System

Abstract

A system is provided.SOLUTION: A system comprising: means for analyzing a question from a user using a natural language processing engine; means for retrieving relevant information from a database based on the analyzed question content; means for using a natural language generation model to generate an answer based on the retrieved information; means for sending the generated answer to the user's terminal; means for receiving user feedback; and means for improving the model using the received feedback.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] New employees are not familiar with the company's internal systems and information, making it difficult for them to obtain information efficiently. This increases the time spent searching for information, leading to problems such as reduced work efficiency. Furthermore, the large number of questions they have increases the burden on other employees and the HR department. Given this background, there is a growing need for a system that allows new employees to quickly and accurately obtain the information they need. [Means for solving the problem]

[0005] This invention provides a system that uses a natural language processing engine to analyze questions from users and searches for related information from a database based on the analyzed question content. This system presents appropriate answers to users by using a natural language generation model that generates answers based on the searched information. It also receives feedback from users and uses that feedback to improve the model, thereby improving the accuracy and reliability of the system. Furthermore, the system has a function that allows users to input questions in natural language through a user interface and a function that learns from question history and feedback to generate better answers, thereby streamlining information acquisition for new employees and improving work efficiency.

[0006] A "natural language processing engine" refers to software or algorithms that analyze text written in natural language, understand its meaning, and extract the necessary information.

[0007] "User" refers to new employees and general users who use the system to search for information and enter questions.

[0008] "Analyzing a question" refers to the process of classifying and structuring the question entered by the user based on specific patterns and keywords.

[0009] A "database" refers to a system or storage in which information to be searched is organized and stored.

[0010] "Searching for relevant information" refers to the process of quickly and efficiently extracting relevant information from a database based on the analyzed question.

[0011] "Answer generation" refers to the process of automatically creating an appropriate response to a user's question based on the retrieved information.

[0012] A "natural language generation model" refers to a machine learning model that learns from large amounts of text data and generates sentences in natural language that humans can understand.

[0013] "User terminal" refers to an electronic device, such as a computer or smartphone, that a user uses to enter questions and receive answers from the system.

[0014] "User feedback" refers to reactions and opinions provided by users, such as evaluations and suggestions for improvement of answers provided by the system.

[0015] "Using feedback to improve models" refers to the process of analyzing user feedback and using it to adjust and retrain natural language processing and generative models to improve their performance.

[0016] "User Interface" means the graphical user interface (GUI) or other means of interaction that allows a user to access and operate a system.

[0017] "Question history" refers to a record of questions that a user has previously asked the system. [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 system is designed to enable new employees to efficiently access company information. It works in cooperation with three parties: the server, the terminal, and the user.

[0040] Server configuration and roles

[0041] The server runs a natural language processing engine and a natural language generation model and is connected to a database. This allows the server to fulfill the following roles:

[0042] 1. Receiving and parsing questions:

[0043] The server receives the user's question sent from the device and uses a natural language processing engine to analyze the question and extract appropriate keywords and meanings.

[0044] 2. Information Search:

[0045] Based on the parsed question, the server retrieves relevant information from a database.

[0046] 3. Generate answers:

[0047] Based on the search results, a natural language generation model is used to generate an answer for the user.

[0048] 4. Submit your response:

[0049] The generated answer is sent to the terminal and displayed to the user.

[0050] 5. Receive feedback and learn:

[0051] It receives user feedback and uses it to improve the AI ​​model, which will generate a better answer the next time you ask a question.

[0052] Device configuration and role

[0053] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[0054] 1. Receiving user input:

[0055] Users can enter questions in natural language by simply typing in the text box and clicking the submit button, which sends the question to the server.

[0056] 2. Show Answer:

[0057] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[0058] 3. Enter your feedback:

[0059] Users can provide feedback on the answer by clicking the feedback button, entering a rating and sending it to the server.

[0060] User operations

[0061] A user follows the steps below to search for the information they need through the system.

[0062] 1. Enter your question:

[0063] The user uses the terminal interface to enter a question in natural language.

[0064] 2. Check your answers:

[0065] The answer sent from the server is confirmed on the device screen.

[0066] 3. Providing Feedback:

[0067] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[0068] Specific examples

[0069] Example 1: How to apply for paid leave

[0070] User: Type into the device, "How do I request paid time off?"

[0071] Server: Receives the question, extracts keywords related to "paid leave" and "how to apply," and searches the database for related information. For example, it generates an answer such as, "Paid leave applications are made through the in-house portal site. Log in to the portal site and apply from the 'Paid Leave Application' menu." and sends it to the device.

[0072] Terminal: Displays the generated answer to the user.

[0073] Users: Rate the answers for accuracy and provide feedback.

[0074] Example 2: New employee training schedule

[0075] User: Type into terminal, "What is the new employee training schedule?"

[0076] Server: Receives the question, extracts the keywords "new employee training" and "schedule," and searches the database for related information. For example, it generates an answer such as "The new employee training schedule is listed on the company calendar. Please check here," and sends it to the device.

[0077] Terminal: Displays the generated answer to the user.

[0078] Users: Rate the answers for accuracy and provide feedback.

[0079] In this way, the system is designed to enable new employees to quickly and accurately obtain the information they need, thereby improving work efficiency.

[0080] The processing flow will be explained below.

[0081] Step 1:

[0082] User: Enter a question in natural language through the device's user interface. Enter a question such as "How do I apply for paid leave?" into the input form and click the submit button.

[0083] Step 2:

[0084] Terminal: The question text entered by the user is structured (for example, converted into JSON format) and sent to the server. Specifically, a program such as JavaScript sends the text data to the server via an AJAX request.

[0085] Step 3:

[0086] Server: Receives the question text sent from the terminal. Specifically, it receives an HTTP request using a web framework such as Flask or Django, and the parser reads the data in JSON format.

[0087] Step 4:

[0088] Server: The question text is analyzed using a natural language processing engine (e.g., spaCy, NLTK). During this analysis process, the text is tokenized and subjected to entity recognition and partial analysis to extract important keywords and phrases such as "paid leave" and "how to apply."

[0089] Step 5:

[0090] Server: Based on the analyzed keywords, it searches for relevant information from a database (e.g., MySQL, PostgreSQL). It generates SQL queries and executes them against the database to retrieve the relevant information.

[0091] Step 6:

[0092] Server: Based on the acquired data, a natural language generation model (e.g., GPT-3) is used to generate an answer. For example, it generates a specific sentence such as, "Paid leave applications can be made through the in-house portal site. Please log in to the portal site and apply from the 'Paid Leave Application' menu."

[0093] Step 7:

[0094] Server: Structure the generated answer (e.g., in JSON format) and send it to the device as an HTTP response, using the Flask or Django response object with the appropriate status code.

[0095] Step 8:

[0096] Terminal: Parses the received answer and displays it in the user interface. Specifically, JavaScript parses the response data and inserts the answer text into an HTML element to visually display it to the user.

[0097] Step 9:

[0098] User: Review the displayed answers and rate them for appropriateness. Once you have finished rating, click the feedback button to provide your feedback.

[0099] Step 10:

[0100] Terminal: Structure the feedback entered by the user (convert it to JSON format) and send it to the server. Again using JavaScript, send the feedback data to the server via an AJAX request.

[0101] Step 11:

[0102] Server: Stores the received feedback in a database. Generates an SQL query to insert into the feedback table in the database.

[0103] Step 12:

[0104] Server: Retrains the NLP engine or NLG model based on the stored feedback, learning to improve the model's performance with new feedback data.

[0105] These are the specific processing steps of the new employee support generation AI system. This system allows new employees to efficiently obtain internal company information and improve work efficiency.

[0106] Example 1

[0107] 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."

[0108] In today's corporate environment, there is a demand for systems that allow new employees to efficiently access internal information and quickly obtain the information they need. Conventional systems have problems such as complicated information search and retrieval, making it difficult for users to operate intuitively, and reducing work efficiency. It has also been difficult to effectively utilize user feedback to improve the system's response accuracy.

[0109] 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.

[0110] In this invention, the server includes: means for analyzing a question from a user using a natural language processing engine; means for searching for related information from an information storage device based on the analyzed question content; means for using a natural language generation model to generate an answer based on the searched information; means for transmitting the generated answer to the user's information display device; means for receiving user evaluation information; means for improving the model using the received evaluation information; means for analyzing the question entered by the user and generating an answer by inputting a prompt sentence into the generation AI model based on the analysis result; means for displaying the generated answer; and means for analyzing the evaluation information provided by the user and reflecting it in the next answer generation. This allows new employees to access internal company information efficiently and intuitively. Furthermore, the accuracy of answer generation can be improved based on user feedback.

[0111] A "natural language processing engine" is a software technology that analyzes questions from users and extracts keywords and meanings.

[0112] A "natural language generation model" is an algorithm for generating appropriate answers in natural language based on the analyzed question content.

[0113] An "information storage device" is hardware or software for storing and managing information, such as a database or file system.

[0114] An "information display device" is a screen, monitor, or similar device for displaying information to a user.

[0115] "Rating Information" refers to feedback and ratings provided by users, and is data used to improve the system's response accuracy.

[0116] A "prompt" is a statement of instructions or a question that is input into a generative AI model to generate a specific answer.

[0117] "Analysis" is the process of understanding a user's question, extracting meaning, and identifying keywords.

[0118] "Generation" is the process of generating new, appropriate answers or text based on the retrieved information.

[0119] "Sending" is the process of sending a response or data from the server to the terminal over a communication path.

[0120] "Learning" refers to the application of machine learning algorithms to improve the system's response accuracy based on user feedback.

[0121] MODE FOR CARRYING OUT THE INVENTION

[0122] The system of the present invention is designed to enable new employees to efficiently access company information and quickly obtain the information they need. It uses a natural language processing engine, a natural language generation model, an information storage device, and an information display device in cooperation with a server, a terminal, and a user.

[0123] First, the user inputs a question into the terminal. The terminal allows the user to input questions in natural language through a user interface. This interface includes a text box and a submit button, and is designed to be intuitive. For example, if the user inputs "How do I apply for paid vacation?", the question will be submitted.

[0124] Next, the device sends the user's question to the server. The server receives the question and analyzes it using a natural language processing engine. Software such as spaCy or NLTK can be used as the natural language processing engine. The question is analyzed to extract keywords such as "paid leave" and "how to apply."

[0125] Based on the parsed query, the server searches for relevant information using an information storage device, such as MySQL or PostgreSQL, and executes a database query to retrieve the appropriate information.

[0126] Next, the server uses a natural language generation model to generate an answer based on the search results. OpenAI's GPT-3 and other generative AI models can be applied. As a specific example, the prompt "Please explain how to apply for paid leave" is entered, and an appropriate answer is generated. The generated answer is something like, "Paid leave applications are made through the in-house portal site. Please log in to the portal site and apply from the 'Paid Leave Application' menu."

[0127] The generated answer is sent from the server to the terminal and displayed to the user, allowing the user to quickly obtain the information they need. The user can also enter evaluation information for the provided answer, i.e., feedback. Feedback can include evaluations such as "It was helpful" or "I'd like more details."

[0128] Finally, the server receives user feedback and analyzes the ratings to improve the natural language generation model, using machine learning algorithms to adjust the model to generate better answers for the next question.

[0129] For example, if the user's evaluation feedback is that the answer was not specific, the server will adjust the model parameters to provide more detailed information the next time it responds, thereby improving the response quality of the entire system.

[0130] This system allows new employees to efficiently access internal company information and quickly obtain the information they need. Furthermore, response accuracy can be improved based on user feedback, allowing for continuous improvement.

[0131] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0132] Step 1: Enter your question

[0133] The user enters a question into the text box on the terminal. For example, the user enters "How do I apply for paid leave?". Specifically, the user enters the question using the keyboard and clicks the send button. Input: User's question. Output: Question data in text format.

[0134] Step 2: Submit your question

[0135] The device sends the question entered by the user to the server. Specifically, when the send button is clicked, the device creates an HTTP request and sends the question to the server's API endpoint. Input: Question data in text format. Output: HTTP request to the server.

[0136] Step 3: Analyzing the Question

[0137] The server analyzes the question received from the device. A natural language processing engine (e.g., spaCy or NLTK) is used to tokenize the question and extract key keywords (e.g., "paid leave" and "how to apply"). Specifically, the question is tokenized and tagged with parts of speech to identify important keywords. Input: Text question data extracted from the HTTP request. Output: Extracted keywords and their analysis results.

[0138] Step 4: Finding information

[0139] The server searches for related information from an information storage device (e.g., MySQL or PostgreSQL) based on the analysis results. Specifically, the server generates an SQL query to retrieve relevant records from the database. Input: Parsed keywords. Output: Related information retrieved from the database.

[0140] Step 5: Generate an answer

[0141] The server inputs a prompt sentence into a natural language generation model (for example, OpenAI's GPT-3) based on the acquired information, and generates an appropriate answer. Specifically, a prompt sentence such as "Please explain how to apply for paid leave" is input into the generative AI model, and the generated text is used as the answer. Input: Relevant information acquired from the database. Output: Generated answer text.

[0142] Step 6: Submit your response

[0143] The server sends the generated response to the terminal. Specifically, it includes the generated text as an HTTP response and sends it to the terminal. Input: Generated response text. Output: HTTP response to the terminal.

[0144] Step 7: Check your answers

[0145] The user checks the answer from the server displayed on the device. For example, the answer displayed is "Paid leave applications can be made through the in-house portal site." In concrete terms, the user looks at the device screen, reads the answer, and understands its contents. Input: Answer text received from the server. Output: User's understanding.

[0146] Step 8: Enter and submit your feedback

[0147] The user inputs feedback for the provided answer. For example, they rate it as "very helpful" or "I'd like more details." Specifically, the user clicks the feedback button, writes their rating, and clicks the submit button. Input: User feedback. Output: Feedback input to the server.

[0148] Step 9: Analyze feedback and learn

[0149] The server receives feedback from users and analyzes it to improve the quality of the AI ​​model. Specifically, the feedback data is stored in a database and evaluated by a machine learning algorithm. This adjusts the AI ​​model to generate a better answer for the next question. Input: User feedback. Output: Improvement and adjustment of the AI ​​model.

[0150] (Application example 1)

[0151] 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."

[0152] In logistics centers, it is important for new employees and workers to quickly obtain the information they need to perform their work in order to carry out their work efficiently. However, accessing complex work procedures and a wide range of inventory information requires advanced knowledge and experience, which places a heavy burden on new employees and first-time workers. A system that solves this issue and simplifies access to information is needed.

[0153] 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.

[0154] In this invention, the server includes means for analyzing instructions from a user using a natural language processing engine, means for searching for related information from an information storage unit based on the analyzed instructions, means for using a natural language generation algorithm to generate an answer based on the searched information, means for transmitting the generated answer to the user's display device, means for receiving user evaluations, means for improving the algorithm using the received evaluations, communication means for the user to request individual information via a wide area network, and dynamic search means for acquiring the requested individual information in real time, thereby enabling new employees and workers to quickly and accurately obtain the information they need.

[0155] A "natural language processing engine" is a software technology for analyzing text data and extracting keywords and meanings from within sentences.

[0156] "User instructions" refer to questions or commands that a user enters into the system.

[0157] The "information storage unit" refers to a database or storage system that stores related information.

[0158] "Related information" refers to the information to be searched for based on a user instruction.

[0159] A "natural language generation algorithm" is an algorithm that generates answers in natural language based on analyzed information.

[0160] A "display device" is a device for visually displaying information to a user.

[0161] "Rating" refers to the feedback or review a user gives to an answer provided by the system.

[0162] "Wide area network" refers to a wide-area communications network such as the Internet.

[0163] "Communication means" refers to the technology for sending user instructions to the server and for sending responses from the server to the user.

[0164] "Dynamic search means" refers to a technology that instantly searches for and retrieves requested information in real time.

[0165] This invention is a system that enables new employees and workers at a logistics center to quickly obtain information necessary for their work. This system operates in cooperation with three parties: a server, a terminal, and a user.

[0166] Server configuration and roles

[0167] The server runs a natural language processing engine and a natural language generation algorithm, and is connected to the information storage unit. The server's roles are as follows:

[0168] 1. Receiving and parsing instructions:

[0169] The server receives user instructions sent from the device and uses a natural language processing engine (e.g., spaCy) to analyze the instructions and extract relevant keywords and meanings.

[0170] 2. Information Search:

[0171] Based on the parsed instructions, the server searches for relevant information from an information storage unit (e.g., a MySQL database).

[0172] 3. Generate answers:

[0173] Based on the search results, a natural language generation algorithm (e.g., GPT-4) is used to generate an answer for the user.

[0174] 4. Submit your response:

[0175] The generated answer is sent to the terminal and displayed to the user.

[0176] 5. Receiving and learning from assessments:

[0177] It receives user ratings and uses them to improve its algorithms, which will generate better answers for the next prompt.

[0178] Device configuration and role

[0179] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[0180] 1. Receiving user input:

[0181] Users can enter instructions in natural language by simply typing in a text box and clicking the submit button, which sends the instructions to the server.

[0182] 2. Show Answer:

[0183] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[0184] 3. Enter your rating:

[0185] Users can provide a rating for an answer by clicking the rating button, which enters their feedback and sends it to the server.

[0186] User operations

[0187] A user follows the steps below to search for the information they need through the system.

[0188] 1. Enter instructions:

[0189] The user uses the terminal's interface to input instructions in natural language.

[0190] For example: "Please let me know the availability of the product."

[0191] 2. Check your answers:

[0192] The answer sent from the server is confirmed on the device screen.

[0193] 3. Providing Evaluations:

[0194] The answer is evaluated for appropriateness and the evaluation is sent to the server via the terminal.

[0195] Providing concrete examples and prompts

[0196] Specific examples

[0197] Consider the case where a new employee enters the following instructions at a logistics center:

[0198] "Please let me know the product availability."

[0199] The server uses a natural language processing engine (spaCy) to extract keywords such as "product" and "stock status," and searches for related information in the information storage unit (MySQL database).Then, it uses a natural language generation algorithm (GPT-4) to generate the following answer:

[0200] "We currently have 50 units of product A in stock. They are ready to ship."

[0201] Prompt Sentence Examples

[0202] The prompt has the following format:

[0203] "Provide a detailed answer for: Product A's stock status"

[0204] This system is designed to enable new employees and workers to quickly and accurately obtain the information they need, thereby improving operational efficiency at logistics centers.

[0205] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0206] Step 1:

[0207] The user uses the device interface to input instructions in natural language, for example, "Please tell me the product availability status," and clicks the submit button. This input is sent from the device to the server.

[0208] Step 2:

[0209] The server receives the user's instructions from the device and analyzes them using a natural language processing engine (e.g., spaCy). Specifically, it tokenizes the instruction text and extracts keywords and meanings. The user's instructions are given as input, and keywords are obtained as output.

[0210] Step 3:

[0211] The server searches for related information from the information storage unit (database example: MySQL) based on the analyzed keywords. For example, if "product inventory status" is extracted as a keyword, the corresponding inventory information is searched for in the database. The extracted keywords are given as input, and the search results (inventory status, etc.) are obtained as output.

[0212] Step 4:

[0213] The server generates an answer for the user using a natural language generation algorithm (e.g., GPT-4) based on the search results. Specifically, the search results are given in the form of a prompt sentence, and an answer in natural language is generated by the generative AI model. The search results are given as input, and the generated answer is obtained as output.

[0214] Step 5:

[0215] The server sends the generated answer to the terminal. The terminal displays the answer received from the server to the user. The generated answer is sent from the server as input, received by the terminal, and displayed to the user as output.

[0216] Step 6:

[0217] The user checks the displayed answers and rates them. The rating is sent from the device to the server. This rating is used to improve the algorithm. The user's rating is sent from the device to the server as input and reflected in the algorithm model as output.

[0218] Step 7:

[0219] The server uses the received evaluations to improve the natural language generation algorithm. Specifically, it adds the evaluation data as training data to improve the accuracy of the next answer generation. The evaluation data is given as input and an improved generation algorithm is obtained as output.

[0220] 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.

[0221] This invention is a system designed to enable new employees to efficiently access company information, and includes an emotion engine that recognizes the user's emotions and adjusts responses. This system operates in cooperation between a server, a terminal, and a user.

[0222] Server configuration and roles

[0223] The server runs a natural language processing engine, a natural language generation model, a database, and an emotion engine, and performs the following roles:

[0224] 1. Receiving and parsing questions:

[0225] The server receives the user's question sent from the device. Using a natural language processing engine, it analyzes the question and extracts appropriate keywords and meanings. For example, it may receive a question such as, "How do I apply for paid leave?"

[0226] 2. Emotion Recognition:

[0227] The emotion engine analyzes the user's emotions from the content of their question, for example, recognizing whether the user is feeling anxious or uncertain.

[0228] 3. Information Search:

[0229] Based on the analyzed keywords and emotion data, the server retrieves relevant information from a database.

[0230] 4. Generate and refine answers:

[0231] Based on the search results and sentiment data, a natural language generation model is used to generate a response for the user. Based on the sentiment engine data, the tone and content of the response are adjusted. For example, if the user is feeling anxious, a phrase such as "Don't worry" is added to the response.

[0232] 5. Submit your response:

[0233] The generated answer is sent to the terminal and displayed to the user.

[0234] 6. Receive feedback and learn:

[0235] It receives user feedback and uses it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[0236] Device configuration and role

[0237] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[0238] 1. Receiving user input:

[0239] Users can enter questions in natural language by simply typing in the text box and clicking the submit button, which sends the question to the server.

[0240] 2. Show Answer:

[0241] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[0242] 3. Enter your feedback:

[0243] Users can provide feedback on the answer by clicking the feedback button, entering a rating and sending it to the server.

[0244] User operations

[0245] A user follows the steps below to search for the information they need through the system.

[0246] 1. Enter your question:

[0247] The user uses the terminal interface to enter a question in natural language.

[0248] 2. Check your answers:

[0249] The answer sent from the server is confirmed on the device screen.

[0250] 3. Providing Feedback:

[0251] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[0252] Specific examples

[0253] Example 1: How to apply for paid leave

[0254] User: Type into the device, "How do I request paid time off?"

[0255] Server: Receives the question and extracts the keywords "paid leave" and "how to apply" as well as the emotion "anxiety." It searches the database for relevant information and uses a natural language generation model to generate an answer such as, "Paid leave applications are made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu." This is then sent to the device.

[0256] Terminal: Displays the generated answer to the user.

[0257] Users: Rate the answers for accuracy and provide feedback.

[0258] Example 2: New employee training schedule

[0259] User: Type into terminal, "What is the new employee training schedule?"

[0260] Server: Receives the question and extracts the keywords "new employee training" and "schedule" as well as the sentiment of "doubt." It searches for related information in the database and uses a natural language generation model to generate an answer such as "The new employee training schedule is listed on the company calendar. Please check here." and sends it to the device.

[0261] Terminal: Displays the generated answer to the user.

[0262] Users: Rate the answers for accuracy and provide feedback.

[0263] In this way, the system not only enables new employees to quickly and accurately obtain the information they need, but also utilizes emotional data to provide more personalized assistance, resulting in improved work efficiency and a greater sense of security for new employees.

[0264] The processing flow will be explained below.

[0265] Step 1:

[0266] User: Enter a question in natural language through the device's user interface, such as "How do I apply for paid leave?", and click the submit button.

[0267] Step 2:

[0268] Terminal: The question text entered by the user is structured (for example, converted into JSON format) and sent to the server. Specifically, a program such as JavaScript sends the text data to the server via an AJAX request.

[0269] Step 3:

[0270] Server: Receives the question text sent from the terminal. Using a web framework such as Flask or Django, it receives an HTTP request and the parser reads the data in JSON format.

[0271] Step 4:

[0272] Server: The question text is analyzed using a natural language processing engine (e.g., spaCy, NLTK). Specifically, the text is tokenized and subjected to entity recognition and partial analysis to extract important keywords and phrases such as "paid leave" and "how to apply."

[0273] Step 5:

[0274] Server: Uses an emotion engine to analyze the emotions contained in the user's question text. For example, it identifies whether the user is feeling "anxiety" or "doubt."

[0275] Step 6:

[0276] Server: Based on the analyzed keywords and sentiment data, it searches for relevant information from a database (e.g., MySQL, PostgreSQL). It generates SQL queries and executes them against the database to retrieve relevant information.

[0277] Step 7:

[0278] Server: Based on the acquired data, a natural language generation model (e.g., GPT-3) is used to generate an answer. The generation process takes into account emotional data and optimizes the tone and content of the answer. For example, it generates a specific sentence such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu."

[0279] Step 8:

[0280] Server: Structure the generated answer (e.g., in JSON format) and send it to the device as an HTTP response, using Flask or Django's response object with the appropriate status code.

[0281] Step 9:

[0282] Terminal: Parses the received answer and displays it in the user interface. JavaScript parses the response data and inserts the answer text into an HTML element to display it visually to the user.

[0283] Step 10:

[0284] User: Review the displayed answers and rate them for appropriateness. Once you have finished rating, click the feedback button to provide your feedback.

[0285] Step 11:

[0286] Terminal: Structure the feedback entered by the user (convert it to JSON format) and send it to the server. Using JavaScript, send the feedback data to the server via an AJAX request.

[0287] Step 12:

[0288] Server: Stores the received feedback in a database. Generates an SQL query to insert into the feedback table in the database.

[0289] Step 13:

[0290] Server: Retrains the NLP engine, NLG model, and sentiment engine based on the stored feedback. Each model learns using new feedback data to improve its performance.

[0291] These are the specific processing steps of the new employee support generation AI system that combines an emotion engine. This system allows new employees to efficiently obtain internal company information and receive personalized support based on their emotions.

[0292] Example 2

[0293] 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."

[0294] When new employees start work, it is urgent to provide them with a means to efficiently access internal company information. However, conventional systems provide mechanical responses without considering the user's feelings, making it difficult to improve the user's psychological sense of security and satisfaction. Furthermore, the system lacks the ability to adaptively improve itself based on feedback, making it difficult to provide information that meets the user's needs.

[0295] 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.

[0296] In this invention, the server includes: means for analyzing user questions using a natural language processing engine; means for analyzing the user's emotions using the analyzed question content and an emotion recognition engine; means for searching for related information from a database based on the analyzed question content and emotion data; means for generating answers based on the searched information and using a generative AI model that adjusts the content according to the emotion; means for sending the generated answers to the user's device; means for receiving user feedback; and means for improving the model using the received feedback. This not only enables new employees to quickly and accurately access the information they need, but also provides personalized answers based on emotion data. Furthermore, by improving the system using feedback, user satisfaction and reliability can be continuously improved.

[0297] A "natural language processing engine" is a software technology that analyzes text entered by a user and understands its content and structure.

[0298] An "emotion recognition engine" is a technology that analyzes emotions from a user's text and identifies emotional states such as anxiety, anger, and joy.

[0299] A "database" is a data management system for storing related information in a structured way that makes it easy to search, store, and update.

[0300] A "generative AI model" is a machine learning model that uses natural language data as input and automatically generates appropriate answers based on the context.

[0301] A "user terminal" is an electronic device that a user uses to access the system, enter questions, and view answers.

[0302] "Feedback" refers to the user's evaluation of the system's answers and suggestions for improvement.

[0303] "Model improvement" is the process of correcting and updating the system's learning data and algorithms based on the feedback received, thereby improving answer accuracy and user satisfaction.

[0304] This invention is a system designed to enable new employees to efficiently access company information, and includes an emotion recognition engine to recognize the user's emotions and adjust responses accordingly. This system operates in cooperation between a server, a terminal, and a user.

[0305] Server configuration and roles

[0306] The following main software modules run on the server:

[0307] 1. Natural Language Processing Engines (e.g. SpaCy, NLTK):

[0308] It is used to analyze user questions and extract keywords and context.

[0309] 2. Emotion Recognition Engine (e.g. Hugging Face Transformers, IBM Watson):

[0310] Analyzes emotions from user text and recognizes feelings such as anxiety and doubt.

[0311] 3. Database (e.g. MySQL, PostgreSQL):

[0312] Store and search relevant information within your company.

[0313] 4. Generative AI models (e.g., OpenAI GPT-3, BERT):

[0314] Generate appropriate answers for users based on search results and sentiment data.

[0315] The server's operating process is as follows:

[0316] 1. Receiving and parsing questions:

[0317] The server receives the user's question sent from the device and analyzes it using a natural language processing engine. For example, it may receive a question such as "How do I apply for paid vacation?"

[0318] 2. Emotion Recognition:

[0319] An emotion recognition engine is used to analyze the emotions of the user based on the content of the question. For example, it can recognize that the user is feeling anxious.

[0320] 3. Information Search:

[0321] Based on the analyzed keywords and emotion data, the server searches the database for relevant information, for example, identifying information on "paid vacation" and "how to apply."

[0322] 4. Generate and refine answers:

[0323] Based on the search results and sentiment data, a generative AI model is used to generate a response for the user. The tone and content of the response are adjusted based on the sentiment data. For example, a response such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu" can be generated.

[0324] 5. Submit your response:

[0325] The generated response is sent to the device.

[0326] 6. Receive feedback and learn:

[0327] It receives user feedback and uses it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[0328] Device configuration and role

[0329] The terminal provides an intuitive and easy-to-use interface for users to access the system, which:

[0330] 1. Receiving user input:

[0331] Allow users to enter their question in natural language, for example by typing their question in a text box and clicking the submit button.

[0332] 2. Show Answer:

[0333] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[0334] 3. Enter your feedback:

[0335] Users can provide feedback on the answers, for example by clicking a feedback button and entering a rating, which is then sent to the server.

[0336] User operations

[0337] A user follows the steps below to search for the information they need through the system.

[0338] 1. Enter your question:

[0339] Users use the device interface to enter questions in natural language.

[0340] 2. Check your answers:

[0341] The answer sent from the server is confirmed on the device screen.

[0342] 3. Providing Feedback:

[0343] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[0344] Specific examples

[0345] Example 1: How to apply for paid leave

[0346] User: Type "How do I request paid time off?" into the device.

[0347] Server: Receives the question and extracts the keywords "paid leave" and "how to apply" as well as the emotion "anxiety." It searches the database for relevant information and uses a generative AI model to generate an answer such as, "Paid leave applications are made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu." This is then sent to the device.

[0348] Terminal: Displays the generated answer to the user.

[0349] Users: Rate the answers for accuracy and provide feedback.

[0350] Example 2: New employee training schedule

[0351] User: Type into terminal, "What is the new employee training schedule?"

[0352] Server: Receives the question and extracts the keywords "new employee training" and "schedule" as well as the emotion "doubt." Searches for related information in the database and uses a generative AI model to generate an answer such as "The new employee training schedule is listed on the company calendar. Please check here." and sends it to the device.

[0353] Terminal: Displays the generated answer to the user.

[0354] Users: Rate the answers for accuracy and provide feedback.

[0355] In this way, the system not only enables new employees to quickly and accurately obtain the information they need, but also utilizes emotional data to provide more personalized support, resulting in improved work efficiency and a greater sense of security for new employees.

[0356] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0357] Step 1:

[0358] The user inputs a question into the device interface. For example, the user inputs "How do I apply for paid leave?" into the text box. This operation saves the question in text format on the device.

[0359] Step 2:

[0360] When the user clicks the submit button, the device sends the question to the server as an HTTP POST request, which includes the question text.

[0361] Step 3:

[0362] The server receives an HTTP POST request from the device. It parses the question text and extracts keywords and contextual information using a natural language processing engine (e.g., SpaCy, NLTK). This process extracts keywords such as "paid vacation" and "how to apply" from the question text.

[0363] Step 4:

[0364] The server analyzes the user's emotions using an emotion recognition engine (e.g., Hugging Face Transformers, IBM Watson) based on the extracted keywords. For example, it detects the user's anxiety from the question text. This process generates emotion data such as "anxiety."

[0365] Step 5:

[0366] The server uses the keywords and sentiment data to search for relevant information from a database (e.g., MySQL, PostgreSQL). For example, it searches for company policies and procedures related to "paid leave" and "how to apply." This process retrieves the relevant information in text format.

[0367] Step 6:

[0368] The server generates an answer using a generative AI model (e.g., OpenAI GPT-3, BERT) based on the search results and emotion data. The generated prompt includes keywords, emotion data, and search results. This process generates an answer text such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu."

[0369] Step 7:

[0370] The server sends the generated answer text to the terminal as an HTTP response, which includes the answer text.

[0371] Step 8:

[0372] The terminal displays the answer text received from the server to the user, for example, by displaying the answer in a text area on the screen.

[0373] Step 9:

[0374] The user inputs feedback on the displayed answer, for example, by clicking a feedback button and inputting an evaluation such as "appropriate" or "inappropriate."

[0375] Step 10:

[0376] The device sends the user-provided feedback to the server as an HTTP POST request.

[0377] Step 11:

[0378] The server receives feedback sent from the device and analyzes the feedback data. Based on this data, it evaluates the performance of the AI ​​model and adjusts the model or updates the training data as needed, thereby improving the accuracy and reliability of the system.

[0379] (Application example 2)

[0380] 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."

[0381] In modern factories, new employees and technicians need a large amount of information and quick access to it in order to adapt to complex machines and processes. However, with previous systems, obtaining information and troubleshooting took time, often causing anxiety and stress. Therefore, an information delivery system that can provide efficient and emotional support is needed.

[0382] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0383] In this invention, the server includes a means for analyzing a user's question using a natural language processing engine, a means for searching for related information from a database based on the analyzed question, and a means for using a natural language generation model to generate an answer based on the searched information. This allows the user to receive an answer tailored to their own emotions. Furthermore, by providing an interface accessible from the robot's display or smart device and adding a function for learning from the user's question history and feedback to generate better answers, the server achieves efficient and stress-free information acquisition.

[0384] A "natural language processing engine" is an engine that analyzes questions entered by users and extracts keywords and context.

[0385] A "database" is a collection of information that stores information in response to user questions and manages it in a searchable format.

[0386] A "natural language generation model" is an algorithm for generating appropriate answers to users based on analyzed questions.

[0387] The "emotion engine" is an engine that analyzes emotions from the user's question and input data and adjusts the response content.

[0388] A "terminal" is a device with an interface that allows a user to access the system and input questions.

[0389] An "interface" is an operation screen or application that allows a user to input questions into the system and receive answers.

[0390] "Feedback" refers to the evaluations and opinions that users provide regarding the system's answers, and is used to improve the system.

[0391] A "robot display" is a device installed on a factory robot that displays operation information and responses from the system.

[0392] A "smart device" is a portable electronic device that can connect to the Internet, such as a smartphone or tablet.

[0393] "Question history" is a record of questions and answers that a user has previously entered into the system.

[0394] "Troubleshooting" is the process of diagnosing errors or problems in a system or machine and providing a solution.

[0395] To implement this invention, an automated information acquisition system must be installed on factory robots and related smart devices. This system works in cooperation with a server, terminals, and users.

[0396] Server configuration and roles

[0397] The server runs a natural language processing engine, a natural language generation model, a database, and an emotion engine. Details are as follows:

[0398] 1. Receiving and analyzing the question: The server receives the user's question sent from the terminal. Using a natural language processing engine, the server analyzes the question and extracts appropriate keywords and meanings. For example, the server may receive a question such as, "Please tell me how to deal with error code E101."

[0399] 2. Emotion Recognition: Using an emotion engine, we analyze the emotions in the user's questions, for example, whether the user is feeling anxious or uncertain.

[0400] 3. Information retrieval: Based on the analyzed keywords and sentiment data, the server retrieves relevant information from the database.

[0401] 4. Answer generation and tailoring: Based on the search results and sentiment data, a natural language generation model is used to generate an answer for the user. The tone and content of the answer are adjusted based on the sentiment engine data. For example, if the user is feeling anxious, a phrase like "Don't worry" is added to the answer.

[0402] 5. Sending the answer: The generated answer is sent to the terminal and displayed to the user.

[0403] 6. Receive and learn feedback: Receive user feedback and use it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[0404] Device configuration and role

[0405] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[0406] 1. Receiving user input: Users can input questions in natural language. Questions can be posed via text or voice using the factory robot's display or smart device.

[0407] 2. Displaying the answer: The answer received from the server is displayed to the user, allowing the user to quickly obtain the required information.

[0408] 3. Enter feedback: Users can provide feedback on the answer by clicking the feedback button, entering their rating and sending it to the server.

[0409] User operations

[0410] A user follows the steps below to search for the information they need through the system.

[0411] 1. Entering a question: Users enter a question in natural language using the factory robot's display or a smart device.

[0412] 2. Confirm the answer: Confirm the answer sent from the server on the device screen.

[0413] 3. Providing feedback: Evaluate the appropriateness of the answer and send feedback to the server via the device.

[0414] Specific examples

[0415] Example 1: How to resolve a robot error code

[0416] User: Enter "Please tell me how to deal with error code E101" into the terminal.

[0417] Server: Receives the question and extracts the keywords "error code E101" and "how to deal with it," as well as the emotion "anxiety." It searches for related information in a database and uses a natural language generation model to generate a response such as "Error code E101 indicates a sensor malfunction. Don't worry, try reconnecting the sensor," and sends it to the device.

[0418] Terminal: Displays the generated answer to the user.

[0419] User: Check the solution and perform the task.

[0420] Prompt Sentence Examples

[0421] "How do I resolve the error code E101?"

[0422] This system enables new employees and technicians to quickly address any problems that arise within the factory, improving work efficiency and safety.

[0423] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0424] Processing Steps

[0425] Step 1:

[0426] The user inputs a question in natural language using the factory robot's display or a smart device. For example, the user might input a question such as, "Please tell me how to deal with error code E101." The user's input is sent to the terminal as text data.

[0427] Step 2:

[0428] The device sends text data from the user to the server. Specifically, the device transfers the user's input data to the server as an API request.

[0429] Step 3:

[0430] The server analyzes the user's question received from the terminal using a natural language processing engine. As a result of the analysis, the keywords "error code E101" and "solution" are extracted. In this analysis process, the input text data is broken down into tokens and semantic analysis is performed.

[0431] Step 4:

[0432] The server uses an emotion engine to recognize emotions from the user's question. For example, the emotion "anxiety" is extracted. In this step, an emotion analysis algorithm is applied to the user's input data to generate emotion data.

[0433] Step 5:

[0434] The server searches for relevant information from the database based on the analyzed keywords and emotion data. For example, it retrieves information about "error code E101" in the database. In this search process, an SQL or NoSQL query is constructed to retrieve the corresponding record from the database.

[0435] Step 6:

[0436] The server generates a response to the user using a natural language generation model based on the search results and sentiment data. For example, a response such as "Error code E101 indicates a sensor malfunction. Please rest assured and try reconnecting the sensor" is generated. In this generation process, a natural language generation algorithm is applied to construct the response text.

[0437] Step 7:

[0438] The server sends the generated response to the device. Specifically, it returns the response data to the device as an API response. In this sending process, the generated text data is packaged in JSON format and sent as an HTTP response.

[0439] Step 8:

[0440] The device displays the answer received from the server to the user. Specifically, the generated answer is displayed on the display. In this display process, the received text data is drawn on the UI.

[0441] Step 9:

[0442] The user evaluates whether the answer is appropriate and sends feedback to the server via the terminal. For example, the user may enter feedback such as "This answer was helpful." In this step, the feedback data is sent to the terminal as text.

[0443] Step 10:

[0444] The server receives feedback from users and uses it to improve the AI ​​model. This improvement process uses the received feedback data as training data to improve the accuracy of the natural language processing engine and natural language generation model.

[0445] These processing steps provide fast and appropriate answers to user questions while continually improving the accuracy and reliability of the overall system.

[0446] 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.

[0447] 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.

[0448] 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.

[0449] [Second embodiment]

[0450] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0451] 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.

[0452] 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).

[0453] 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.

[0454] 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.

[0455] 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).

[0456] 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. 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.

[0457] 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.

[0458] 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.

[0459] 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.

[0460] In the smart glasses 214, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0461] 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."

[0462] This system is designed to enable new employees to efficiently access company information. It works in cooperation with three parties: the server, the terminal, and the user.

[0463] Server configuration and roles

[0464] The server runs a natural language processing engine and a natural language generation model and is connected to a database. This allows the server to fulfill the following roles:

[0465] 1. Receiving and parsing questions:

[0466] The server receives the user's question sent from the device and uses a natural language processing engine to analyze the question and extract appropriate keywords and meanings.

[0467] 2. Information Search:

[0468] Based on the parsed question, the server retrieves relevant information from a database.

[0469] 3. Generate answers:

[0470] Based on the search results, a natural language generation model is used to generate an answer for the user.

[0471] 4. Submit your response:

[0472] The generated answer is sent to the terminal and displayed to the user.

[0473] 5. Receive feedback and learn:

[0474] It receives user feedback and uses it to improve the AI ​​model, which will generate a better answer the next time you ask a question.

[0475] Device configuration and role

[0476] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[0477] 1. Receiving user input:

[0478] Users can enter questions in natural language by simply typing in the text box and clicking the submit button, which sends the question to the server.

[0479] 2. Show Answer:

[0480] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[0481] 3. Enter your feedback:

[0482] Users can provide feedback on the answer by clicking the feedback button, entering a rating and sending it to the server.

[0483] User operations

[0484] A user follows the steps below to search for the information they need through the system.

[0485] 1. Enter your question:

[0486] The user uses the terminal interface to enter a question in natural language.

[0487] 2. Check your answers:

[0488] The answer sent from the server is confirmed on the device screen.

[0489] 3. Providing Feedback:

[0490] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[0491] Specific examples

[0492] Example 1: How to apply for paid leave

[0493] User: Type into the device, "How do I request paid time off?"

[0494] Server: Receives the question, extracts keywords related to "paid leave" and "how to apply," and searches the database for related information. For example, it generates an answer such as, "Paid leave applications are made through the in-house portal site. Log in to the portal site and apply from the 'Paid Leave Application' menu." and sends it to the device.

[0495] Terminal: Displays the generated answer to the user.

[0496] Users: Rate the answers for accuracy and provide feedback.

[0497] Example 2: New employee training schedule

[0498] User: Type into terminal, "What is the new employee training schedule?"

[0499] Server: Receives the question, extracts the keywords "new employee training" and "schedule," and searches the database for related information. For example, it generates an answer such as "The new employee training schedule is listed on the company calendar. Please check here," and sends it to the device.

[0500] Terminal: Displays the generated answer to the user.

[0501] Users: Rate the answers for accuracy and provide feedback.

[0502] In this way, the system is designed to enable new employees to quickly and accurately obtain the information they need, thereby improving work efficiency.

[0503] The processing flow will be explained below.

[0504] Step 1:

[0505] User: Enter a question in natural language through the device's user interface. Enter a question such as "How do I apply for paid leave?" into the input form and click the submit button.

[0506] Step 2:

[0507] Terminal: The question text entered by the user is structured (for example, converted into JSON format) and sent to the server. Specifically, a program such as JavaScript sends the text data to the server via an AJAX request.

[0508] Step 3:

[0509] Server: Receives the question text sent from the terminal. Specifically, it receives an HTTP request using a web framework such as Flask or Django, and the parser reads the data in JSON format.

[0510] Step 4:

[0511] Server: The question text is analyzed using a natural language processing engine (e.g., spaCy, NLTK). During this analysis process, the text is tokenized and subjected to entity recognition and partial analysis to extract important keywords and phrases such as "paid leave" and "how to apply."

[0512] Step 5:

[0513] Server: Based on the analyzed keywords, it searches for relevant information from a database (e.g., MySQL, PostgreSQL). It generates SQL queries and executes them against the database to retrieve the relevant information.

[0514] Step 6:

[0515] Server: Based on the acquired data, a natural language generation model (e.g., GPT-3) is used to generate an answer. For example, it generates a specific sentence such as, "Paid leave applications can be made through the in-house portal site. Please log in to the portal site and apply from the 'Paid Leave Application' menu."

[0516] Step 7:

[0517] Server: Structure the generated answer (e.g., in JSON format) and send it to the device as an HTTP response, using the Flask or Django response object with the appropriate status code.

[0518] Step 8:

[0519] Terminal: Parses the received answer and displays it in the user interface. Specifically, JavaScript parses the response data and inserts the answer text into an HTML element to visually display it to the user.

[0520] Step 9:

[0521] User: Review the displayed answers and rate them for appropriateness. Once you have finished rating, click the feedback button to provide your feedback.

[0522] Step 10:

[0523] Terminal: Structure the feedback entered by the user (convert it to JSON format) and send it to the server. Again using JavaScript, send the feedback data to the server via an AJAX request.

[0524] Step 11:

[0525] Server: Stores the received feedback in a database. Generates an SQL query to insert into the feedback table in the database.

[0526] Step 12:

[0527] Server: Retrains the NLP engine or NLG model based on the stored feedback, learning to improve the model's performance with new feedback data.

[0528] These are the specific processing steps of the new employee support generation AI system. This system allows new employees to efficiently obtain internal company information and improve work efficiency.

[0529] Example 1

[0530] 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."

[0531] In today's corporate environment, there is a demand for systems that allow new employees to efficiently access internal information and quickly obtain the information they need. Conventional systems have problems such as complicated information search and retrieval, making it difficult for users to operate intuitively, and reducing work efficiency. It has also been difficult to effectively utilize user feedback to improve the system's response accuracy.

[0532] 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.

[0533] In this invention, the server includes: means for analyzing a question from a user using a natural language processing engine; means for searching for related information from an information storage device based on the analyzed question content; means for using a natural language generation model to generate an answer based on the searched information; means for transmitting the generated answer to the user's information display device; means for receiving user evaluation information; means for improving the model using the received evaluation information; means for analyzing the question entered by the user and generating an answer by inputting a prompt sentence into the generation AI model based on the analysis result; means for displaying the generated answer; and means for analyzing the evaluation information provided by the user and reflecting it in the next answer generation. This allows new employees to access internal company information efficiently and intuitively. Furthermore, the accuracy of answer generation can be improved based on user feedback.

[0534] A "natural language processing engine" is a software technology that analyzes questions from users and extracts keywords and meanings.

[0535] A "natural language generation model" is an algorithm for generating appropriate answers in natural language based on the analyzed question content.

[0536] An "information storage device" is hardware or software for storing and managing information, such as a database or file system.

[0537] An "information display device" is a screen, monitor, or similar device for displaying information to a user.

[0538] "Rating Information" refers to feedback and ratings provided by users, and is data used to improve the system's response accuracy.

[0539] A "prompt" is a statement of instructions or a question that is input into a generative AI model to generate a specific answer.

[0540] "Analysis" is the process of understanding a user's question, extracting meaning, and identifying keywords.

[0541] "Generation" is the process of generating new, appropriate answers or text based on the retrieved information.

[0542] "Sending" is the process of sending a response or data from the server to the terminal over a communication path.

[0543] "Learning" refers to the application of machine learning algorithms to improve the system's response accuracy based on user feedback.

[0544] MODE FOR CARRYING OUT THE INVENTION

[0545] The system of the present invention is designed to enable new employees to efficiently access company information and quickly obtain the information they need. It uses a natural language processing engine, a natural language generation model, an information storage device, and an information display device in cooperation with a server, a terminal, and a user.

[0546] First, the user inputs a question into the terminal. The terminal allows the user to input questions in natural language through a user interface. This interface includes a text box and a submit button, and is designed to be intuitive. For example, if the user inputs "How do I apply for paid vacation?", the question will be submitted.

[0547] Next, the device sends the user's question to the server. The server receives the question and analyzes it using a natural language processing engine. Software such as spaCy or NLTK can be used as the natural language processing engine. The question is analyzed to extract keywords such as "paid leave" and "how to apply."

[0548] Based on the parsed query, the server searches for relevant information using an information storage device, such as MySQL or PostgreSQL, and executes a database query to retrieve the appropriate information.

[0549] Next, the server uses a natural language generation model to generate an answer based on the search results. OpenAI's GPT-3 and other generative AI models can be applied. As a specific example, the prompt "Please explain how to apply for paid leave" is entered, and an appropriate answer is generated. The generated answer is something like, "Paid leave applications are made through the in-house portal site. Please log in to the portal site and apply from the 'Paid Leave Application' menu."

[0550] The generated answer is sent from the server to the terminal and displayed to the user, allowing the user to quickly obtain the information they need. The user can also enter evaluation information for the provided answer, i.e., feedback. Feedback can include evaluations such as "It was helpful" or "I'd like more details."

[0551] Finally, the server receives user feedback and analyzes the ratings to improve the natural language generation model, using machine learning algorithms to adjust the model to generate better answers for the next question.

[0552] For example, if the user's evaluation feedback is that the answer was not specific, the server will adjust the model parameters to provide more detailed information the next time it responds, thereby improving the response quality of the entire system.

[0553] This system allows new employees to efficiently access internal company information and quickly obtain the information they need. Furthermore, response accuracy can be improved based on user feedback, allowing for continuous improvement.

[0554] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0555] Step 1: Enter your question

[0556] The user enters a question into the text box on the terminal. For example, the user enters "How do I apply for paid leave?". Specifically, the user enters the question using the keyboard and clicks the send button. Input: User's question. Output: Question data in text format.

[0557] Step 2: Submit your question

[0558] The device sends the question entered by the user to the server. Specifically, when the send button is clicked, the device creates an HTTP request and sends the question to the server's API endpoint. Input: Question data in text format. Output: HTTP request to the server.

[0559] Step 3: Analyzing the Question

[0560] The server analyzes the question received from the device. A natural language processing engine (e.g., spaCy or NLTK) is used to tokenize the question and extract key keywords (e.g., "paid leave" and "how to apply"). Specifically, the question is tokenized and tagged with parts of speech to identify important keywords. Input: Text question data extracted from the HTTP request. Output: Extracted keywords and their analysis results.

[0561] Step 4: Finding information

[0562] The server searches for related information from an information storage device (e.g., MySQL or PostgreSQL) based on the analysis results. Specifically, the server generates an SQL query to retrieve relevant records from the database. Input: Parsed keywords. Output: Related information retrieved from the database.

[0563] Step 5: Generate an answer

[0564] The server inputs a prompt sentence into a natural language generation model (for example, OpenAI's GPT-3) based on the acquired information, and generates an appropriate answer. Specifically, a prompt sentence such as "Please explain how to apply for paid leave" is input into the generative AI model, and the generated text is used as the answer. Input: Relevant information acquired from the database. Output: Generated answer text.

[0565] Step 6: Submit your response

[0566] The server sends the generated response to the terminal. Specifically, it includes the generated text as an HTTP response and sends it to the terminal. Input: Generated response text. Output: HTTP response to the terminal.

[0567] Step 7: Check your answers

[0568] The user checks the answer from the server displayed on the device. For example, the answer displayed is "Paid leave applications can be made through the in-house portal site." In concrete terms, the user looks at the device screen, reads the answer, and understands its contents. Input: Answer text received from the server. Output: User's understanding.

[0569] Step 8: Enter and submit your feedback

[0570] The user inputs feedback for the provided answer. For example, they rate it as "very helpful" or "I'd like more details." Specifically, the user clicks the feedback button, writes their rating, and clicks the submit button. Input: User feedback. Output: Feedback input to the server.

[0571] Step 9: Analyze feedback and learn

[0572] The server receives feedback from users and analyzes it to improve the quality of the AI ​​model. Specifically, the feedback data is stored in a database and evaluated by a machine learning algorithm. This adjusts the AI ​​model to generate a better answer for the next question. Input: User feedback. Output: Improvement and adjustment of the AI ​​model.

[0573] (Application example 1)

[0574] 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."

[0575] In logistics centers, it is important for new employees and workers to quickly obtain the information they need to perform their work in order to carry out their work efficiently. However, accessing complex work procedures and a wide range of inventory information requires advanced knowledge and experience, which places a heavy burden on new employees and first-time workers. A system that solves this issue and simplifies access to information is needed.

[0576] 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.

[0577] In this invention, the server includes means for analyzing instructions from a user using a natural language processing engine, means for searching for related information from an information storage unit based on the analyzed instructions, means for using a natural language generation algorithm to generate an answer based on the searched information, means for transmitting the generated answer to the user's display device, means for receiving user evaluations, means for improving the algorithm using the received evaluations, communication means for the user to request individual information via a wide area network, and dynamic search means for acquiring the requested individual information in real time, thereby enabling new employees and workers to quickly and accurately obtain the information they need.

[0578] A "natural language processing engine" is a software technology for analyzing text data and extracting keywords and meanings from within sentences.

[0579] "User instructions" refer to questions or commands that a user enters into the system.

[0580] The "information storage unit" refers to a database or storage system that stores related information.

[0581] "Related information" refers to the information to be searched for based on a user instruction.

[0582] A "natural language generation algorithm" is an algorithm that generates answers in natural language based on analyzed information.

[0583] A "display device" is a device for visually displaying information to a user.

[0584] "Rating" refers to the feedback or review a user gives to an answer provided by the system.

[0585] "Wide area network" refers to a wide-area communications network such as the Internet.

[0586] "Communication means" refers to the technology for sending user instructions to the server and for sending responses from the server to the user.

[0587] "Dynamic search means" refers to a technology that instantly searches for and retrieves requested information in real time.

[0588] This invention is a system that enables new employees and workers at a logistics center to quickly obtain information necessary for their work. This system operates in cooperation with three parties: a server, a terminal, and a user.

[0589] Server configuration and roles

[0590] The server runs a natural language processing engine and a natural language generation algorithm, and is connected to the information storage unit. The server's roles are as follows:

[0591] 1. Receiving and parsing instructions:

[0592] The server receives user instructions sent from the device and uses a natural language processing engine (e.g., spaCy) to analyze the instructions and extract relevant keywords and meanings.

[0593] 2. Information Search:

[0594] Based on the parsed instructions, the server searches for relevant information from an information storage unit (e.g., a MySQL database).

[0595] 3. Generate answers:

[0596] Based on the search results, a natural language generation algorithm (e.g., GPT-4) is used to generate an answer for the user.

[0597] 4. Submit your response:

[0598] The generated answer is sent to the terminal and displayed to the user.

[0599] 5. Receiving and learning from assessments:

[0600] It receives user ratings and uses them to improve its algorithms, which will generate better answers for the next prompt.

[0601] Device configuration and role

[0602] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[0603] 1. Receiving user input:

[0604] Users can enter instructions in natural language by simply typing in a text box and clicking the submit button, which sends the instructions to the server.

[0605] 2. Show Answer:

[0606] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[0607] 3. Enter your rating:

[0608] Users can provide a rating for an answer by clicking the rating button, which enters their feedback and sends it to the server.

[0609] User operations

[0610] A user follows the steps below to search for the information they need through the system.

[0611] 1. Enter instructions:

[0612] The user uses the terminal's interface to input instructions in natural language.

[0613] For example: "Please let me know the availability of the product."

[0614] 2. Check your answers:

[0615] The answer sent from the server is confirmed on the device screen.

[0616] 3. Providing Evaluations:

[0617] The answer is evaluated for appropriateness and the evaluation is sent to the server via the terminal.

[0618] Providing concrete examples and prompts

[0619] Specific examples

[0620] Consider the case where a new employee enters the following instructions at a logistics center:

[0621] "Please let me know the product availability."

[0622] The server uses a natural language processing engine (spaCy) to extract keywords such as "product" and "stock status," and searches for related information in the information storage unit (MySQL database).Then, it uses a natural language generation algorithm (GPT-4) to generate the following answer:

[0623] "We currently have 50 units of product A in stock. They are ready to ship."

[0624] Prompt Sentence Examples

[0625] The prompt has the following format:

[0626] "Provide a detailed answer for: Product A's stock status"

[0627] This system is designed to enable new employees and workers to quickly and accurately obtain the information they need, thereby improving operational efficiency at logistics centers.

[0628] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0629] Step 1:

[0630] The user uses the device interface to input instructions in natural language, for example, "Please tell me the product availability status," and clicks the submit button. This input is sent from the device to the server.

[0631] Step 2:

[0632] The server receives the user's instructions from the device and analyzes them using a natural language processing engine (e.g., spaCy). Specifically, it tokenizes the instruction text and extracts keywords and meanings. The user's instructions are given as input, and keywords are obtained as output.

[0633] Step 3:

[0634] The server searches for related information from the information storage unit (database example: MySQL) based on the analyzed keywords. For example, if "product inventory status" is extracted as a keyword, the corresponding inventory information is searched for in the database. The extracted keywords are given as input, and the search results (inventory status, etc.) are obtained as output.

[0635] Step 4:

[0636] The server generates an answer for the user using a natural language generation algorithm (e.g., GPT-4) based on the search results. Specifically, the search results are given in the form of a prompt sentence, and an answer in natural language is generated by the generative AI model. The search results are given as input, and the generated answer is obtained as output.

[0637] Step 5:

[0638] The server sends the generated answer to the terminal. The terminal displays the answer received from the server to the user. The generated answer is sent from the server as input, received by the terminal, and displayed to the user as output.

[0639] Step 6:

[0640] The user checks the displayed answers and rates them. The rating is sent from the device to the server. This rating is used to improve the algorithm. The user's rating is sent from the device to the server as input and reflected in the algorithm model as output.

[0641] Step 7:

[0642] The server uses the received evaluations to improve the natural language generation algorithm. Specifically, it adds the evaluation data as training data to improve the accuracy of the next answer generation. The evaluation data is given as input and an improved generation algorithm is obtained as output.

[0643] 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.

[0644] This invention is a system designed to enable new employees to efficiently access company information, and includes an emotion engine that recognizes the user's emotions and adjusts responses. This system operates in cooperation between a server, a terminal, and a user.

[0645] Server configuration and roles

[0646] The server runs a natural language processing engine, a natural language generation model, a database, and an emotion engine, and performs the following roles:

[0647] 1. Receiving and parsing questions:

[0648] The server receives the user's question sent from the device. Using a natural language processing engine, it analyzes the question and extracts appropriate keywords and meanings. For example, it may receive a question such as, "How do I apply for paid leave?"

[0649] 2. Emotion Recognition:

[0650] The emotion engine analyzes the user's emotions from the content of their question, for example, recognizing whether the user is feeling anxious or uncertain.

[0651] 3. Information Search:

[0652] Based on the analyzed keywords and emotion data, the server retrieves relevant information from a database.

[0653] 4. Generate and refine answers:

[0654] Based on the search results and sentiment data, a natural language generation model is used to generate a response for the user. Based on the sentiment engine data, the tone and content of the response are adjusted. For example, if the user is feeling anxious, a phrase like "Don't worry" is added to the response.

[0655] 5. Submit your response:

[0656] The generated answer is sent to the terminal and displayed to the user.

[0657] 6. Receive feedback and learn:

[0658] It receives user feedback and uses it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[0659] Device configuration and role

[0660] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[0661] 1. Receiving user input:

[0662] Users can enter questions in natural language by simply typing in the text box and clicking the submit button, which sends the question to the server.

[0663] 2. Show Answer:

[0664] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[0665] 3. Enter your feedback:

[0666] Users can provide feedback on the answer by clicking the feedback button, entering a rating and sending it to the server.

[0667] User operations

[0668] A user follows the steps below to search for the information they need through the system.

[0669] 1. Enter your question:

[0670] The user uses the terminal interface to enter a question in natural language.

[0671] 2. Check your answers:

[0672] The answer sent from the server is confirmed on the device screen.

[0673] 3. Providing Feedback:

[0674] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[0675] Specific examples

[0676] Example 1: How to apply for paid leave

[0677] User: Type into the device, "How do I request paid time off?"

[0678] Server: Receives the question and extracts the keywords "paid leave" and "how to apply" as well as the emotion "anxiety." It searches the database for relevant information and uses a natural language generation model to generate an answer such as, "Paid leave applications are made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu." This is then sent to the device.

[0679] Terminal: Displays the generated answer to the user.

[0680] Users: Rate the answers for accuracy and provide feedback.

[0681] Example 2: New employee training schedule

[0682] User: Type into terminal, "What is the new employee training schedule?"

[0683] Server: Receives the question and extracts the keywords "new employee training" and "schedule" as well as the sentiment of "doubt." It searches for related information in the database and uses a natural language generation model to generate an answer such as "The new employee training schedule is listed on the company calendar. Please check here." and sends it to the device.

[0684] Terminal: Displays the generated answer to the user.

[0685] Users: Rate the answers for accuracy and provide feedback.

[0686] In this way, the system not only enables new employees to quickly and accurately obtain the information they need, but also utilizes emotional data to provide more personalized assistance, resulting in improved work efficiency and a greater sense of security for new employees.

[0687] The processing flow will be explained below.

[0688] Step 1:

[0689] User: Enter a question in natural language through the device's user interface, such as "How do I apply for paid leave?", and click the submit button.

[0690] Step 2:

[0691] Terminal: The question text entered by the user is structured (for example, converted into JSON format) and sent to the server. Specifically, a program such as JavaScript sends the text data to the server via an AJAX request.

[0692] Step 3:

[0693] Server: Receives the question text sent from the terminal. Using a web framework such as Flask or Django, it receives an HTTP request and the parser reads the data in JSON format.

[0694] Step 4:

[0695] Server: The question text is analyzed using a natural language processing engine (e.g., spaCy, NLTK). Specifically, the text is tokenized and subjected to entity recognition and partial analysis to extract important keywords and phrases such as "paid leave" and "how to apply."

[0696] Step 5:

[0697] Server: Uses an emotion engine to analyze the emotions contained in the user's question text. For example, it identifies whether the user is feeling "anxiety" or "doubt."

[0698] Step 6:

[0699] Server: Based on the analyzed keywords and sentiment data, it searches for relevant information from a database (e.g., MySQL, PostgreSQL). It generates SQL queries and executes them against the database to retrieve relevant information.

[0700] Step 7:

[0701] Server: Based on the acquired data, a natural language generation model (e.g., GPT-3) is used to generate an answer. The generation process takes into account emotional data and optimizes the tone and content of the answer. For example, it generates a specific sentence such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu."

[0702] Step 8:

[0703] Server: Structure the generated answer (e.g., in JSON format) and send it to the device as an HTTP response, using Flask or Django's response object with the appropriate status code.

[0704] Step 9:

[0705] Terminal: Parses the received answer and displays it in the user interface. JavaScript parses the response data and inserts the answer text into an HTML element to display it visually to the user.

[0706] Step 10:

[0707] User: Review the displayed answers and rate them for appropriateness. Once you have finished rating, click the feedback button to provide your feedback.

[0708] Step 11:

[0709] Terminal: Structure the feedback entered by the user (convert it to JSON format) and send it to the server. Using JavaScript, send the feedback data to the server via an AJAX request.

[0710] Step 12:

[0711] Server: Stores the received feedback in a database. Generates an SQL query to insert into the feedback table in the database.

[0712] Step 13:

[0713] Server: Retrains the NLP engine, NLG model, and sentiment engine based on the stored feedback. Each model learns using new feedback data to improve its performance.

[0714] These are the specific processing steps of the new employee support generation AI system that combines an emotion engine. This system allows new employees to efficiently obtain internal company information and receive personalized support based on their emotions.

[0715] Example 2

[0716] 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."

[0717] When new employees start work, it is urgent to provide them with a means to efficiently access internal company information. However, conventional systems provide mechanical responses without considering the user's feelings, making it difficult to improve the user's psychological sense of security and satisfaction. Furthermore, the system lacks the ability to adaptively improve itself based on feedback, making it difficult to provide information that meets the user's needs.

[0718] 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.

[0719] In this invention, the server includes: means for analyzing user questions using a natural language processing engine; means for analyzing the user's emotions using the analyzed question content and an emotion recognition engine; means for searching for related information from a database based on the analyzed question content and emotion data; means for generating answers based on the searched information and using a generative AI model that adjusts the content according to the emotion; means for sending the generated answers to the user's device; means for receiving user feedback; and means for improving the model using the received feedback. This not only enables new employees to quickly and accurately access the information they need, but also provides personalized answers based on emotion data. Furthermore, by improving the system using feedback, user satisfaction and reliability can be continuously improved.

[0720] A "natural language processing engine" is a software technology that analyzes text entered by a user and understands its content and structure.

[0721] An "emotion recognition engine" is a technology that analyzes emotions from a user's text and identifies emotional states such as anxiety, anger, and joy.

[0722] A "database" is a data management system for storing related information in a structured way that makes it easy to search, store, and update.

[0723] A "generative AI model" is a machine learning model that uses natural language data as input and automatically generates appropriate answers based on the context.

[0724] A "user terminal" is an electronic device that a user uses to access the system, enter questions, and view answers.

[0725] "Feedback" refers to the user's evaluation of the system's answers and suggestions for improvement.

[0726] "Model improvement" is the process of correcting and updating the system's learning data and algorithms based on the feedback received, thereby improving answer accuracy and user satisfaction.

[0727] This invention is a system designed to enable new employees to efficiently access company information, and includes an emotion recognition engine to recognize the user's emotions and adjust responses accordingly. This system operates in cooperation between a server, a terminal, and a user.

[0728] Server configuration and roles

[0729] The following main software modules run on the server:

[0730] 1. Natural Language Processing Engines (e.g. SpaCy, NLTK):

[0731] It is used to analyze user questions and extract keywords and context.

[0732] 2. Emotion Recognition Engine (e.g. Hugging Face Transformers, IBM Watson):

[0733] Analyzes emotions from user text and recognizes feelings such as anxiety and doubt.

[0734] 3. Database (e.g. MySQL, PostgreSQL):

[0735] Store and search relevant information within your company.

[0736] 4. Generative AI models (e.g., OpenAI GPT-3, BERT):

[0737] Generate appropriate answers for users based on search results and sentiment data.

[0738] The server's operating process is as follows:

[0739] 1. Receiving and parsing questions:

[0740] The server receives the user's question sent from the device and analyzes it using a natural language processing engine. For example, it may receive a question such as "How do I apply for paid vacation?"

[0741] 2. Emotion Recognition:

[0742] An emotion recognition engine is used to analyze the emotions of the user based on the content of the question. For example, it can recognize that the user is feeling anxious.

[0743] 3. Information Search:

[0744] Based on the analyzed keywords and emotion data, the server searches the database for relevant information, for example, identifying information on "paid vacation" and "how to apply."

[0745] 4. Generate and refine answers:

[0746] Based on the search results and sentiment data, a generative AI model is used to generate a response for the user. The tone and content of the response are adjusted based on the sentiment data. For example, a response such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu" can be generated.

[0747] 5. Submit your response:

[0748] The generated response is sent to the device.

[0749] 6. Receive feedback and learn:

[0750] It receives user feedback and uses it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[0751] Device configuration and role

[0752] The terminal provides an intuitive and easy-to-use interface for users to access the system, which:

[0753] 1. Receiving user input:

[0754] Allow users to enter their question in natural language, for example by typing their question in a text box and clicking the submit button.

[0755] 2. Show Answer:

[0756] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[0757] 3. Enter your feedback:

[0758] Users can provide feedback on the answers, for example by clicking a feedback button and entering a rating, which is then sent to the server.

[0759] User operations

[0760] A user follows the steps below to search for the information they need through the system.

[0761] 1. Enter your question:

[0762] Users use the device interface to enter questions in natural language.

[0763] 2. Check your answers:

[0764] The answer sent from the server is confirmed on the device screen.

[0765] 3. Providing Feedback:

[0766] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[0767] Specific examples

[0768] Example 1: How to apply for paid leave

[0769] User: Type "How do I request paid time off?" into the device.

[0770] Server: Receives the question and extracts the keywords "paid leave" and "how to apply" as well as the emotion "anxiety." It searches the database for relevant information and uses a generative AI model to generate an answer such as, "Paid leave applications are made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu." This is then sent to the device.

[0771] Terminal: Displays the generated answer to the user.

[0772] Users: Rate the answers for accuracy and provide feedback.

[0773] Example 2: New employee training schedule

[0774] User: Type into terminal, "What is the new employee training schedule?"

[0775] Server: Receives the question and extracts the keywords "new employee training" and "schedule" as well as the emotion "doubt." Searches for related information in the database and uses a generative AI model to generate an answer such as "The new employee training schedule is listed on the company calendar. Please check here." and sends it to the device.

[0776] Terminal: Displays the generated answer to the user.

[0777] Users: Rate the answers for accuracy and provide feedback.

[0778] In this way, the system not only enables new employees to quickly and accurately obtain the information they need, but also utilizes emotional data to provide more personalized support, resulting in improved work efficiency and a greater sense of security for new employees.

[0779] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0780] Step 1:

[0781] The user inputs a question into the device interface. For example, the user inputs "How do I apply for paid leave?" into the text box. This operation saves the question in text format on the device.

[0782] Step 2:

[0783] When the user clicks the submit button, the device sends the question to the server as an HTTP POST request, which includes the question text.

[0784] Step 3:

[0785] The server receives an HTTP POST request from the device. It parses the question text and extracts keywords and contextual information using a natural language processing engine (e.g., SpaCy, NLTK). This process extracts keywords such as "paid vacation" and "how to apply" from the question text.

[0786] Step 4:

[0787] The server analyzes the user's emotions using an emotion recognition engine (e.g., Hugging Face Transformers, IBM Watson) based on the extracted keywords. For example, it detects the user's anxiety from the question text. This process generates emotion data such as "anxiety."

[0788] Step 5:

[0789] The server uses the keywords and sentiment data to search for relevant information from a database (e.g., MySQL, PostgreSQL). For example, it searches for company policies and procedures related to "paid leave" and "how to apply." This process retrieves the relevant information in text format.

[0790] Step 6:

[0791] The server generates an answer using a generative AI model (e.g., OpenAI GPT-3, BERT) based on the search results and emotion data. The generated prompt includes keywords, emotion data, and search results. This process generates an answer text such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu."

[0792] Step 7:

[0793] The server sends the generated answer text to the terminal as an HTTP response, which includes the answer text.

[0794] Step 8:

[0795] The terminal displays the answer text received from the server to the user, for example, by displaying the answer in a text area on the screen.

[0796] Step 9:

[0797] The user inputs feedback on the displayed answer, for example, by clicking a feedback button and inputting an evaluation such as "appropriate" or "inappropriate."

[0798] Step 10:

[0799] The device sends the user-provided feedback to the server as an HTTP POST request.

[0800] Step 11:

[0801] The server receives feedback sent from the device and analyzes the feedback data. Based on this data, it evaluates the performance of the AI ​​model and adjusts the model or updates the training data as needed, thereby improving the accuracy and reliability of the system.

[0802] (Application example 2)

[0803] 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."

[0804] In modern factories, new employees and technicians need a large amount of information and quick access to it in order to adapt to complex machines and processes. However, with previous systems, obtaining information and troubleshooting took time, often causing anxiety and stress. Therefore, an information delivery system that can provide efficient and emotional support is needed.

[0805] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0806] In this invention, the server includes a means for analyzing a user's question using a natural language processing engine, a means for searching for related information from a database based on the analyzed question, and a means for using a natural language generation model to generate an answer based on the searched information. This allows the user to receive an answer tailored to their own emotions. Furthermore, by providing an interface accessible from the robot's display or smart device and adding a function for learning from the user's question history and feedback to generate better answers, the server achieves efficient and stress-free information acquisition.

[0807] A "natural language processing engine" is an engine that analyzes questions entered by users and extracts keywords and context.

[0808] A "database" is a collection of information that stores information in response to user questions and manages it in a searchable format.

[0809] A "natural language generation model" is an algorithm for generating appropriate answers to users based on analyzed questions.

[0810] The "emotion engine" is an engine that analyzes emotions from the user's question and input data and adjusts the response content.

[0811] A "terminal" is a device with an interface that allows a user to access the system and input questions.

[0812] An "interface" is an operation screen or application that allows a user to input questions into the system and receive answers.

[0813] "Feedback" refers to the evaluations and opinions that users provide regarding the system's answers, and is used to improve the system.

[0814] A "robot display" is a device installed on a factory robot that displays operation information and responses from the system.

[0815] A "smart device" is a portable electronic device that can connect to the Internet, such as a smartphone or tablet.

[0816] "Question history" is a record of questions and answers that a user has previously entered into the system.

[0817] "Troubleshooting" is the process of diagnosing errors or problems in a system or machine and providing a solution.

[0818] To implement this invention, an automated information acquisition system must be installed on factory robots and related smart devices. This system works in cooperation with a server, terminals, and users.

[0819] Server configuration and roles

[0820] The server runs a natural language processing engine, a natural language generation model, a database, and an emotion engine. Details are as follows:

[0821] 1. Receiving and analyzing the question: The server receives the user's question sent from the terminal. Using a natural language processing engine, the server analyzes the question and extracts appropriate keywords and meanings. For example, the server may receive a question such as, "Please tell me how to deal with error code E101."

[0822] 2. Emotion Recognition: Using an emotion engine, we analyze the emotions in the user's questions, for example, whether the user is feeling anxious or uncertain.

[0823] 3. Information retrieval: Based on the analyzed keywords and sentiment data, the server retrieves relevant information from the database.

[0824] 4. Answer generation and tailoring: Based on the search results and sentiment data, a natural language generation model is used to generate an answer for the user. The tone and content of the answer are adjusted based on the sentiment engine data. For example, if the user is feeling anxious, a phrase like "Don't worry" is added to the answer.

[0825] 5. Sending the answer: The generated answer is sent to the terminal and displayed to the user.

[0826] 6. Receive and learn feedback: Receive user feedback and use it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[0827] Device configuration and role

[0828] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[0829] 1. Receiving user input: Users can input questions in natural language. Questions can be posed via text or voice using the factory robot's display or smart device.

[0830] 2. Displaying the answer: The answer received from the server is displayed to the user, allowing the user to quickly obtain the required information.

[0831] 3. Enter feedback: Users can provide feedback on the answer by clicking the feedback button, entering their rating and sending it to the server.

[0832] User operations

[0833] A user follows the steps below to search for the information they need through the system.

[0834] 1. Entering a question: Users enter a question in natural language using the factory robot's display or a smart device.

[0835] 2. Confirm the answer: Confirm the answer sent from the server on the device screen.

[0836] 3. Providing feedback: Evaluate the appropriateness of the answer and send feedback to the server via the device.

[0837] Specific examples

[0838] Example 1: How to resolve a robot error code

[0839] User: Enter "Please tell me how to deal with error code E101" into the terminal.

[0840] Server: Receives the question and extracts the keywords "error code E101" and "how to deal with it," as well as the emotion "anxiety." It searches for related information in a database and uses a natural language generation model to generate a response such as "Error code E101 indicates a sensor malfunction. Don't worry, try reconnecting the sensor," and sends it to the device.

[0841] Terminal: Displays the generated answer to the user.

[0842] User: Check the solution and perform the task.

[0843] Prompt Sentence Examples

[0844] "How do I resolve the error code E101?"

[0845] This system enables new employees and technicians to quickly address any problems that arise within the factory, improving work efficiency and safety.

[0846] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0847] Processing Steps

[0848] Step 1:

[0849] The user inputs a question in natural language using the factory robot's display or a smart device. For example, the user might input a question such as, "Please tell me how to deal with error code E101." The user's input is sent to the terminal as text data.

[0850] Step 2:

[0851] The device sends text data from the user to the server. Specifically, the device transfers the user's input data to the server as an API request.

[0852] Step 3:

[0853] The server analyzes the user's question received from the terminal using a natural language processing engine. As a result of the analysis, the keywords "error code E101" and "solution" are extracted. In this analysis process, the input text data is broken down into tokens and semantic analysis is performed.

[0854] Step 4:

[0855] The server uses an emotion engine to recognize emotions from the user's question. For example, the emotion "anxiety" is extracted. In this step, an emotion analysis algorithm is applied to the user's input data to generate emotion data.

[0856] Step 5:

[0857] The server searches for relevant information from the database based on the analyzed keywords and emotion data. For example, it retrieves information about "error code E101" in the database. In this search process, an SQL or NoSQL query is constructed to retrieve the corresponding record from the database.

[0858] Step 6:

[0859] The server generates a response to the user using a natural language generation model based on the search results and sentiment data. For example, a response such as "Error code E101 indicates a sensor malfunction. Please rest assured and try reconnecting the sensor" is generated. In this generation process, a natural language generation algorithm is applied to construct the response text.

[0860] Step 7:

[0861] The server sends the generated response to the device. Specifically, it returns the response data to the device as an API response. In this sending process, the generated text data is packaged in JSON format and sent as an HTTP response.

[0862] Step 8:

[0863] The device displays the answer received from the server to the user. Specifically, the generated answer is displayed on the display. In this display process, the received text data is drawn on the UI.

[0864] Step 9:

[0865] The user evaluates whether the answer is appropriate and sends feedback to the server via the terminal. For example, the user may enter feedback such as "This answer was helpful." In this step, the feedback data is sent to the terminal as text.

[0866] Step 10:

[0867] The server receives feedback from users and uses it to improve the AI ​​model. This improvement process uses the received feedback data as training data to improve the accuracy of the natural language processing engine and natural language generation model.

[0868] These processing steps provide fast and appropriate answers to user questions while continually improving the accuracy and reliability of the overall system.

[0869] 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.

[0870] 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.

[0871] 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.

[0872] [Third embodiment]

[0873] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0874] 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.

[0875] 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).

[0876] 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.

[0877] 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.

[0878] 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).

[0879] 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. 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.

[0880] 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.

[0881] 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.

[0882] 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.

[0883] 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.

[0884] 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."

[0885] This system is designed to enable new employees to efficiently access company information. It works in cooperation with three parties: the server, the terminal, and the user.

[0886] Server configuration and roles

[0887] The server runs a natural language processing engine and a natural language generation model and is connected to a database. This allows the server to fulfill the following roles:

[0888] 1. Receiving and parsing questions:

[0889] The server receives the user's question sent from the device and uses a natural language processing engine to analyze the question and extract appropriate keywords and meanings.

[0890] 2. Information Search:

[0891] Based on the parsed question, the server retrieves relevant information from a database.

[0892] 3. Generate answers:

[0893] Based on the search results, a natural language generation model is used to generate an answer for the user.

[0894] 4. Submit your response:

[0895] The generated answer is sent to the terminal and displayed to the user.

[0896] 5. Receive feedback and learn:

[0897] It receives user feedback and uses it to improve the AI ​​model, which will generate a better answer the next time you ask a question.

[0898] Device configuration and role

[0899] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[0900] 1. Receiving user input:

[0901] Users can enter questions in natural language by simply typing in the text box and clicking the submit button, which sends the question to the server.

[0902] 2. Show Answer:

[0903] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[0904] 3. Enter your feedback:

[0905] Users can provide feedback on the answer by clicking the feedback button, entering a rating and sending it to the server.

[0906] User operations

[0907] A user follows the steps below to search for the information they need through the system.

[0908] 1. Enter your question:

[0909] The user uses the terminal interface to enter a question in natural language.

[0910] 2. Check your answers:

[0911] The answer sent from the server is confirmed on the device screen.

[0912] 3. Providing Feedback:

[0913] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[0914] Specific examples

[0915] Example 1: How to apply for paid leave

[0916] User: Type into the device, "How do I request paid time off?"

[0917] Server: Receives the question, extracts keywords related to "paid leave" and "how to apply," and searches the database for related information. For example, it generates an answer such as, "Paid leave applications are made through the in-house portal site. Log in to the portal site and apply from the 'Paid Leave Application' menu." and sends it to the device.

[0918] Terminal: Displays the generated answer to the user.

[0919] Users: Rate the answers for accuracy and provide feedback.

[0920] Example 2: New employee training schedule

[0921] User: Type into terminal, "What is the new employee training schedule?"

[0922] Server: Receives the question, extracts the keywords "new employee training" and "schedule," and searches the database for related information. For example, it generates an answer such as "The new employee training schedule is listed on the company calendar. Please check here," and sends it to the device.

[0923] Terminal: Displays the generated answer to the user.

[0924] Users: Rate the answers for accuracy and provide feedback.

[0925] In this way, the system is designed to enable new employees to quickly and accurately obtain the information they need, thereby improving work efficiency.

[0926] The processing flow will be explained below.

[0927] Step 1:

[0928] User: Enter a question in natural language through the device's user interface. Enter a question such as "How do I apply for paid leave?" into the input form and click the submit button.

[0929] Step 2:

[0930] Terminal: The question text entered by the user is structured (for example, converted into JSON format) and sent to the server. Specifically, a program such as JavaScript sends the text data to the server via an AJAX request.

[0931] Step 3:

[0932] Server: Receives the question text sent from the terminal. Specifically, it receives an HTTP request using a web framework such as Flask or Django, and the parser reads the data in JSON format.

[0933] Step 4:

[0934] Server: The question text is analyzed using a natural language processing engine (e.g., spaCy, NLTK). During this analysis process, the text is tokenized and subjected to entity recognition and partial analysis to extract important keywords and phrases such as "paid leave" and "how to apply."

[0935] Step 5:

[0936] Server: Based on the analyzed keywords, it searches for relevant information from a database (e.g., MySQL, PostgreSQL). It generates SQL queries and executes them against the database to retrieve the relevant information.

[0937] Step 6:

[0938] Server: Based on the acquired data, a natural language generation model (e.g., GPT-3) is used to generate an answer. For example, it generates a specific sentence such as, "Paid leave applications can be made through the in-house portal site. Please log in to the portal site and apply from the 'Paid Leave Application' menu."

[0939] Step 7:

[0940] Server: Structure the generated answer (e.g., in JSON format) and send it to the device as an HTTP response, using the Flask or Django response object with the appropriate status code.

[0941] Step 8:

[0942] Terminal: Parses the received answer and displays it in the user interface. Specifically, JavaScript parses the response data and inserts the answer text into an HTML element to visually display it to the user.

[0943] Step 9:

[0944] User: Review the displayed answers and rate them for appropriateness. Once you have finished rating, click the feedback button to provide your feedback.

[0945] Step 10:

[0946] Terminal: Structure the feedback entered by the user (convert it to JSON format) and send it to the server. Again using JavaScript, send the feedback data to the server via an AJAX request.

[0947] Step 11:

[0948] Server: Stores the received feedback in a database. Generates an SQL query to insert into the feedback table in the database.

[0949] Step 12:

[0950] Server: Retrains the NLP engine or NLG model based on the stored feedback, learning to improve the model's performance with new feedback data.

[0951] These are the specific processing steps of the new employee support generation AI system. This system allows new employees to efficiently obtain internal company information and improve work efficiency.

[0952] Example 1

[0953] 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."

[0954] In today's corporate environment, there is a demand for systems that allow new employees to efficiently access internal information and quickly obtain the information they need. Conventional systems have problems such as complicated information search and retrieval, making it difficult for users to operate intuitively, and reducing work efficiency. It has also been difficult to effectively utilize user feedback to improve the system's response accuracy.

[0955] 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.

[0956] In this invention, the server includes: means for analyzing a question from a user using a natural language processing engine; means for searching for related information from an information storage device based on the analyzed question content; means for using a natural language generation model to generate an answer based on the searched information; means for transmitting the generated answer to the user's information display device; means for receiving user evaluation information; means for improving the model using the received evaluation information; means for analyzing the question entered by the user and generating an answer by inputting a prompt sentence into the generation AI model based on the analysis result; means for displaying the generated answer; and means for analyzing the evaluation information provided by the user and reflecting it in the next answer generation. This allows new employees to access internal company information efficiently and intuitively. Furthermore, the accuracy of answer generation can be improved based on user feedback.

[0957] A "natural language processing engine" is a software technology that analyzes questions from users and extracts keywords and meanings.

[0958] A "natural language generation model" is an algorithm for generating appropriate answers in natural language based on the analyzed question content.

[0959] An "information storage device" is hardware or software for storing and managing information, such as a database or file system.

[0960] An "information display device" is a screen, monitor, or similar device for displaying information to a user.

[0961] "Rating Information" refers to feedback and ratings provided by users, and is data used to improve the system's response accuracy.

[0962] A "prompt" is a statement of instructions or a question that is input into a generative AI model to generate a specific answer.

[0963] "Analysis" is the process of understanding a user's question, extracting meaning, and identifying keywords.

[0964] "Generation" is the process of generating new, appropriate answers or text based on the retrieved information.

[0965] "Sending" is the process of sending a response or data from the server to the terminal over a communication path.

[0966] "Learning" refers to the application of machine learning algorithms to improve the system's response accuracy based on user feedback.

[0967] MODE FOR CARRYING OUT THE INVENTION

[0968] The system of the present invention is designed to enable new employees to efficiently access company information and quickly obtain the information they need. It uses a natural language processing engine, a natural language generation model, an information storage device, and an information display device in cooperation with a server, a terminal, and a user.

[0969] First, the user inputs a question into the terminal. The terminal allows the user to input questions in natural language through a user interface. This interface includes a text box and a submit button, and is designed to be intuitive. For example, if the user inputs "How do I apply for paid vacation?", the question will be submitted.

[0970] Next, the device sends the user's question to the server. The server receives the question and analyzes it using a natural language processing engine. Software such as spaCy or NLTK can be used as the natural language processing engine. The question is analyzed to extract keywords such as "paid leave" and "how to apply."

[0971] Based on the parsed query, the server searches for relevant information using an information storage device, such as MySQL or PostgreSQL, and executes a database query to retrieve the appropriate information.

[0972] Next, the server uses a natural language generation model to generate an answer based on the search results. OpenAI's GPT-3 and other generative AI models can be applied. As a specific example, the prompt "Please explain how to apply for paid leave" is entered, and an appropriate answer is generated. The generated answer is something like, "Paid leave applications are made through the in-house portal site. Please log in to the portal site and apply from the 'Paid Leave Application' menu."

[0973] The generated answer is sent from the server to the terminal and displayed to the user, allowing the user to quickly obtain the information they need. The user can also enter evaluation information for the provided answer, i.e., feedback. Feedback can include evaluations such as "It was helpful" or "I'd like more details."

[0974] Finally, the server receives user feedback and analyzes the ratings to improve the natural language generation model, using machine learning algorithms to adjust the model to generate better answers for the next question.

[0975] For example, if the user's evaluation feedback is that the answer was not specific, the server will adjust the model parameters to provide more detailed information the next time it responds, thereby improving the response quality of the entire system.

[0976] This system allows new employees to efficiently access internal company information and quickly obtain the information they need. Furthermore, response accuracy can be improved based on user feedback, allowing for continuous improvement.

[0977] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0978] Step 1: Enter your question

[0979] The user enters a question into the text box on the terminal. For example, the user enters "How do I apply for paid leave?". Specifically, the user enters the question using the keyboard and clicks the send button. Input: User's question. Output: Question data in text format.

[0980] Step 2: Submit your question

[0981] The device sends the question entered by the user to the server. Specifically, when the send button is clicked, the device creates an HTTP request and sends the question to the server's API endpoint. Input: Question data in text format. Output: HTTP request to the server.

[0982] Step 3: Analyzing the Question

[0983] The server analyzes the question received from the device. A natural language processing engine (e.g., spaCy or NLTK) is used to tokenize the question and extract key keywords (e.g., "paid leave" and "how to apply"). Specifically, the question is tokenized and tagged with parts of speech to identify important keywords. Input: Text question data extracted from the HTTP request. Output: Extracted keywords and their analysis results.

[0984] Step 4: Finding information

[0985] The server searches for related information from an information storage device (e.g., MySQL or PostgreSQL) based on the analysis results. Specifically, the server generates an SQL query to retrieve relevant records from the database. Input: Parsed keywords. Output: Related information retrieved from the database.

[0986] Step 5: Generate an answer

[0987] The server inputs a prompt sentence into a natural language generation model (for example, OpenAI's GPT-3) based on the acquired information, and generates an appropriate answer. Specifically, a prompt sentence such as "Please explain how to apply for paid leave" is input into the generative AI model, and the generated text is used as the answer. Input: Relevant information acquired from the database. Output: Generated answer text.

[0988] Step 6: Submit your response

[0989] The server sends the generated response to the terminal. Specifically, it includes the generated text as an HTTP response and sends it to the terminal. Input: Generated response text. Output: HTTP response to the terminal.

[0990] Step 7: Check your answers

[0991] The user checks the answer from the server displayed on the device. For example, the answer displayed is "Paid leave applications can be made through the in-house portal site." In concrete terms, the user looks at the device screen, reads the answer, and understands its contents. Input: Answer text received from the server. Output: User's understanding.

[0992] Step 8: Enter and submit your feedback

[0993] The user inputs feedback for the provided answer. For example, they rate it as "very helpful" or "I'd like more details." Specifically, the user clicks the feedback button, writes their rating, and clicks the submit button. Input: User feedback. Output: Feedback input to the server.

[0994] Step 9: Analyze feedback and learn

[0995] The server receives feedback from users and analyzes it to improve the quality of the AI ​​model. Specifically, the feedback data is stored in a database and evaluated by a machine learning algorithm. This adjusts the AI ​​model to generate a better answer for the next question. Input: User feedback. Output: Improvement and adjustment of the AI ​​model.

[0996] (Application example 1)

[0997] 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."

[0998] In logistics centers, it is important for new employees and workers to quickly obtain the information they need to perform their work in order to carry out their work efficiently. However, accessing complex work procedures and a wide range of inventory information requires advanced knowledge and experience, which places a heavy burden on new employees and first-time workers. A system that solves this issue and simplifies access to information is needed.

[0999] 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.

[1000] In this invention, the server includes means for analyzing instructions from a user using a natural language processing engine, means for searching for related information from an information storage unit based on the analyzed instructions, means for using a natural language generation algorithm to generate an answer based on the searched information, means for transmitting the generated answer to the user's display device, means for receiving user evaluations, means for improving the algorithm using the received evaluations, communication means for the user to request individual information via a wide area network, and dynamic search means for acquiring the requested individual information in real time, thereby enabling new employees and workers to quickly and accurately obtain the information they need.

[1001] A "natural language processing engine" is a software technology for analyzing text data and extracting keywords and meanings from within sentences.

[1002] "User instructions" refer to questions or commands that a user enters into the system.

[1003] The "information storage unit" refers to a database or storage system that stores related information.

[1004] "Related information" refers to the information to be searched for based on a user instruction.

[1005] A "natural language generation algorithm" is an algorithm that generates answers in natural language based on analyzed information.

[1006] A "display device" is a device for visually displaying information to a user.

[1007] "Rating" refers to the feedback or review a user gives to an answer provided by the system.

[1008] "Wide area network" refers to a wide-area communications network such as the Internet.

[1009] "Communication means" refers to the technology for sending user instructions to the server and for sending responses from the server to the user.

[1010] "Dynamic search means" refers to a technology that instantly searches for and retrieves requested information in real time.

[1011] This invention is a system that enables new employees and workers at a logistics center to quickly obtain information necessary for their work. This system operates in cooperation with three parties: a server, a terminal, and a user.

[1012] Server configuration and roles

[1013] The server runs a natural language processing engine and a natural language generation algorithm, and is connected to the information storage unit. The server's roles are as follows:

[1014] 1. Receiving and parsing instructions:

[1015] The server receives user instructions sent from the device and uses a natural language processing engine (e.g., spaCy) to analyze the instructions and extract relevant keywords and meanings.

[1016] 2. Information Search:

[1017] Based on the parsed instructions, the server searches for relevant information from an information storage unit (e.g., a MySQL database).

[1018] 3. Generate answers:

[1019] Based on the search results, a natural language generation algorithm (e.g., GPT-4) is used to generate an answer for the user.

[1020] 4. Submit your response:

[1021] The generated answer is sent to the terminal and displayed to the user.

[1022] 5. Receiving and learning from assessments:

[1023] It receives user ratings and uses them to improve its algorithms, which will generate better answers for the next prompt.

[1024] Device configuration and role

[1025] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[1026] 1. Receiving user input:

[1027] Users can enter instructions in natural language by simply typing in a text box and clicking the submit button, which sends the instructions to the server.

[1028] 2. Show Answer:

[1029] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[1030] 3. Enter your rating:

[1031] Users can provide a rating for an answer by clicking the rating button, which enters their feedback and sends it to the server.

[1032] User operations

[1033] A user follows the steps below to search for the information they need through the system.

[1034] 1. Enter instructions:

[1035] The user uses the terminal's interface to input instructions in natural language.

[1036] For example: "Please let me know the availability of the product."

[1037] 2. Check your answers:

[1038] The answer sent from the server is confirmed on the device screen.

[1039] 3. Providing Evaluations:

[1040] The answer is evaluated for appropriateness and the evaluation is sent to the server via the terminal.

[1041] Providing concrete examples and prompts

[1042] Specific examples

[1043] Consider the case where a new employee enters the following instructions at a logistics center:

[1044] "Please let me know the product availability."

[1045] The server uses a natural language processing engine (spaCy) to extract keywords such as "product" and "stock status," and searches for related information in the information storage unit (MySQL database).Then, it uses a natural language generation algorithm (GPT-4) to generate the following answer:

[1046] "We currently have 50 units of product A in stock. They are ready to ship."

[1047] Prompt Sentence Examples

[1048] The prompt has the following format:

[1049] "Provide a detailed answer for: Product A's stock status"

[1050] This system is designed to enable new employees and workers to quickly and accurately obtain the information they need, thereby improving operational efficiency at logistics centers.

[1051] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1052] Step 1:

[1053] The user uses the device interface to input instructions in natural language, for example, "Please tell me the product availability status," and clicks the submit button. This input is sent from the device to the server.

[1054] Step 2:

[1055] The server receives the user's instructions from the device and analyzes them using a natural language processing engine (e.g., spaCy). Specifically, it tokenizes the instruction text and extracts keywords and meanings. The user's instructions are given as input, and keywords are obtained as output.

[1056] Step 3:

[1057] The server searches for related information from the information storage unit (database example: MySQL) based on the analyzed keywords. For example, if "product inventory status" is extracted as a keyword, the corresponding inventory information is searched for in the database. The extracted keywords are given as input, and the search results (inventory status, etc.) are obtained as output.

[1058] Step 4:

[1059] The server generates an answer for the user using a natural language generation algorithm (e.g., GPT-4) based on the search results. Specifically, the search results are given in the form of a prompt sentence, and an answer in natural language is generated by the generative AI model. The search results are given as input, and the generated answer is obtained as output.

[1060] Step 5:

[1061] The server sends the generated answer to the terminal. The terminal displays the answer received from the server to the user. The generated answer is sent from the server as input, received by the terminal, and displayed to the user as output.

[1062] Step 6:

[1063] The user checks the displayed answers and rates them. The rating is sent from the device to the server. This rating is used to improve the algorithm. The user's rating is sent from the device to the server as input and reflected in the algorithm model as output.

[1064] Step 7:

[1065] The server uses the received evaluations to improve the natural language generation algorithm. Specifically, it adds the evaluation data as training data to improve the accuracy of the next answer generation. The evaluation data is given as input and an improved generation algorithm is obtained as output.

[1066] 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.

[1067] This invention is a system designed to enable new employees to efficiently access company information, and includes an emotion engine that recognizes the user's emotions and adjusts responses. This system operates in cooperation between a server, a terminal, and a user.

[1068] Server configuration and roles

[1069] The server runs a natural language processing engine, a natural language generation model, a database, and an emotion engine, and performs the following roles:

[1070] 1. Receiving and parsing questions:

[1071] The server receives the user's question sent from the device. Using a natural language processing engine, it analyzes the question and extracts appropriate keywords and meanings. For example, it may receive a question such as, "How do I apply for paid leave?"

[1072] 2. Emotion Recognition:

[1073] The emotion engine analyzes the user's emotions from the content of their question, for example, recognizing whether the user is feeling anxious or uncertain.

[1074] 3. Information Search:

[1075] Based on the analyzed keywords and emotion data, the server retrieves relevant information from a database.

[1076] 4. Generate and refine answers:

[1077] Based on the search results and sentiment data, a natural language generation model is used to generate a response for the user. Based on the sentiment engine data, the tone and content of the response are adjusted. For example, if the user is feeling anxious, a phrase like "Don't worry" is added to the response.

[1078] 5. Submit your response:

[1079] The generated answer is sent to the terminal and displayed to the user.

[1080] 6. Receive feedback and learn:

[1081] It receives user feedback and uses it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[1082] Device configuration and role

[1083] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[1084] 1. Receiving user input:

[1085] Users can enter questions in natural language by simply typing in the text box and clicking the submit button, which sends the question to the server.

[1086] 2. Show Answer:

[1087] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[1088] 3. Enter your feedback:

[1089] Users can provide feedback on the answer by clicking the feedback button, entering a rating and sending it to the server.

[1090] User operations

[1091] A user follows the steps below to search for the information they need through the system.

[1092] 1. Enter your question:

[1093] The user uses the terminal interface to enter a question in natural language.

[1094] 2. Check your answers:

[1095] The answer sent from the server is confirmed on the device screen.

[1096] 3. Providing Feedback:

[1097] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[1098] Specific examples

[1099] Example 1: How to apply for paid leave

[1100] User: Type into the device, "How do I request paid time off?"

[1101] Server: Receives the question and extracts the keywords "paid leave" and "how to apply" as well as the emotion "anxiety." It searches the database for relevant information and uses a natural language generation model to generate an answer such as, "Paid leave applications are made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu." This is then sent to the device.

[1102] Terminal: Displays the generated answer to the user.

[1103] Users: Rate the answers for accuracy and provide feedback.

[1104] Example 2: New employee training schedule

[1105] User: Type into terminal, "What is the new employee training schedule?"

[1106] Server: Receives the question and extracts the keywords "new employee training" and "schedule" as well as the sentiment of "doubt." It searches for related information in the database and uses a natural language generation model to generate an answer such as "The new employee training schedule is listed on the company calendar. Please check here." and sends it to the device.

[1107] Terminal: Displays the generated answer to the user.

[1108] Users: Rate the answers for accuracy and provide feedback.

[1109] In this way, the system not only enables new employees to quickly and accurately obtain the information they need, but also utilizes emotional data to provide more personalized assistance, resulting in improved work efficiency and a greater sense of security for new employees.

[1110] The processing flow will be explained below.

[1111] Step 1:

[1112] User: Enter a question in natural language through the device's user interface, such as "How do I apply for paid leave?", and click the submit button.

[1113] Step 2:

[1114] Terminal: The question text entered by the user is structured (for example, converted into JSON format) and sent to the server. Specifically, a program such as JavaScript sends the text data to the server via an AJAX request.

[1115] Step 3:

[1116] Server: Receives the question text sent from the terminal. Using a web framework such as Flask or Django, it receives an HTTP request and the parser reads the data in JSON format.

[1117] Step 4:

[1118] Server: The question text is analyzed using a natural language processing engine (e.g., spaCy, NLTK). Specifically, the text is tokenized and subjected to entity recognition and partial analysis to extract important keywords and phrases such as "paid leave" and "how to apply."

[1119] Step 5:

[1120] Server: Uses an emotion engine to analyze the emotions contained in the user's question text. For example, it identifies whether the user is feeling "anxiety" or "doubt."

[1121] Step 6:

[1122] Server: Based on the analyzed keywords and sentiment data, it searches for relevant information from a database (e.g., MySQL, PostgreSQL). It generates SQL queries and executes them against the database to retrieve relevant information.

[1123] Step 7:

[1124] Server: Based on the acquired data, a natural language generation model (e.g., GPT-3) is used to generate an answer. The generation process takes into account emotional data and optimizes the tone and content of the answer. For example, it generates a specific sentence such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu."

[1125] Step 8:

[1126] Server: Structure the generated answer (e.g., in JSON format) and send it to the device as an HTTP response, using Flask or Django's response object with the appropriate status code.

[1127] Step 9:

[1128] Terminal: Parses the received answer and displays it in the user interface. JavaScript parses the response data and inserts the answer text into an HTML element to display it visually to the user.

[1129] Step 10:

[1130] User: Review the displayed answers and rate them for appropriateness. Once you have finished rating, click the feedback button to provide your feedback.

[1131] Step 11:

[1132] Terminal: Structure the feedback entered by the user (convert it to JSON format) and send it to the server. Using JavaScript, send the feedback data to the server via an AJAX request.

[1133] Step 12:

[1134] Server: Stores the received feedback in a database. Generates an SQL query to insert into the feedback table in the database.

[1135] Step 13:

[1136] Server: Retrains the NLP engine, NLG model, and sentiment engine based on the stored feedback. Each model learns using new feedback data to improve its performance.

[1137] These are the specific processing steps of the new employee support generation AI system that combines an emotion engine. This system allows new employees to efficiently obtain internal company information and receive personalized support based on their emotions.

[1138] Example 2

[1139] 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."

[1140] When new employees start work, it is urgent to provide them with a means to efficiently access internal company information. However, conventional systems provide mechanical responses without considering the user's feelings, making it difficult to improve the user's psychological sense of security and satisfaction. Furthermore, the system lacks the ability to adaptively improve itself based on feedback, making it difficult to provide information that meets the user's needs.

[1141] 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.

[1142] In this invention, the server includes: means for analyzing user questions using a natural language processing engine; means for analyzing the user's emotions using the analyzed question content and an emotion recognition engine; means for searching for related information from a database based on the analyzed question content and emotion data; means for generating answers based on the searched information and using a generative AI model that adjusts the content according to the emotion; means for sending the generated answers to the user's device; means for receiving user feedback; and means for improving the model using the received feedback. This not only enables new employees to quickly and accurately access the information they need, but also provides personalized answers based on emotion data. Furthermore, by improving the system using feedback, user satisfaction and reliability can be continuously improved.

[1143] A "natural language processing engine" is a software technology that analyzes text entered by a user and understands its content and structure.

[1144] An "emotion recognition engine" is a technology that analyzes emotions from a user's text and identifies emotional states such as anxiety, anger, and joy.

[1145] A "database" is a data management system for storing related information in a structured way that makes it easy to search, store, and update.

[1146] A "generative AI model" is a machine learning model that uses natural language data as input and automatically generates appropriate answers based on the context.

[1147] A "user terminal" is an electronic device that a user uses to access the system, enter questions, and view answers.

[1148] "Feedback" refers to the user's evaluation of the system's answers and suggestions for improvement.

[1149] "Model improvement" is the process of correcting and updating the system's learning data and algorithms based on the feedback received, thereby improving answer accuracy and user satisfaction.

[1150] This invention is a system designed to enable new employees to efficiently access company information, and includes an emotion recognition engine to recognize the user's emotions and adjust responses accordingly. This system operates in cooperation between a server, a terminal, and a user.

[1151] Server configuration and roles

[1152] The following main software modules run on the server:

[1153] 1. Natural Language Processing Engines (e.g. SpaCy, NLTK):

[1154] It is used to analyze user questions and extract keywords and context.

[1155] 2. Emotion Recognition Engine (e.g. Hugging Face Transformers, IBM Watson):

[1156] Analyzes emotions from user text and recognizes feelings such as anxiety and doubt.

[1157] 3. Database (e.g. MySQL, PostgreSQL):

[1158] Store and search relevant information within your company.

[1159] 4. Generative AI models (e.g., OpenAI GPT-3, BERT):

[1160] Generate appropriate answers for users based on search results and sentiment data.

[1161] The server's operating process is as follows:

[1162] 1. Receiving and parsing questions:

[1163] The server receives the user's question sent from the device and analyzes it using a natural language processing engine. For example, it may receive a question such as "How do I apply for paid vacation?"

[1164] 2. Emotion Recognition:

[1165] An emotion recognition engine is used to analyze the emotions of the user based on the content of the question. For example, it can recognize that the user is feeling anxious.

[1166] 3. Information Search:

[1167] Based on the analyzed keywords and emotion data, the server searches the database for relevant information, for example, identifying information on "paid vacation" and "how to apply."

[1168] 4. Generate and refine answers:

[1169] Based on the search results and sentiment data, a generative AI model is used to generate a response for the user. The tone and content of the response are adjusted based on the sentiment data. For example, a response such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu" can be generated.

[1170] 5. Submit your response:

[1171] The generated response is sent to the device.

[1172] 6. Receive feedback and learn:

[1173] It receives user feedback and uses it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[1174] Device configuration and role

[1175] The terminal provides an intuitive and easy-to-use interface for users to access the system, which:

[1176] 1. Receiving user input:

[1177] Allow users to enter their question in natural language, for example by typing their question in a text box and clicking the submit button.

[1178] 2. Show Answer:

[1179] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[1180] 3. Enter your feedback:

[1181] Users can provide feedback on the answers, for example by clicking a feedback button and entering a rating, which is then sent to the server.

[1182] User operations

[1183] A user follows the steps below to search for the information they need through the system.

[1184] 1. Enter your question:

[1185] Users use the device interface to enter questions in natural language.

[1186] 2. Check your answers:

[1187] The answer sent from the server is confirmed on the device screen.

[1188] 3. Providing Feedback:

[1189] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[1190] Specific examples

[1191] Example 1: How to apply for paid leave

[1192] User: Type "How do I request paid time off?" into the device.

[1193] Server: Receives the question and extracts the keywords "paid leave" and "how to apply" as well as the emotion "anxiety." It searches the database for relevant information and uses a generative AI model to generate an answer such as, "Paid leave applications are made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu." This is then sent to the device.

[1194] Terminal: Displays the generated answer to the user.

[1195] Users: Rate the answers for accuracy and provide feedback.

[1196] Example 2: New employee training schedule

[1197] User: Type into terminal, "What is the new employee training schedule?"

[1198] Server: Receives the question and extracts the keywords "new employee training" and "schedule" as well as the emotion "doubt." Searches for related information in the database and uses a generative AI model to generate an answer such as "The new employee training schedule is listed on the company calendar. Please check here." and sends it to the device.

[1199] Terminal: Displays the generated answer to the user.

[1200] Users: Rate the answers for accuracy and provide feedback.

[1201] In this way, the system not only enables new employees to quickly and accurately obtain the information they need, but also utilizes emotional data to provide more personalized support, resulting in improved work efficiency and a greater sense of security for new employees.

[1202] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1203] Step 1:

[1204] The user inputs a question into the device interface. For example, the user inputs "How do I apply for paid leave?" into the text box. This operation saves the question in text format on the device.

[1205] Step 2:

[1206] When the user clicks the submit button, the device sends the question to the server as an HTTP POST request, which includes the question text.

[1207] Step 3:

[1208] The server receives an HTTP POST request from the device. It parses the question text and extracts keywords and contextual information using a natural language processing engine (e.g., SpaCy, NLTK). This process extracts keywords such as "paid vacation" and "how to apply" from the question text.

[1209] Step 4:

[1210] The server analyzes the user's emotions using an emotion recognition engine (e.g., Hugging Face Transformers, IBM Watson) based on the extracted keywords. For example, it detects the user's anxiety from the question text. This process generates emotion data such as "anxiety."

[1211] Step 5:

[1212] The server uses the keywords and sentiment data to search for relevant information from a database (e.g., MySQL, PostgreSQL). For example, it searches for company policies and procedures related to "paid leave" and "how to apply." This process retrieves the relevant information in text format.

[1213] Step 6:

[1214] The server generates an answer using a generative AI model (e.g., OpenAI GPT-3, BERT) based on the search results and emotion data. The generated prompt includes keywords, emotion data, and search results. This process generates an answer text such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu."

[1215] Step 7:

[1216] The server sends the generated answer text to the terminal as an HTTP response, which includes the answer text.

[1217] Step 8:

[1218] The terminal displays the answer text received from the server to the user, for example, by displaying the answer in a text area on the screen.

[1219] Step 9:

[1220] The user inputs feedback on the displayed answer, for example, by clicking a feedback button and inputting an evaluation such as "appropriate" or "inappropriate."

[1221] Step 10:

[1222] The device sends the user-provided feedback to the server as an HTTP POST request.

[1223] Step 11:

[1224] The server receives feedback sent from the device and analyzes the feedback data. Based on this data, it evaluates the performance of the AI ​​model and adjusts the model or updates the training data as needed, thereby improving the accuracy and reliability of the system.

[1225] (Application example 2)

[1226] 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."

[1227] In modern factories, new employees and technicians need a large amount of information and quick access to it in order to adapt to complex machines and processes. However, with previous systems, obtaining information and troubleshooting took time, often causing anxiety and stress. Therefore, an information delivery system that can provide efficient and emotional support is needed.

[1228] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1229] In this invention, the server includes a means for analyzing a user's question using a natural language processing engine, a means for searching for related information from a database based on the analyzed question, and a means for using a natural language generation model to generate an answer based on the searched information. This allows the user to receive an answer tailored to their own emotions. Furthermore, by providing an interface accessible from the robot's display or smart device and adding a function for learning from the user's question history and feedback to generate better answers, the server achieves efficient and stress-free information acquisition.

[1230] A "natural language processing engine" is an engine that analyzes questions entered by users and extracts keywords and context.

[1231] A "database" is a collection of information that stores information in response to user questions and manages it in a searchable format.

[1232] A "natural language generation model" is an algorithm for generating appropriate answers to users based on analyzed questions.

[1233] The "emotion engine" is an engine that analyzes emotions from the user's question and input data and adjusts the response content.

[1234] A "terminal" is a device with an interface that allows a user to access the system and input questions.

[1235] An "interface" is an operation screen or application that allows a user to input questions into the system and receive answers.

[1236] "Feedback" refers to the evaluations and opinions that users provide regarding the system's answers, and is used to improve the system.

[1237] A "robot display" is a device installed on a factory robot that displays operation information and responses from the system.

[1238] A "smart device" is a portable electronic device that can connect to the Internet, such as a smartphone or tablet.

[1239] "Question history" is a record of questions and answers that a user has previously entered into the system.

[1240] "Troubleshooting" is the process of diagnosing errors or problems in a system or machine and providing a solution.

[1241] To implement this invention, an automated information acquisition system must be installed on factory robots and related smart devices. This system works in cooperation with a server, terminals, and users.

[1242] Server configuration and roles

[1243] The server runs a natural language processing engine, a natural language generation model, a database, and an emotion engine. Details are as follows:

[1244] 1. Receiving and analyzing the question: The server receives the user's question sent from the terminal. Using a natural language processing engine, the server analyzes the question and extracts appropriate keywords and meanings. For example, the server may receive a question such as, "Please tell me how to deal with error code E101."

[1245] 2. Emotion Recognition: Using an emotion engine, we analyze the emotions in the user's questions, for example, whether the user is feeling anxious or uncertain.

[1246] 3. Information retrieval: Based on the analyzed keywords and sentiment data, the server retrieves relevant information from the database.

[1247] 4. Answer generation and tailoring: Based on the search results and sentiment data, a natural language generation model is used to generate an answer for the user. The tone and content of the answer are adjusted based on the sentiment engine data. For example, if the user is feeling anxious, a phrase like "Don't worry" is added to the answer.

[1248] 5. Sending the answer: The generated answer is sent to the terminal and displayed to the user.

[1249] 6. Receive and learn feedback: Receive user feedback and use it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[1250] Device configuration and role

[1251] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[1252] 1. Receiving user input: Users can input questions in natural language. Questions can be posed via text or voice using the factory robot's display or smart device.

[1253] 2. Displaying the answer: The answer received from the server is displayed to the user, allowing the user to quickly obtain the required information.

[1254] 3. Enter feedback: Users can provide feedback on the answer by clicking the feedback button, entering their rating and sending it to the server.

[1255] User operations

[1256] A user follows the steps below to search for the information they need through the system.

[1257] 1. Entering a question: Users enter a question in natural language using the factory robot's display or a smart device.

[1258] 2. Confirm the answer: Confirm the answer sent from the server on the device screen.

[1259] 3. Providing feedback: Evaluate the appropriateness of the answer and send feedback to the server via the device.

[1260] Specific examples

[1261] Example 1: How to resolve a robot error code

[1262] User: Enter "Please tell me how to deal with error code E101" into the terminal.

[1263] Server: Receives the question and extracts the keywords "error code E101" and "how to deal with it," as well as the emotion "anxiety." It searches for related information in a database and uses a natural language generation model to generate a response such as "Error code E101 indicates a sensor malfunction. Don't worry, try reconnecting the sensor," and sends it to the device.

[1264] Terminal: Displays the generated answer to the user.

[1265] User: Check the solution and perform the task.

[1266] Prompt Sentence Examples

[1267] "How do I resolve the error code E101?"

[1268] This system enables new employees and technicians to quickly address any problems that arise within the factory, improving work efficiency and safety.

[1269] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1270] Processing Steps

[1271] Step 1:

[1272] The user inputs a question in natural language using the factory robot's display or a smart device. For example, the user might input a question such as, "Please tell me how to deal with error code E101." The user's input is sent to the terminal as text data.

[1273] Step 2:

[1274] The device sends text data from the user to the server. Specifically, the device transfers the user's input data to the server as an API request.

[1275] Step 3:

[1276] The server analyzes the user's question received from the terminal using a natural language processing engine. As a result of the analysis, the keywords "error code E101" and "solution" are extracted. In this analysis process, the input text data is broken down into tokens and semantic analysis is performed.

[1277] Step 4:

[1278] The server uses an emotion engine to recognize emotions from the user's question. For example, the emotion "anxiety" is extracted. In this step, an emotion analysis algorithm is applied to the user's input data to generate emotion data.

[1279] Step 5:

[1280] The server searches for relevant information from the database based on the analyzed keywords and emotion data. For example, it retrieves information about "error code E101" in the database. In this search process, an SQL or NoSQL query is constructed to retrieve the corresponding record from the database.

[1281] Step 6:

[1282] The server generates a response to the user using a natural language generation model based on the search results and sentiment data. For example, a response such as "Error code E101 indicates a sensor malfunction. Please rest assured and try reconnecting the sensor" is generated. In this generation process, a natural language generation algorithm is applied to construct the response text.

[1283] Step 7:

[1284] The server sends the generated response to the device. Specifically, it returns the response data to the device as an API response. In this sending process, the generated text data is packaged in JSON format and sent as an HTTP response.

[1285] Step 8:

[1286] The device displays the answer received from the server to the user. Specifically, the generated answer is displayed on the display. In this display process, the received text data is drawn on the UI.

[1287] Step 9:

[1288] The user evaluates whether the answer is appropriate and sends feedback to the server via the terminal. For example, the user may enter feedback such as "This answer was helpful." In this step, the feedback data is sent to the terminal as text.

[1289] Step 10:

[1290] The server receives feedback from users and uses it to improve the AI ​​model. This improvement process uses the received feedback data as training data to improve the accuracy of the natural language processing engine and natural language generation model.

[1291] These processing steps provide fast and appropriate answers to user questions while continually improving the accuracy and reliability of the overall system.

[1292] 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.

[1293] 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.

[1294] 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.

[1295] [Fourth embodiment]

[1296] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1297] 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.

[1298] 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).

[1299] 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.

[1300] 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.

[1301] 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).

[1302] 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.

[1303] 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.

[1304] 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.

[1305] 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.

[1306] 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.

[1307] 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.

[1308] 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."

[1309] This system is designed to enable new employees to efficiently access company information. It works in cooperation with three parties: the server, the terminal, and the user.

[1310] Server configuration and roles

[1311] The server runs a natural language processing engine and a natural language generation model and is connected to a database. This allows the server to fulfill the following roles:

[1312] 1. Receiving and parsing questions:

[1313] The server receives the user's question sent from the device and uses a natural language processing engine to analyze the question and extract appropriate keywords and meanings.

[1314] 2. Information Search:

[1315] Based on the parsed question, the server retrieves relevant information from a database.

[1316] 3. Generate answers:

[1317] Based on the search results, a natural language generation model is used to generate an answer for the user.

[1318] 4. Submit your response:

[1319] The generated answer is sent to the terminal and displayed to the user.

[1320] 5. Receive feedback and learn:

[1321] It receives user feedback and uses it to improve the AI ​​model, which will generate a better answer the next time you ask a question.

[1322] Device configuration and role

[1323] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[1324] 1. Receiving user input:

[1325] Users can enter questions in natural language by simply typing in the text box and clicking the submit button, which sends the question to the server.

[1326] 2. Show Answer:

[1327] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[1328] 3. Enter your feedback:

[1329] Users can provide feedback on the answer by clicking the feedback button, entering a rating and sending it to the server.

[1330] User operations

[1331] A user follows the steps below to search for the information they need through the system.

[1332] 1. Enter your question:

[1333] The user uses the terminal interface to enter a question in natural language.

[1334] 2. Check your answers:

[1335] The answer sent from the server is confirmed on the device screen.

[1336] 3. Providing Feedback:

[1337] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[1338] Specific examples

[1339] Example 1: How to apply for paid leave

[1340] User: Type into the device, "How do I request paid time off?"

[1341] Server: Receives the question, extracts keywords related to "paid leave" and "how to apply," and searches the database for related information. For example, it generates an answer such as, "Paid leave applications are made through the in-house portal site. Log in to the portal site and apply from the 'Paid Leave Application' menu." and sends it to the device.

[1342] Terminal: Displays the generated answer to the user.

[1343] Users: Rate the answers for accuracy and provide feedback.

[1344] Example 2: New employee training schedule

[1345] User: Type into terminal, "What is the new employee training schedule?"

[1346] Server: Receives the question, extracts the keywords "new employee training" and "schedule," and searches the database for related information. For example, it generates an answer such as "The new employee training schedule is listed on the company calendar. Please check here," and sends it to the device.

[1347] Terminal: Displays the generated answer to the user.

[1348] Users: Rate the answers for accuracy and provide feedback.

[1349] In this way, the system is designed to enable new employees to quickly and accurately obtain the information they need, thereby improving work efficiency.

[1350] The processing flow will be explained below.

[1351] Step 1:

[1352] User: Enter a question in natural language through the device's user interface. Enter a question such as "How do I apply for paid leave?" into the input form and click the submit button.

[1353] Step 2:

[1354] Terminal: The question text entered by the user is structured (for example, converted into JSON format) and sent to the server. Specifically, a program such as JavaScript sends the text data to the server via an AJAX request.

[1355] Step 3:

[1356] Server: Receives the question text sent from the terminal. Specifically, it receives an HTTP request using a web framework such as Flask or Django, and the parser reads the data in JSON format.

[1357] Step 4:

[1358] Server: The question text is analyzed using a natural language processing engine (e.g., spaCy, NLTK). During this analysis process, the text is tokenized and subjected to entity recognition and partial analysis to extract important keywords and phrases such as "paid leave" and "how to apply."

[1359] Step 5:

[1360] Server: Based on the analyzed keywords, it searches for relevant information from a database (e.g., MySQL, PostgreSQL). It generates SQL queries and executes them against the database to retrieve the relevant information.

[1361] Step 6:

[1362] Server: Based on the acquired data, a natural language generation model (e.g., GPT-3) is used to generate an answer. For example, it generates a specific sentence such as, "Paid leave applications can be made through the in-house portal site. Please log in to the portal site and apply from the 'Paid Leave Application' menu."

[1363] Step 7:

[1364] Server: Structure the generated answer (e.g., in JSON format) and send it to the device as an HTTP response, using the Flask or Django response object with the appropriate status code.

[1365] Step 8:

[1366] Terminal: Parses the received answer and displays it in the user interface. Specifically, JavaScript parses the response data and inserts the answer text into an HTML element to visually display it to the user.

[1367] Step 9:

[1368] User: Review the displayed answers and rate them for appropriateness. Once you have finished rating, click the feedback button to provide your feedback.

[1369] Step 10:

[1370] Terminal: Structure the feedback entered by the user (convert it to JSON format) and send it to the server. Again using JavaScript, send the feedback data to the server via an AJAX request.

[1371] Step 11:

[1372] Server: Stores the received feedback in a database. Generates an SQL query to insert into the feedback table in the database.

[1373] Step 12:

[1374] Server: Retrains the NLP engine or NLG model based on the stored feedback, learning to improve the model's performance with new feedback data.

[1375] These are the specific processing steps of the new employee support generation AI system. This system allows new employees to efficiently obtain internal company information and improve work efficiency.

[1376] Example 1

[1377] 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."

[1378] In today's corporate environment, there is a demand for systems that allow new employees to efficiently access internal information and quickly obtain the information they need. Conventional systems have problems such as complicated information search and retrieval, making it difficult for users to operate intuitively, and reducing work efficiency. It has also been difficult to effectively utilize user feedback to improve the system's response accuracy.

[1379] 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.

[1380] In this invention, the server includes: means for analyzing a question from a user using a natural language processing engine; means for searching for related information from an information storage device based on the analyzed question content; means for using a natural language generation model to generate an answer based on the searched information; means for transmitting the generated answer to the user's information display device; means for receiving user evaluation information; means for improving the model using the received evaluation information; means for analyzing the question entered by the user and generating an answer by inputting a prompt sentence into the generation AI model based on the analysis result; means for displaying the generated answer; and means for analyzing the evaluation information provided by the user and reflecting it in the next answer generation. This allows new employees to access internal company information efficiently and intuitively. Furthermore, the accuracy of answer generation can be improved based on user feedback.

[1381] A "natural language processing engine" is a software technology that analyzes questions from users and extracts keywords and meanings.

[1382] A "natural language generation model" is an algorithm for generating appropriate answers in natural language based on the analyzed question content.

[1383] An "information storage device" is hardware or software for storing and managing information, such as a database or file system.

[1384] An "information display device" is a screen, monitor, or similar device for displaying information to a user.

[1385] "Rating Information" refers to feedback and ratings provided by users, and is data used to improve the system's response accuracy.

[1386] A "prompt" is a statement of instructions or a question that is input into a generative AI model to generate a specific answer.

[1387] "Analysis" is the process of understanding a user's question, extracting meaning, and identifying keywords.

[1388] "Generation" is the process of generating new, appropriate answers or text based on the retrieved information.

[1389] "Sending" is the process of sending a response or data from the server to the terminal over a communication path.

[1390] "Learning" refers to the application of machine learning algorithms to improve the system's response accuracy based on user feedback.

[1391] MODE FOR CARRYING OUT THE INVENTION

[1392] The system of the present invention is designed to enable new employees to efficiently access company information and quickly obtain the information they need. It uses a natural language processing engine, a natural language generation model, an information storage device, and an information display device in cooperation with a server, a terminal, and a user.

[1393] First, the user inputs a question into the terminal. The terminal allows the user to input questions in natural language through a user interface. This interface includes a text box and a submit button, and is designed to be intuitive. For example, if the user inputs "How do I apply for paid vacation?", the question will be submitted.

[1394] Next, the device sends the user's question to the server. The server receives the question and analyzes it using a natural language processing engine. Software such as spaCy or NLTK can be used as the natural language processing engine. The question is analyzed to extract keywords such as "paid leave" and "how to apply."

[1395] Based on the parsed query, the server searches for relevant information using an information storage device, such as MySQL or PostgreSQL, and executes a database query to retrieve the appropriate information.

[1396] Next, the server uses a natural language generation model to generate an answer based on the search results. OpenAI's GPT-3 and other generative AI models can be applied. As a specific example, the prompt "Please explain how to apply for paid leave" is entered, and an appropriate answer is generated. The generated answer is something like, "Paid leave applications are made through the in-house portal site. Please log in to the portal site and apply from the 'Paid Leave Application' menu."

[1397] The generated answer is sent from the server to the terminal and displayed to the user, allowing the user to quickly obtain the information they need. The user can also enter evaluation information for the provided answer, i.e., feedback. Feedback can include evaluations such as "It was helpful" or "I'd like more details."

[1398] Finally, the server receives user feedback and analyzes the ratings to improve the natural language generation model, using machine learning algorithms to adjust the model to generate better answers for the next question.

[1399] For example, if the user's evaluation feedback is that the answer was not specific, the server will adjust the model parameters to provide more detailed information the next time it responds, thereby improving the response quality of the entire system.

[1400] This system allows new employees to efficiently access internal company information and quickly obtain the information they need. Furthermore, response accuracy can be improved based on user feedback, allowing for continuous improvement.

[1401] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1402] Step 1: Enter your question

[1403] The user enters a question into the text box on the terminal. For example, the user enters "How do I apply for paid leave?". Specifically, the user enters the question using the keyboard and clicks the send button. Input: User's question. Output: Question data in text format.

[1404] Step 2: Submit your question

[1405] The device sends the question entered by the user to the server. Specifically, when the send button is clicked, the device creates an HTTP request and sends the question to the server's API endpoint. Input: Question data in text format. Output: HTTP request to the server.

[1406] Step 3: Analyzing the Question

[1407] The server analyzes the question received from the device. A natural language processing engine (e.g., spaCy or NLTK) is used to tokenize the question and extract key keywords (e.g., "paid leave" and "how to apply"). Specifically, the question is tokenized and tagged with parts of speech to identify important keywords. Input: Text question data extracted from the HTTP request. Output: Extracted keywords and their analysis results.

[1408] Step 4: Finding information

[1409] The server searches for related information from an information storage device (e.g., MySQL or PostgreSQL) based on the analysis results. Specifically, the server generates an SQL query to retrieve relevant records from the database. Input: Parsed keywords. Output: Related information retrieved from the database.

[1410] Step 5: Generate an answer

[1411] The server inputs a prompt sentence into a natural language generation model (for example, OpenAI's GPT-3) based on the acquired information, and generates an appropriate answer. Specifically, a prompt sentence such as "Please explain how to apply for paid leave" is input into the generative AI model, and the generated text is used as the answer. Input: Relevant information acquired from the database. Output: Generated answer text.

[1412] Step 6: Submit your response

[1413] The server sends the generated response to the terminal. Specifically, it includes the generated text as an HTTP response and sends it to the terminal. Input: Generated response text. Output: HTTP response to the terminal.

[1414] Step 7: Check your answers

[1415] The user checks the answer from the server displayed on the device. For example, the answer displayed is "Paid leave applications can be made through the in-house portal site." In concrete terms, the user looks at the device screen, reads the answer, and understands its contents. Input: Answer text received from the server. Output: User's understanding.

[1416] Step 8: Enter and submit your feedback

[1417] The user inputs feedback for the provided answer. For example, they rate it as "very helpful" or "I'd like more details." Specifically, the user clicks the feedback button, writes their rating, and clicks the submit button. Input: User feedback. Output: Feedback input to the server.

[1418] Step 9: Analyze feedback and learn

[1419] The server receives feedback from users and analyzes it to improve the quality of the AI ​​model. Specifically, the feedback data is stored in a database and evaluated by a machine learning algorithm. This adjusts the AI ​​model to generate a better answer for the next question. Input: User feedback. Output: Improvement and adjustment of the AI ​​model.

[1420] (Application example 1)

[1421] 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."

[1422] In logistics centers, it is important for new employees and workers to quickly obtain the information they need to perform their work in order to carry out their work efficiently. However, accessing complex work procedures and a wide range of inventory information requires advanced knowledge and experience, which places a heavy burden on new employees and first-time workers. A system that solves this issue and simplifies access to information is needed.

[1423] 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.

[1424] In this invention, the server includes means for analyzing instructions from a user using a natural language processing engine, means for searching for related information from an information storage unit based on the analyzed instructions, means for using a natural language generation algorithm to generate an answer based on the searched information, means for transmitting the generated answer to the user's display device, means for receiving user evaluations, means for improving the algorithm using the received evaluations, communication means for the user to request individual information via a wide area network, and dynamic search means for acquiring the requested individual information in real time, thereby enabling new employees and workers to quickly and accurately obtain the information they need.

[1425] A "natural language processing engine" is a software technology for analyzing text data and extracting keywords and meanings from within sentences.

[1426] "User instructions" refer to questions or commands that a user enters into the system.

[1427] The "information storage unit" refers to a database or storage system that stores related information.

[1428] "Related information" refers to the information to be searched for based on a user instruction.

[1429] A "natural language generation algorithm" is an algorithm that generates answers in natural language based on analyzed information.

[1430] A "display device" is a device for visually displaying information to a user.

[1431] "Rating" refers to the feedback or review a user gives to an answer provided by the system.

[1432] "Wide area network" refers to a wide-area communications network such as the Internet.

[1433] "Communication means" refers to the technology for sending user instructions to the server and for sending responses from the server to the user.

[1434] "Dynamic search means" refers to a technology that instantly searches for and retrieves requested information in real time.

[1435] This invention is a system that enables new employees and workers at a logistics center to quickly obtain information necessary for their work. This system operates in cooperation with three parties: a server, a terminal, and a user.

[1436] Server configuration and roles

[1437] The server runs a natural language processing engine and a natural language generation algorithm, and is connected to the information storage unit. The server's roles are as follows:

[1438] 1. Receiving and parsing instructions:

[1439] The server receives user instructions sent from the device and uses a natural language processing engine (e.g., spaCy) to analyze the instructions and extract relevant keywords and meanings.

[1440] 2. Information Search:

[1441] Based on the parsed instructions, the server searches for relevant information from an information storage unit (e.g., a MySQL database).

[1442] 3. Generate answers:

[1443] Based on the search results, a natural language generation algorithm (e.g., GPT-4) is used to generate an answer for the user.

[1444] 4. Submit your response:

[1445] The generated answer is sent to the terminal and displayed to the user.

[1446] 5. Receiving and learning from assessments:

[1447] It receives user ratings and uses them to improve its algorithms, which will generate better answers for the next prompt.

[1448] Device configuration and role

[1449] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[1450] 1. Receiving user input:

[1451] Users can enter instructions in natural language by simply typing in a text box and clicking the submit button, which sends the instructions to the server.

[1452] 2. Show Answer:

[1453] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[1454] 3. Enter your rating:

[1455] Users can provide a rating for an answer by clicking the rating button, which enters their feedback and sends it to the server.

[1456] User operations

[1457] A user follows the steps below to search for the information they need through the system.

[1458] 1. Enter instructions:

[1459] The user uses the terminal's interface to input instructions in natural language.

[1460] For example: "Please let me know the availability of the product."

[1461] 2. Check your answers:

[1462] The answer sent from the server is confirmed on the device screen.

[1463] 3. Providing Evaluations:

[1464] The answer is evaluated for appropriateness and the evaluation is sent to the server via the terminal.

[1465] Providing concrete examples and prompts

[1466] Specific examples

[1467] Consider the case where a new employee enters the following instructions at a logistics center:

[1468] "Please let me know the product availability."

[1469] The server uses a natural language processing engine (spaCy) to extract keywords such as "product" and "stock status," and searches for related information in the information storage unit (MySQL database).Then, it uses a natural language generation algorithm (GPT-4) to generate the following answer:

[1470] "We currently have 50 units of product A in stock. They are ready to ship."

[1471] Prompt Sentence Examples

[1472] The prompt has the following format:

[1473] "Provide a detailed answer for: Product A's stock status"

[1474] This system is designed to enable new employees and workers to quickly and accurately obtain the information they need, thereby improving operational efficiency at logistics centers.

[1475] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1476] Step 1:

[1477] The user uses the device interface to input instructions in natural language, for example, "Please tell me the product availability status," and clicks the submit button. This input is sent from the device to the server.

[1478] Step 2:

[1479] The server receives the user's instructions from the device and analyzes them using a natural language processing engine (e.g., spaCy). Specifically, it tokenizes the instruction text and extracts keywords and meanings. The user's instructions are given as input, and keywords are obtained as output.

[1480] Step 3:

[1481] The server searches for related information from the information storage unit (database example: MySQL) based on the analyzed keywords. For example, if "product inventory status" is extracted as a keyword, the corresponding inventory information is searched for in the database. The extracted keywords are given as input, and the search results (inventory status, etc.) are obtained as output.

[1482] Step 4:

[1483] The server generates an answer for the user using a natural language generation algorithm (e.g., GPT-4) based on the search results. Specifically, the search results are given in the form of a prompt sentence, and an answer in natural language is generated by the generative AI model. The search results are given as input, and the generated answer is obtained as output.

[1484] Step 5:

[1485] The server sends the generated answer to the terminal. The terminal displays the answer received from the server to the user. The generated answer is sent from the server as input, received by the terminal, and displayed to the user as output.

[1486] Step 6:

[1487] The user checks the displayed answers and rates them. The rating is sent from the device to the server. This rating is used to improve the algorithm. The user's rating is sent from the device to the server as input and reflected in the algorithm model as output.

[1488] Step 7:

[1489] The server uses the received evaluations to improve the natural language generation algorithm. Specifically, it adds the evaluation data as training data to improve the accuracy of the next answer generation. The evaluation data is given as input and an improved generation algorithm is obtained as output.

[1490] 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.

[1491] This invention is a system designed to enable new employees to efficiently access company information, and includes an emotion engine that recognizes the user's emotions and adjusts responses. This system operates in cooperation between a server, a terminal, and a user.

[1492] Server configuration and roles

[1493] The server runs a natural language processing engine, a natural language generation model, a database, and an emotion engine, and performs the following roles:

[1494] 1. Receiving and parsing questions:

[1495] The server receives the user's question sent from the device. Using a natural language processing engine, it analyzes the question and extracts appropriate keywords and meanings. For example, it may receive a question such as, "How do I apply for paid leave?"

[1496] 2. Emotion Recognition:

[1497] The emotion engine analyzes the user's emotions from the content of their question, for example, recognizing whether the user is feeling anxious or uncertain.

[1498] 3. Information Search:

[1499] Based on the analyzed keywords and emotion data, the server retrieves relevant information from a database.

[1500] 4. Generate and refine answers:

[1501] Based on the search results and sentiment data, a natural language generation model is used to generate a response for the user. Based on the sentiment engine data, the tone and content of the response are adjusted. For example, if the user is feeling anxious, a phrase such as "Don't worry" is added to the response.

[1502] 5. Submit your response:

[1503] The generated answer is sent to the terminal and displayed to the user.

[1504] 6. Receive feedback and learn:

[1505] It receives user feedback and uses it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[1506] Device configuration and role

[1507] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[1508] 1. Receiving user input:

[1509] Users can enter questions in natural language by simply typing in the text box and clicking the submit button, which sends the question to the server.

[1510] 2. Show Answer:

[1511] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[1512] 3. Enter your feedback:

[1513] Users can provide feedback on the answer by clicking the feedback button, entering a rating and sending it to the server.

[1514] User operations

[1515] A user follows the steps below to search for the information they need through the system.

[1516] 1. Enter your question:

[1517] The user uses the terminal interface to enter a question in natural language.

[1518] 2. Check your answers:

[1519] The answer sent from the server is confirmed on the device screen.

[1520] 3. Providing Feedback:

[1521] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[1522] Specific examples

[1523] Example 1: How to apply for paid leave

[1524] User: Type into the device, "How do I request paid time off?"

[1525] Server: Receives the question and extracts the keywords "paid leave" and "how to apply" as well as the emotion "anxiety." It searches the database for relevant information and uses a natural language generation model to generate an answer such as, "Paid leave applications are made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu." This is then sent to the device.

[1526] Terminal: Displays the generated answer to the user.

[1527] Users: Rate the answers for accuracy and provide feedback.

[1528] Example 2: New employee training schedule

[1529] User: Type into terminal, "What is the new employee training schedule?"

[1530] Server: Receives the question and extracts the keywords "new employee training" and "schedule" as well as the sentiment of "doubt." It searches for related information in the database and uses a natural language generation model to generate an answer such as "The new employee training schedule is listed on the company calendar. Please check here." and sends it to the device.

[1531] Terminal: Displays the generated answer to the user.

[1532] Users: Rate the answers for accuracy and provide feedback.

[1533] In this way, the system not only enables new employees to quickly and accurately obtain the information they need, but also utilizes emotional data to provide more personalized assistance, resulting in improved work efficiency and a greater sense of security for new employees.

[1534] The processing flow will be explained below.

[1535] Step 1:

[1536] User: Enter a question in natural language through the device's user interface, such as "How do I apply for paid leave?", and click the submit button.

[1537] Step 2:

[1538] Terminal: The question text entered by the user is structured (for example, converted into JSON format) and sent to the server. Specifically, a program such as JavaScript sends the text data to the server via an AJAX request.

[1539] Step 3:

[1540] Server: Receives the question text sent from the terminal. Using a web framework such as Flask or Django, it receives an HTTP request and the parser reads the data in JSON format.

[1541] Step 4:

[1542] Server: The question text is analyzed using a natural language processing engine (e.g., spaCy, NLTK). Specifically, the text is tokenized and subjected to entity recognition and partial analysis to extract important keywords and phrases such as "paid leave" and "how to apply."

[1543] Step 5:

[1544] Server: Uses an emotion engine to analyze the emotions contained in the user's question text. For example, it identifies whether the user is feeling "anxiety" or "doubt."

[1545] Step 6:

[1546] Server: Based on the analyzed keywords and sentiment data, it searches for relevant information from a database (e.g., MySQL, PostgreSQL). It generates SQL queries and executes them against the database to retrieve relevant information.

[1547] Step 7:

[1548] Server: Based on the acquired data, a natural language generation model (e.g., GPT-3) is used to generate an answer. The generation process takes into account emotional data and optimizes the tone and content of the answer. For example, it generates a specific sentence such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu."

[1549] Step 8:

[1550] Server: Structure the generated answer (e.g., in JSON format) and send it to the device as an HTTP response, using Flask or Django's response object with the appropriate status code.

[1551] Step 9:

[1552] Terminal: Parses the received answer and displays it in the user interface. JavaScript parses the response data and inserts the answer text into an HTML element to display it visually to the user.

[1553] Step 10:

[1554] User: Review the displayed answers and rate them for appropriateness. Once you have finished rating, click the feedback button to provide your feedback.

[1555] Step 11:

[1556] Terminal: Structure the feedback entered by the user (convert it to JSON format) and send it to the server. Using JavaScript, send the feedback data to the server via an AJAX request.

[1557] Step 12:

[1558] Server: Stores the received feedback in a database. Generates an SQL query to insert into the feedback table in the database.

[1559] Step 13:

[1560] Server: Retrains the NLP engine, NLG model, and sentiment engine based on the stored feedback. Each model learns using new feedback data to improve its performance.

[1561] These are the specific processing steps of the new employee support generation AI system that combines an emotion engine. This system allows new employees to efficiently obtain internal company information and receive personalized support based on their emotions.

[1562] Example 2

[1563] 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."

[1564] When new employees start work, it is urgent to provide them with a means to efficiently access internal company information. However, conventional systems provide mechanical responses without considering the user's feelings, making it difficult to improve the user's psychological sense of security and satisfaction. Furthermore, the system lacks the ability to adaptively improve itself based on feedback, making it difficult to provide information that meets the user's needs.

[1565] 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.

[1566] In this invention, the server includes: means for analyzing user questions using a natural language processing engine; means for analyzing the user's emotions using the analyzed question content and an emotion recognition engine; means for searching for related information from a database based on the analyzed question content and emotion data; means for generating answers based on the searched information and using a generative AI model that adjusts the content according to the emotion; means for sending the generated answers to the user's device; means for receiving user feedback; and means for improving the model using the received feedback. This not only enables new employees to quickly and accurately access the information they need, but also provides personalized answers based on emotion data. Furthermore, by improving the system using feedback, user satisfaction and reliability can be continuously improved.

[1567] A "natural language processing engine" is a software technology that analyzes text entered by a user and understands its content and structure.

[1568] An "emotion recognition engine" is a technology that analyzes emotions from a user's text and identifies emotional states such as anxiety, anger, and joy.

[1569] A "database" is a data management system for storing related information in a structured way that makes it easy to search, store, and update.

[1570] A "generative AI model" is a machine learning model that uses natural language data as input and automatically generates appropriate answers based on the context.

[1571] A "user terminal" is an electronic device that a user uses to access the system, enter questions, and view answers.

[1572] "Feedback" refers to the user's evaluation of the system's answers and suggestions for improvement.

[1573] "Model improvement" is the process of correcting and updating the system's learning data and algorithms based on the feedback received, thereby improving answer accuracy and user satisfaction.

[1574] This invention is a system designed to enable new employees to efficiently access company information, and includes an emotion recognition engine to recognize the user's emotions and adjust responses accordingly. This system operates in cooperation between a server, a terminal, and a user.

[1575] Server configuration and roles

[1576] The following main software modules run on the server:

[1577] 1. Natural Language Processing Engines (e.g. SpaCy, NLTK):

[1578] It is used to analyze user questions and extract keywords and context.

[1579] 2. Emotion Recognition Engine (e.g. Hugging Face Transformers, IBM Watson):

[1580] Analyzes emotions from user text and recognizes feelings such as anxiety and doubt.

[1581] 3. Database (e.g. MySQL, PostgreSQL):

[1582] Store and search relevant information within your company.

[1583] 4. Generative AI models (e.g., OpenAI GPT-3, BERT):

[1584] Generate appropriate answers for users based on search results and sentiment data.

[1585] The server's operating process is as follows:

[1586] 1. Receiving and parsing questions:

[1587] The server receives the user's question sent from the device and analyzes it using a natural language processing engine. For example, it may receive a question such as "How do I apply for paid vacation?"

[1588] 2. Emotion Recognition:

[1589] An emotion recognition engine is used to analyze the emotions of the user based on the content of the question. For example, it can recognize that the user is feeling anxious.

[1590] 3. Information Search:

[1591] Based on the analyzed keywords and emotion data, the server searches the database for relevant information, for example, identifying information on "paid vacation" and "how to apply."

[1592] 4. Generate and refine answers:

[1593] Based on the search results and sentiment data, a generative AI model is used to generate a response for the user. The tone and content of the response are adjusted based on the sentiment data. For example, a response such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu" can be generated.

[1594] 5. Submit your response:

[1595] The generated response is sent to the device.

[1596] 6. Receive feedback and learn:

[1597] It receives user feedback and uses it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[1598] Device configuration and role

[1599] The terminal provides an intuitive and easy-to-use interface for users to access the system, which:

[1600] 1. Receiving user input:

[1601] Allow users to enter their question in natural language, for example by typing their question in a text box and clicking the submit button.

[1602] 2. Show Answer:

[1603] The answer received from the server is displayed to the user, allowing the user to quickly obtain the information they need.

[1604] 3. Enter your feedback:

[1605] Users can provide feedback on the answers, for example by clicking a feedback button and entering a rating, which is then sent to the server.

[1606] User operations

[1607] A user follows the steps below to search for the information they need through the system.

[1608] 1. Enter your question:

[1609] Users use the device interface to enter questions in natural language.

[1610] 2. Check your answers:

[1611] The answer sent from the server is confirmed on the device screen.

[1612] 3. Providing Feedback:

[1613] The answer is evaluated for appropriateness and feedback is sent to the server via the device.

[1614] Specific examples

[1615] Example 1: How to apply for paid leave

[1616] User: Type "How do I request paid time off?" into the device.

[1617] Server: Receives the question and extracts the keywords "paid leave" and "how to apply" as well as the emotion "anxiety." It searches the database for relevant information and uses a generative AI model to generate an answer such as, "Paid leave applications are made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu." This is then sent to the device.

[1618] Terminal: Displays the generated answer to the user.

[1619] Users: Rate the answers for accuracy and provide feedback.

[1620] Example 2: New employee training schedule

[1621] User: Type into terminal, "What is the new employee training schedule?"

[1622] Server: Receives the question and extracts the keywords "new employee training" and "schedule" as well as the emotion "doubt." Searches for related information in the database and uses a generative AI model to generate an answer such as "The new employee training schedule is listed on the company calendar. Please check here." and sends it to the device.

[1623] Terminal: Displays the generated answer to the user.

[1624] Users: Rate the answers for accuracy and provide feedback.

[1625] In this way, the system not only enables new employees to quickly and accurately obtain the information they need, but also utilizes emotional data to provide more personalized support, resulting in improved work efficiency and a greater sense of security for new employees.

[1626] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1627] Step 1:

[1628] The user inputs a question into the device interface. For example, the user inputs "How do I apply for paid leave?" into the text box. This operation saves the question in text format on the device.

[1629] Step 2:

[1630] When the user clicks the submit button, the device sends the question to the server as an HTTP POST request, which includes the question text.

[1631] Step 3:

[1632] The server receives an HTTP POST request from the device. It parses the question text and extracts keywords and contextual information using a natural language processing engine (e.g., SpaCy, NLTK). This process extracts keywords such as "paid vacation" and "how to apply" from the question text.

[1633] Step 4:

[1634] The server analyzes the user's emotions using an emotion recognition engine (e.g., Hugging Face Transformers, IBM Watson) based on the extracted keywords. For example, it detects the user's anxiety from the question text. This process generates emotion data such as "anxiety."

[1635] Step 5:

[1636] The server uses the keywords and sentiment data to search for relevant information from a database (e.g., MySQL, PostgreSQL). For example, it searches for company policies and procedures related to "paid leave" and "how to apply." This process retrieves the relevant information in text format.

[1637] Step 6:

[1638] The server generates an answer using a generative AI model (e.g., OpenAI GPT-3, BERT) based on the search results and emotion data. The generated prompt includes keywords, emotion data, and search results. This process generates an answer text such as, "Paid leave applications can be made through the in-house portal site. Don't worry, just log in to the portal site and apply from the 'Paid Leave Application' menu."

[1639] Step 7:

[1640] The server sends the generated answer text to the terminal as an HTTP response, which includes the answer text.

[1641] Step 8:

[1642] The terminal displays the answer text received from the server to the user, for example, by displaying the answer in a text area on the screen.

[1643] Step 9:

[1644] The user inputs feedback on the displayed answer, for example, by clicking a feedback button and inputting an evaluation such as "appropriate" or "inappropriate."

[1645] Step 10:

[1646] The device sends the user-provided feedback to the server as an HTTP POST request.

[1647] Step 11:

[1648] The server receives feedback sent from the device and analyzes the feedback data. Based on this data, it evaluates the performance of the AI ​​model and adjusts the model or updates the training data as needed, thereby improving the accuracy and reliability of the system.

[1649] (Application example 2)

[1650] 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."

[1651] In modern factories, new employees and technicians need a large amount of information and quick access to it in order to adapt to complex machines and processes. However, with previous systems, obtaining information and troubleshooting took time, often causing anxiety and stress. Therefore, an information delivery system that can provide efficient and emotional support is needed.

[1652] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1653] In this invention, the server includes a means for analyzing a user's question using a natural language processing engine, a means for searching for related information from a database based on the analyzed question, and a means for using a natural language generation model to generate an answer based on the searched information. This allows the user to receive an answer tailored to their own emotions. Furthermore, by providing an interface accessible from the robot's display or smart device and adding a function for learning from the user's question history and feedback to generate better answers, the server achieves efficient and stress-free information acquisition.

[1654] A "natural language processing engine" is an engine that analyzes questions entered by users and extracts keywords and context.

[1655] A "database" is a collection of information that stores information in response to user questions and manages it in a searchable format.

[1656] A "natural language generation model" is an algorithm for generating appropriate answers to users based on analyzed questions.

[1657] The "emotion engine" is an engine that analyzes emotions from the user's question and input data and adjusts the response content.

[1658] A "terminal" is a device with an interface that allows a user to access the system and input questions.

[1659] An "interface" is an operation screen or application that allows a user to input questions into the system and receive answers.

[1660] "Feedback" refers to the evaluations and opinions that users provide regarding the system's answers, and is used to improve the system.

[1661] A "robot display" is a device installed on a factory robot that displays operation information and responses from the system.

[1662] A "smart device" is a portable electronic device that can connect to the Internet, such as a smartphone or tablet.

[1663] "Question history" is a record of questions and answers that a user has previously entered into the system.

[1664] "Troubleshooting" is the process of diagnosing errors or problems in a system or machine and providing a solution.

[1665] To implement this invention, an automated information acquisition system must be installed on factory robots and related smart devices. This system works in cooperation with a server, terminals, and users.

[1666] Server configuration and roles

[1667] The server runs a natural language processing engine, a natural language generation model, a database, and an emotion engine. Details are as follows:

[1668] 1. Receiving and analyzing the question: The server receives the user's question sent from the terminal. Using a natural language processing engine, the server analyzes the question and extracts appropriate keywords and meanings. For example, the server may receive a question such as, "Please tell me how to deal with error code E101."

[1669] 2. Emotion Recognition: Using an emotion engine, we analyze the emotions in the user's questions, for example, whether the user is feeling anxious or uncertain.

[1670] 3. Information retrieval: Based on the analyzed keywords and sentiment data, the server retrieves relevant information from the database.

[1671] 4. Answer generation and tailoring: Based on the search results and sentiment data, a natural language generation model is used to generate an answer for the user. The tone and content of the answer are adjusted based on the sentiment engine data. For example, if the user is feeling anxious, a phrase like "Don't worry" is added to the answer.

[1672] 5. Sending the answer: The generated answer is sent to the terminal and displayed to the user.

[1673] 6. Receive and learn feedback: Receive user feedback and use it to improve the AI ​​model, thereby increasing the accuracy and reliability of the system.

[1674] Device configuration and role

[1675] The terminal provides the user with an interface to access the system. This interface is designed to be intuitive and easy to use, and performs the following functions:

[1676] 1. Receiving user input: Users can input questions in natural language. Questions can be posed via text or voice using the factory robot's display or smart device.

[1677] 2. Displaying the answer: The answer received from the server is displayed to the user, allowing the user to quickly obtain the required information.

[1678] 3. Enter feedback: Users can provide feedback on the answer by clicking the feedback button, entering their rating and sending it to the server.

[1679] User operations

[1680] A user follows the steps below to search for the information they need through the system.

[1681] 1. Entering a question: Users enter a question in natural language using the factory robot's display or a smart device.

[1682] 2. Confirm the answer: Confirm the answer sent from the server on the device screen.

[1683] 3. Providing feedback: Evaluate the appropriateness of the answer and send feedback to the server via the device.

[1684] Specific examples

[1685] Example 1: How to resolve a robot error code

[1686] User: Enter "Please tell me how to deal with error code E101" into the terminal.

[1687] Server: Receives the question and extracts the keywords "error code E101" and "how to deal with it," as well as the emotion "anxiety." It searches for related information in a database and uses a natural language generation model to generate a response such as "Error code E101 indicates a sensor malfunction. Don't worry, try reconnecting the sensor," and sends it to the device.

[1688] Terminal: Displays the generated answer to the user.

[1689] User: Check the solution and perform the task.

[1690] Prompt Sentence Examples

[1691] "How do I resolve the error code E101?"

[1692] This system enables new employees and technicians to quickly address any problems that arise within the factory, improving work efficiency and safety.

[1693] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1694] Processing Steps

[1695] Step 1:

[1696] The user inputs a question in natural language using the factory robot's display or a smart device. For example, the user might input a question such as, "Please tell me how to deal with error code E101." The user's input is sent to the terminal as text data.

[1697] Step 2:

[1698] The device sends text data from the user to the server. Specifically, the device transfers the user's input data to the server as an API request.

[1699] Step 3:

[1700] The server analyzes the user's question received from the terminal using a natural language processing engine. As a result of the analysis, the keywords "error code E101" and "solution" are extracted. In this analysis process, the input text data is broken down into tokens and semantic analysis is performed.

[1701] Step 4:

[1702] The server uses an emotion engine to recognize emotions from the user's question. For example, the emotion "anxiety" is extracted. In this step, an emotion analysis algorithm is applied to the user's input data to generate emotion data.

[1703] Step 5:

[1704] The server searches for relevant information from the database based on the analyzed keywords and emotion data. For example, it retrieves information about "error code E101" in the database. In this search process, an SQL or NoSQL query is constructed to retrieve the corresponding record from the database.

[1705] Step 6:

[1706] The server generates a response to the user using a natural language generation model based on the search results and sentiment data. For example, a response such as "Error code E101 indicates a sensor malfunction. Please rest assured and try reconnecting the sensor" is generated. In this generation process, a natural language generation algorithm is applied to construct the response text.

[1707] Step 7:

[1708] The server sends the generated response to the device. Specifically, it returns the response data to the device as an API response. In this sending process, the generated text data is packaged in JSON format and sent as an HTTP response.

[1709] Step 8:

[1710] The device displays the answer received from the server to the user. Specifically, the generated answer is displayed on the display. In this display process, the received text data is drawn on the UI.

[1711] Step 9:

[1712] The user evaluates whether the answer is appropriate and sends feedback to the server via the terminal. For example, the user may enter feedback such as "This answer was helpful." In this step, the feedback data is sent to the terminal as text.

[1713] Step 10:

[1714] The server receives feedback from users and uses it to improve the AI ​​model. This improvement process uses the received feedback data as training data to improve the accuracy of the natural language processing engine and natural language generation model.

[1715] These processing steps provide fast and appropriate answers to user questions while continually improving the accuracy and reliability of the overall system.

[1716] 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.

[1717] 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.

[1718] 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.

[1719] 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.

[1720] 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.

[1721] 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.

[1722] 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).

[1723] 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.

[1724] 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."

[1725] 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.

[1726] 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).

[1727] 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.

[1728] 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.

[1729] 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.

[1730] 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.

[1731] 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.

[1732] 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.

[1733] 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.

[1734] 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.

[1735] 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.

[1736] 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.

[1737] The following is further disclosed regarding the above embodiment.

[1738] (Claim 1)

[1739] A means for analyzing a question from a user using a natural language processing engine;

[1740] A means for searching for relevant information from a database based on the analyzed question content;

[1741] a means for using a natural language generation model to generate answers based on the retrieved information;

[1742] means for transmitting the generated answer to the user's terminal;

[1743] a means for receiving user feedback;

[1744] A system that includes a means of using received feedback to improve the model.

[1745] (Claim 2)

[1746] 10. The system of claim 1, wherein the user interface allows a user to enter a question in natural language.

[1747] (Claim 3)

[1748] 2. The system according to claim 1, having a function of learning from a user's question history and feedback to generate better answers.

[1749] "Example 1"

[1750] (Claim 1)

[1751] A means for analyzing a question from a user using a natural language processing engine;

[1752] A means for searching for related information from an information storage device based on the analyzed question content;

[1753] a means for using a natural language generation model to generate answers based on the retrieved information;

[1754] means for transmitting the generated answer to the user's information display device;

[1755] means for receiving user evaluation information;

[1756] a means for using the received evaluation information to improve the model; and

[1757] A means for analyzing a question input by a user and inputting a prompt sentence into a generative AI model based on the analysis result to generate an answer;

[1758] a means for displaying the generated answers;

[1759] A means for analyzing the evaluation information provided by the user and reflecting it in generating the next answer;

[1760] A system including:

[1761] (Claim 2)

[1762] 10. The system of claim 1, wherein the user interface allows a user to enter a question in natural language.

[1763] (Claim 3)

[1764] 2. The system according to claim 1, having a function of learning a user's question history and evaluation information to generate better answers.

[1765] "Application Example 1"

[1766] (Claim 1)

[1767] means for analyzing instructions from a user using a natural language processing engine;

[1768] a means for searching for related information from an information storage unit based on the analyzed instruction content;

[1769] a means for using a natural language generation algorithm to generate an answer based on the sought information;

[1770] means for transmitting the generated answer to a user's display device;

[1771] means for receiving user ratings;

[1772] a means of using the received evaluations to improve the algorithm;

[1773] communication means for a user to request the individual information via a wide area network;

[1774] A system including dynamic search means for obtaining requested specific information in real time.

[1775] (Claim 2)

[1776] 10. The system of claim 1, wherein the user can input instructions in natural language through a user interface.

[1777] (Claim 3)

[1778] 2. The system according to claim 1, having a function of learning a user's instruction history and evaluations to generate more appropriate answers.

[1779] "Example 2: Combining Emotion Engines"

[1780] (Claim 1)

[1781] A means for analyzing a question from a user using a natural language processing engine;

[1782] means for analyzing the user's emotions using the analyzed question content and an emotion recognition engine;

[1783] A means for searching for relevant information from a database based on the analyzed question content and emotion data;

[1784] Using a generative AI model to generate answers based on searched information and adjust content based on emotion;

[1785] means for transmitting the generated answer to the user's terminal;

[1786] a means for receiving user feedback;

[1787] A system that includes a means of using received feedback to improve the model.

[1788] (Claim 2)

[1789] 10. The system of claim 1, wherein the user interface allows a user to enter questions in natural language, view answers, and enter feedback.

[1790] (Claim 3)

[1791] 10. The system of claim 1, having the ability to learn from a user's question history and feedback to generate better answers, and to personalize the answers using emotional data.

[1792] "Application example 2 when combining emotion engines"

[1793] (Claim 1)

[1794] A means for analyzing a question from a user using a natural language processing engine;

[1795] A means for searching for relevant information from a database based on the analyzed question content;

[1796] a means for using a natural language generation model to generate answers based on the retrieved information;

[1797] means for transmitting the generated answer to the user's terminal;

[1798] a means for receiving user feedback;

[1799] a means of using the feedback received to improve the model;

[1800] a means for using an emotion engine to recognize the user's emotions and adjust the response content;

[1801] A means for providing an interface that can be accessed from the robot's display or smart device; and

[1802] A system including:

[1803] (Claim 2)

[1804] 10. The system of claim 1, wherein the user interface allows a user to enter a question in natural language.

[1805] (Claim 3)

[1806] 2. The system according to claim 1, having a function of learning from a user's question history and feedback to generate better answers. [Explanation of symbols]

[1807] 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 analyzing a question from a user using a natural language processing engine; A means for searching for relevant information from a database based on the analyzed question content; a means for using a natural language generation model to generate answers based on the retrieved information; means for transmitting the generated answer to the user's terminal; a means for receiving user feedback; A system that includes a means of using received feedback to improve the model.

2. 10. The system of claim 1, wherein the user interface allows a user to input a question in natural language.

3. The system according to claim 1, further comprising a function for learning from a user's question history and feedback to generate better answers.

Images