user. For example, detailed steps are shown in the form of "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your superior. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[1495] Specific examples

[1496] For example, if a user asks, "Tell me about your recent sales report?":

[1497] 1. The server first collects the latest data on sales reports from the customer relationship management system.

[1498] 2. The server filters and cleans the data, removing duplicates and extracting the information you need.

[1499] 3. The server trains the generative AI model using the cleaned data, for example, incorporating the latest sales results and customer feedback.

[1500] 4. When the user enters a question, the device recognizes the emotion data using the emotion engine and sends it to the server. For example, if the user is feeling impatient, that emotion will be recorded.

[1501] 5. The server receives the question, analyzes keywords such as "sales report" and "recent" using natural language processing technology, and searches for related information in the knowledge base.

[1502] 6. The server uses a generative AI model to generate the optimal answer, taking into account the user's emotional data (impatience), and provides a friendly response such as, "Our recent sales report shows that after the mid-October version upgrade, feedback from major customers has been positive, and sales have increased by 20%."

[1503] 7. The server sends the generated answer to the terminal, which displays it to the user.

[1504] In this way, the "Internal Knowledge GPT" system, which combines an emotion engine, provides quick and accurate answers while taking into account the user's emotions, thereby improving work efficiency and the reliability of information.

[1505] Prompt Sentence Examples

[1506] To ask "Tell me about recent sales reports," a user would type the following prompt into the terminal:

[1507] Please tell me about your recent sales report.

[1508] The system analyzes this prompt, collects, organizes, and analyzes the necessary information, and provides a satisfactory answer to the user.

[1509] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1510] Step 1:

[1511] The server collects data from each business system (e.g. CRM, ERP, mail server, chat system, etc.). The input is the API or data file from each business system, and the output is data saved in temporary storage. Specific operations include sending requests to the API endpoints of each system and receiving data as a response.

[1512] Step 2:

[1513] The server filters the collected data to remove unnecessary and duplicated information. The input is the raw data collected in step 1, and the output is the filtered data. Specifically, it identifies duplicate records and keeps only the most recent data. It also includes the removal of unnecessary temporary notes and duplicate information.

[1514] Step 3:

[1515] The server cleans and formats the filtered data. The input is the data filtered in step 2, and the output is the cleaned data. Specific operations include standardizing date formats, correcting typos, and imputing missing values. For example, standardizing date data in different formats to the "YYYY-MM-DD" format.

[1516] Step 4:

[1517] The server trains a generative AI model (GPT) using the cleaned data. The input is the cleaned data from step 3, and the output is a trained generative AI model. Specifically, the data is tokenized and the model is trained using a machine learning library such as TensorFlow or PyTorch. The model accuracy is improved through 100 epochs of training.

[1518] Step 5:

[1519] The server stores the learned knowledge of the generative AI model in a knowledge base. The input is the generative AI model trained in step 4, and the output is the updated knowledge base. Specifically, the server saves the model parameters in a database and updates the knowledge base accordingly.

[1520] Step 6:

[1521] The user uses a terminal to input a question into the system. The input is the user's question, and the output is the inquiry displayed on the terminal. For example, the user might input, "Please tell me about the application flow for a new project." At this time, the emotion engine recognizes the user's emotion and records the emotion data.

[1522] Step 7:

[1523] The device passes the input question to the emotion engine and records the emotion data. The input is the question entered by the user, and the output is the recognized emotion data. Specifically, if the user is feeling impatient, that emotion is recorded as "impatience."

[1524] Step 8:

[1525] The server receives a question from the user and analyzes it using natural language processing technology along with the emotional data detected by the emotion engine. The input is the question received in step 6 and the emotional data obtained in step 7, and the output is keywords and request details as the analysis results. For example, the keywords "new project" and "application flow" are extracted.

[1526] Step 9:

[1527] The server uses the generative AI model to generate the optimal answer to the question, taking the user's emotional data into consideration. The input is the analysis results and emotional data obtained in step 8, and the output is the generated answer. Specific operations include generating an answer that includes specific steps and information based on the analysis results. For example, if the user is feeling impatient, the tone of the answer will be softened.

[1528] Step 10:

[1529] The server sends the generated answer to the terminal. The input is the answer generated in step 9, and the output is the answer sent to the terminal. Specific operations include the server sending the answer to the terminal and the user confirming it.

[1530] Step 11:

[1531] The terminal displays the received response to the user. The input is the response sent in step 10, and the output is the response displayed on the user's screen. For example, detailed steps might be displayed, such as, "The new project application flow is as follows: 1. Create a project proposal. 2. Review the proposal with the project manager. 3. Obtain approval from your supervisor. 4. Enter the application into the project management system. 5. Wait for final approval from the management department."

[1532] (Application example 2)

[1533] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1534] In conventional factory production lines, it was difficult to detect malfunctions or abnormalities early on, and it was also difficult to immediately provide workers with appropriate countermeasures. Furthermore, there was a lack of means to reduce the stress workers felt when malfunctions occurred. This prevented efficient production management and maintenance, leading to problems that led to production line shutdowns and a decline in quality.

[1535] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting data from a data collection system, means for filtering and cleaning the collected data, means for training a generative AI model using the filtered and cleaned data, means for saving data trained by the generative AI model in a knowledge base, means for accepting questions from users, means for analyzing the accepted questions, means for generating answers based on the analyzed questions, means for recognizing user emotion data, means for adjusting the tone of the answer taking the recognized emotion data into consideration, and means for providing the generated answer to the user. This makes it possible to detect abnormalities in the production line early and quickly provide appropriate countermeasures in a tone that corresponds to the emotion.

[1536] A "data collection system" is a system that automatically collects necessary data from various sensor devices and other information systems within a factory.

[1537] "Filtering" is the process of removing unnecessary information and noise from collected data, making it suitable for analysis and processing.

[1538] "Cleaning" is the process of further refining the filtered data, correcting inconsistencies and inappropriate formats, and compiling it into a unified format.

[1539] A "generative AI model" is an AI model that can perform natural language processing and other tasks based on large amounts of data, specifically referring to generative artificial intelligence models such as GPT.

[1540] A "knowledge base" is a database that stores the knowledge and information of trained generative AI models and makes them easily accessible.

[1541] "Emotion data" is information about emotions inferred from the user's input and actions, and is data that indicates an emotional state such as happiness, dissatisfaction, or impatience.

[1542] "Tone" refers to the style of writing and expression of the generated response, which is appropriately adjusted depending on the user's emotional state.

[1543] The system embodying this invention monitors and controls production lines in a factory, and detects and responds to malfunctions and abnormalities at an early stage. The detailed configuration and operation of this system will be described below.

[1544] System configuration

[1545] Hardware

[1546] Sensors: Attached to various equipment in the factory, they collect data in real time.

[1547] Server: A central location responsible for collecting, filtering, and cleaning data, training generative AI models, and managing the knowledge base.

[1548] User device: A tablet or PC used by workers to enter questions and receive answers.

[1549] software

[1550] CRM, ERP system: Manages data such as production plans and inventory status.

[1551] Natural Language Processing library (NLP): Parse user questions.

[1552] Generative AI model (GPT-3.5 or GPT-4): Generates appropriate answers to user questions.

[1553] Emotion Recognition Engine (Affectiva): Recognizes the user's emotional data and adjusts the tone of the response based on the analysis results.

[1554] Data collection

[1555] The server automatically collects necessary data from sensors in the factory and ERP systems, such as temperature sensors, vibration sensors, and operating hours, in real time.

[1556] Data Filtering and Cleaning

[1557] The collected data is filtered by the server to remove unnecessary information and noise, and then data cleaning is performed to correct inconsistencies and incorrect formats and standardize the data into a consistent format.

[1558] Training generative AI models

[1559] The filtered and cleaned data is used to train a generative AI model (such as GPT-3.5 or GPT-4) that understands the production line's operational information and management procedures to provide optimal answers.

[1560] Knowledge Base Updates

[1561] The knowledge of the learned generative AI model is stored in a knowledge base by the server, which is always updated to the latest version and can be referenced by users as needed.

[1562] Emotional Data Recognition

[1563] When a user uses a terminal to input a question into the system, the emotion recognition engine collects and analyzes emotion data from the user's facial expressions and voice, which is then sent to the server along with the question.

[1564] Generate and provide answers

[1565] The server analyzes the user's question using natural language processing technology and retrieves relevant information from a knowledge base. The generative AI model takes into account the user's emotional data to generate the optimal answer. The generated answer is then adjusted in tone according to the emotional data and provided to the user's device.

[1566] Specific examples

[1567] When a worker at a factory asks, "There is an error with XX device. Please tell me how to fix it," the server processes the following:

[1568] 1. Collect real-time anomaly data from sensors.

[1569] 2. Filter and clean the data.

[1570] 3. Use a trained AI model to identify the cause of the anomaly and how to address it.

[1571] 4. Detect user impatience with an emotion recognition engine.

[1572] 5. Start by saying "Please stay calm" in a gentle tone, then provide detailed instructions on how to proceed.

[1573] Example prompt sentence:

[1574] "Please tell me the maintenance procedures for the production line."

[1575] "XX device has an error. Please tell me how to deal with it."

[1576] "Please tell me the production schedule for next week."

[1577] In this way, the present invention makes it possible to detect abnormalities on the production line early and quickly provide appropriate countermeasures in a tone that reflects the worker's emotions, thereby improving the operating efficiency of the factory and reducing the stress of workers.

[1578] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1579] Step 1:

[1580] The server collects the necessary data from various sensors in the factory and from the ERP system. Specifically, it collects real-time information from temperature sensors, vibration sensors, operating time data, etc. This data collection makes it possible to grasp the latest production status. The input is real-time data from each sensor, and the output is raw data stored on the server.

[1581] Step 2:

[1582] The server filters the collected data to remove unnecessary information and noise, for example, removing sensor error data and redundant data. This process generates a data set with fewer errors. The input is the collected raw data, and the output is the filtered, clean data.

[1583] Step 3:

[1584] The server then cleans the filtered data, correcting any inconsistencies or incorrect formats and standardizing the data (for example, changing all date and time data to the same format). The input is filtered data, and the output is clean, uniform data.

[1585] Step 4:

[1586] The server uses the cleaned data to train a generative AI model. In this process, a generative AI model (such as GPT-3.5 or GPT-4) is retrained using a large amount of data to understand the operational information and management procedures of the production line. The input is clean, unified data, and the output is an updated generative AI model.

[1587] Step 5:

[1588] Using the learned knowledge of the generative AI model, the server updates the knowledge base. This knowledge base stores new knowledge acquired by the generative AI model and allows it to be referenced as needed. The input is the updated knowledge of the generative AI model, and the output is the latest knowledge base.

[1589] Step 6:

[1590] A user uses a terminal to input a question into the system. For example, "An error has occurred with XX device. Please tell me how to deal with it." The input is the user's question, and the output is the question data sent from the terminal to the server.

[1591] Step 7:

[1592] The server uses an emotion recognition engine to recognize the user's emotional data and analyzes it along with the question. This allows the server to understand the user's emotional state (e.g., impatience or dissatisfaction). The input is the user's question and emotional data, and the output is the question data with the emotional data added.

[1593] Step 8:

[1594] The server uses natural language processing technology to analyze the received questions. It extracts the keywords and intent of the questions and searches for related information from a knowledge base. For example, it extracts keywords such as "XX device," "abnormality," and "how to deal with it." The input is question data with emotional data added, and the output is the analyzed question data.

[1595] Step 9:

[1596] The server takes into account the user's emotional data when generating the optimal answer to a question using a generative AI model. For example, if the user is impatient, the tone of the answer can be adjusted to be calmer. The input is the analyzed question data and emotional data, and the output is an answer with the tone adjusted according to the emotion.

[1597] Step 10:

[1598] The server sends the generated answer to the terminal, which then displays it to the user. Specifically, instructions such as "An abnormality has been detected in XX equipment. Please follow the steps below: 1. Check the abnormality. 2. Pause the system. 3. Perform maintenance on the affected area. 4. Perform a restart test" are provided in a friendly tone. The input is an answer with a tone adjusted according to the emotion, and the output is specific instructions displayed on the user's terminal.

[1599] In this way, a series of processing steps can be used to quickly and appropriately respond to abnormalities in the factory production line. In addition, the use of emotion data can reduce worker stress.

[1600] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1601] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1602] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1603] The emotion identification model 59 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 an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1604] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1605] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1606] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1607] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1608] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs 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 a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1609] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1610] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1611] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1612] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1613] 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.

[1614] It is not necessary to store all 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 all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1615] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1616] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with 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). Also, the hardware resource that executes the specific processing may be a single processor.

[1617] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1618] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1619] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1620] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1621] The following is further disclosed regarding the above embodiment.

[1622] (Claim 1)

[1623] a means for collecting data from a data collection system;

[1624] a means of filtering and cleaning the collected data;

[1625] a means for training a generative AI model using the filtered and cleaned data; and

[1626] a means for storing data learned by the generative AI model in a knowledge base;

[1627] means for accepting questions from users;

[1628] a means for analyzing received questions;

[1629] means for generating an answer based on the parsed question;

[1630] means for providing the generated answer to the user;

[1631] A system including:

[1632] (Claim 2)

[1633] 2. The system according to claim 1, further comprising means for analyzing a question from a user using natural language processing technology.

[1634] (Claim 3)

[1635] 10. The system of claim 1, further comprising means for removing unnecessary information from the collected data.

[1636] "Example 1"

[1637] (Claim 1)

[1638] a means for collecting data from a data collection system;

[1639] a means of filtering and cleaning the collected data;

[1640] a means for training a generative AI model using the filtered and cleaned data; and

[1641] a means for storing the knowledge learned by the generative AI model in a knowledge base;

[1642] means for accepting questions from users;

[1643] A means for analyzing the received questions using natural language processing technology;

[1644] means for retrieving information from a knowledge base and generating an answer based on the parsed question;

[1645] means for providing the generated answer to the user;

[1646] A system including:

[1647] (Claim 2)

[1648] 2. The system according to claim 1, further comprising means for searching a knowledge base for relevant information based on the content of a user's question.

[1649] (Claim 3)

[1650] 2. The system according to claim 1, further comprising means for deleting unnecessary information from the collected data to standardize the data format.

[1651] "Application Example 1"

[1652] (Claim 1)

[1653] a means for collecting data from a data collection system;

[1654] a means of filtering and cleaning the collected data;

[1655] a means for training a generative AI model using the filtered and cleaned data; and

[1656] a means for storing data learned by the generative AI model in a knowledge base;

[1657] means for accepting questions from users;

[1658] a means for analyzing received questions;

[1659] means for generating an answer based on the parsed question;

[1660] a means for collecting manufacturing process data and maintenance history from a factory management system and generating answers to questions based thereon;

[1661] means for providing the generated answer to the user;

[1662] A system including:

[1663] (Claim 2)

[1664] 2. The system according to claim 1, further comprising means for analyzing a question from a user using natural language processing technology.

[1665] (Claim 3)

[1666] 10. The system of claim 1, further comprising means for removing unnecessary information from the collected data.

[1667] "Example 2: Combining Emotion Engines"

[1668] (Claim 1)

[1669] a means for collecting data from a data collection system;

[1670] a means of filtering and cleaning the collected data;

[1671] a means for training a generative AI model using the filtered and cleaned data; and

[1672] a means for storing data learned by the generative AI model in a knowledge base;

[1673] means for accepting questions from users;

[1674] A means of analyzing emotions,

[1675] A means for analyzing the received questions using natural language processing technology;

[1676] means for generating an answer based on the parsed question and sentiment data;

[1677] means for providing the generated answer to the user;

[1678] A system including:

[1679] (Claim 2)

[1680] 10. The system of claim 1, further comprising means for analyzing user emotions using an emotion engine.

[1681] (Claim 3)

[1682] 10. The system of claim 1, further comprising means for adjusting the tone of the response by taking into account the emotional data.

[1683] "Application example 2 when combining emotion engines"

[1684] (Claim 1)

[1685] a means for collecting data from a data collection system;

[1686] a means of filtering and cleaning the collected data;

[1687] a means for training a generative AI model using the filtered and cleaned data; and

[1688] a means for storing data learned by the generative AI model in a knowledge base;

[1689] means for accepting questions from users;

[1690] a means for analyzing received questions;

[1691] means for generating an answer based on the parsed question;

[1692] means for recognizing user emotion data;

[1693] a means of adjusting the tone of responses taking into account perceived emotional data;

[1694] means for providing the generated answer to the user;

[1695] A system including:

[1696] (Claim 2)

[1697] 2. The system according to claim 1, further comprising means for analyzing a question from a user using natural language processing technology.

[1698] (Claim 3)

[1699] 10. The system of claim 1, further comprising means for removing unnecessary information from the collected data. [Explanation of symbols]

[1700] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. a means for collecting data from a data collection system; a means of filtering and cleaning the collected data; a means for training a generative AI model using the filtered and cleaned data; and a means for storing data learned by the generative AI model in a knowledge base; means for accepting questions from users; a means for analyzing received questions; means for generating an answer based on the parsed question; means for providing the generated answer to the user; A system including:

2. 2. The system according to claim 1, further comprising means for analyzing a question from a user using natural language processing technology.

3. 10. The system of claim 1, further comprising means for removing unnecessary information from the collected data.

Images