System

Abstract

A system is provided.SOLUTION: A system including means for uploading expertise from a user's terminal to a server, means for analyzing the uploaded data to generate metadata and storing the metadata in a database, means for a user to customize a chatbot through a web interface, means for searching for relevant information from the database for an input question, means for generating an appropriate response using a large language model based on the search result, and means for transmitting the generated response back to the end user's terminal.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] In recent years, services using generative AI have rapidly become more widespread, and there has been an increase in use cases that require sophisticated responses and specialized knowledge that cannot be handled by conventional large-scale language models (LLMs). However, there is a lack of platforms that can efficiently utilize data containing specialized knowledge and enable users to easily create sophisticated chatbots. Furthermore, there are no fully developed processes for automating the management of expertise, data analysis, and generation of appropriate responses from search results. This results in expertise remaining unused, creating a problem where user needs cannot be met. [Means for solving the problem]

[0005] The present invention solves the above-mentioned problems by combining a means for uploading expert knowledge from a user's terminal to a server, a means for analyzing the uploaded data to generate metadata and storing it in a database, a means for the user to customize a chatbot through a web interface, a means for searching for related information from a database in response to an input question, a means for generating an appropriate response using a large-scale language model based on the search results, and a means for returning the generated response to the end user's terminal. This enables users to easily generate advanced chatbots that effectively utilize expert knowledge and can appropriately respond to advanced questions in specialized fields.

[0006] "Expertise" is a collection of advanced information and knowledge in a particular field.

[0007] "User" means an individual or legal entity that uses the system, uploads specialized knowledge, and generates a chatbot.

[0008] A "terminal" is a device used by a user, such as a computer, smartphone, or tablet.

[0009] A "server" is a centralized computer system that receives uploaded data, analyzes it, and stores it in a database.

[0010] "Upload" is the act of a user sending data from their own device to a server.

[0011] "Analysis" is the process of breaking down uploaded data based on its content to make it easier to understand.

[0012] "Metadata" is additional information associated with uploaded data, such as title, author, category, etc.

[0013] A "database" is a system that systematically stores and manages analyzed data and its metadata.

[0014] A "web interface" is a screen or operating means that allows a user to access and operate a system through a web browser.

[0015] A "chatbot" is a conversational application that automatically generates responses using a user-configured knowledge base and large-scale language models.

[0016] "Customization" refers to the act of a user changing the chatbot's functions and settings to suit their own needs.

[0017] A "question" is an inquiry that an end user enters into a chatbot.

[0018] "Searching" is the process of locating relevant information from a database based on an entered question.

[0019] A "large-scale language model (LLM)" is a machine learning model that is trained based on large amounts of text data and is used to perform natural language processing.

[0020] A "response" is answer information for the end user that is generated based on search results and a large-scale language model.

[0021] "Returning" is the act of sending the generated response back to the end user's terminal. [Brief explanation of the drawings]

[0022] [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

[0023] 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.

[0024] First, the terms used in the following description will be explained.

[0025] 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).

[0026] 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.

[0027] 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.

[0028] 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.

[0029] 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."

[0030] [First embodiment]

[0031] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0032] 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.

[0033] 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).

[0034] 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.

[0035] 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.

[0036] 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.

[0037] 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.

[0038] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0039] 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.

[0040] 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.

[0041] 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.

[0042] 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."

[0043] The system of the present invention provides a chatbot that generates intelligent responses using data with specialized knowledge. Specific embodiments of the system are described below in natural language.

[0044] Upload your expertise

[0045] Users first upload data files containing specialized knowledge from their own devices to the server. Examples include medical papers, technical manuals, and legal documents. Using a web interface, users select the files they want to upload and press the send button to begin the process. At this point, the file format is verified and saved as a temporary file on the server.

[0046] Data analysis and metadata generation

[0047] The server receives the uploaded data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items. As a result of the analysis, metadata such as the file title, author, and category is generated. This metadata is organized and stored in a database for use by search engines.

[0048] Chatbot generation

[0049] Users can customize their chatbot through a web interface, selecting the database they want to use as a knowledge base, the large-scale language model (LLM) they want to use, and interface settings, and then a chatbot is built on the server that generates responses based on their expert knowledge.

[0050] Enter a question and search

[0051] When an end user types a question into the chatbot, the question is sent to the server, which extracts keywords from the question and searches for related information in the database. Entries related to the question are identified through keyword and ID searches.

[0052] Response Generation

[0053] The server generates a response based on the search results using a large-scale language model (LLM). The LLM interprets the information based on expert knowledge and composes it in natural language, presenting it in a format that is easy for the end user to understand. For example, in response to a question such as "What are the latest cancer treatments?", the server searches the database for the latest medical papers, and the LLM generates an appropriate response based on their content.

[0054] Providing a response

[0055] Finally, the server sends the generated response back to the end user's device. The end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has the ability to provide feedback, which is used to improve the accuracy of future responses.

[0056] Specific examples

[0057] Example 1: Medical knowledge-based chatbot

[0058] 1. The user uploads medical paper data from their device to the server.

[0059] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0060] 3. The user configures the "medical chatbot" through a web interface.

[0061] 4. When an end user asks, "Tell me about the latest cancer treatments," the server searches the database for relevant information and the LLM generates a response.

[0062] 5. The server generates a response and sends it back to the end user.

[0063] Example 2: Legal advice chatbot

[0064] 1. A user uploads a legal document to the server.

[0065] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0066] 3. Users customize the "Legal Advice Chatbot."

[0067] 4. When an end user asks, "What legal details are required for a new business contract?", the server retrieves the information from the relevant documents and the LLM generates the appropriate response.

[0068] 5. The server generates a response and sends it back to the end user.

[0069] As a result, the system of the present invention efficiently utilizes specialized knowledge, enabling users to easily create advanced chatbots and appropriately respond to advanced questions related to specialized fields.

[0070] The processing flow will be explained below.

[0071] Step 1:

[0072] Users upload document files containing their specialized knowledge from their own devices to the server. Using a web interface, users select the files to upload and press the send button. At this time, the file format is confirmed and saved as a temporary file on the server.

[0073] Step 2:

[0074] The server receives the received data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items.

[0075] Step 3:

[0076] The server generates metadata such as file title, author, category, etc. from the analysis results, which are then organized and stored in a database for search engine compatibility.

[0077] Step 4:

[0078] Users configure the chatbot generation through a web interface, choosing the database they want to use as a knowledge base, the large-scale language model (LLM) they want to use, and interface settings to customize the chatbot.

[0079] Step 5:

[0080] The server receives the user's configuration and builds a chatbot that integrates the specified knowledge base and LLM model.

[0081] Step 6:

[0082] The user or end user types a question for the chatbot, which is then sent to the server.

[0083] Step 7:

[0084] The server analyzes the received question and extracts keywords. The extracted keywords are used to search for related information in the database. Entries related to the question are identified through keyword and ID searches.

[0085] Step 8:

[0086] The server generates a response based on the identified search results using a large-scale language model (LLM), which interprets the information based on expert knowledge and formats it in natural language in a format that is easy for the end user to understand.

[0087] Step 9:

[0088] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application.

[0089] Step 10:

[0090] If the response does not meet expectations, users can provide feedback, which is aggregated on the server and used to improve the accuracy of future responses.

[0091] Example 1

[0092] 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."

[0093] Current chatbot systems face the challenge of being unable to efficiently generate responses that incorporate specialized knowledge. Another problem is that it takes a lot of time and effort for users to upload and customize the knowledge base. In particular, when dealing with specialized knowledge, systems that can provide reliable responses instantly are required, but current technology is unable to adequately meet this demand.

[0094] 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.

[0095] In this invention, the server includes means for transmitting specialized knowledge from a user's communication device to the server, means for analyzing the transmitted data to generate metadata and storing it in a storage device, and means for the user to configure an information processing device through an interface, thereby enabling efficient construction of an advanced chatbot based on specialized knowledge and instantaneous generation of specialized responses.

[0096] "Specialized knowledge" is information that includes advanced knowledge and skills in a specific field.

[0097] "Users" are individuals or businesses that use the system to provide their expertise to the chatbot.

[0098] "Communication equipment" refers to a device for sending and receiving data via the Internet, including personal computers and smartphones.

[0099] A "server" is a computer system that receives, processes, and stores data submitted by users.

[0100] "Transmission means" means the method or process for transmitting data from the user's communication device to the server.

[0101] "Analysis means" refers to the method or technology by which the server analyzes the specialized knowledge and generates the metadata.

[0102] "Metadata" is information used to identify and organize the content of data, including title, author, category, etc.

[0103] "Storage device" refers to a device or system for storing analyzed data and metadata.

[0104] "Interface" means the user interface through which a user interacts with the system, including a web browser or dedicated application.

[0105] An "information processing device" is a program and hardware for processing data, such as a chatbot or search engine customized by the user.

[0106] An "enquiry" is a question or request that an end user types into a chatbot.

[0107] An "extraction means" is a method or technique for selecting relevant information from a storage device in response to a query.

[0108] A "generative AI model" is an artificial intelligence model that generates natural language responses based on input data.

[0109] A "response generation means" is a method or technique for generating a response based on the extracted information.

[0110] "Return mechanism" means the method or process for returning the generated response to the end user's communication device.

[0111] The system of the present invention provides a chatbot that generates sophisticated responses using data with specialized knowledge. Specific embodiments of the system are described in detail below.

[0112] Upload your expertise

[0113] Users use their communication devices to send data files containing specialized knowledge to the server. The data files can be, for example, medical papers, technical manuals, or legal documents. The user clicks the upload button through the interface, selects the file, and then presses the send button. The server temporarily stores the received file and checks whether the file format is appropriate.

[0114] Data analysis and metadata generation

[0115] The server analyzes the temporarily stored data files. Specifically, it uses Apache Tika to extract text data from the file contents. The extracted text data is then analyzed using a natural language processing tool (e.g., NLTK) to identify keywords and important items. Based on the results of this analysis, metadata such as title, author, and classification is generated and stored in a storage device. Generating and storing metadata allows for efficient subsequent data searches.

[0116] Chatbot generation

[0117] Users can customize their chatbot through the interface by selecting the data stored in the memory to use as a knowledge base, a large-scale language model (e.g., GPT-4), and interface settings. Based on these settings, the server builds a chatbot that generates responses based on expert knowledge. This chatbot utilizes generative AI models to generate responses to questions, allowing it to provide highly accurate responses even in situations requiring expert knowledge.

[0118] Enter a question and search

[0119] When an end user inputs a question into the chatbot, the question is sent to the server. The server analyzes the received question and extracts keywords. Natural language processing tools are used here as well. Based on the extracted keywords, related information is searched for in the storage device. A search engine such as Elasticsearch is used for this search.

[0120] Response Generation

[0121] The server generates a response based on the search results using a generative AI model (e.g., GPT-4), which leverages pre-trained models to translate the received data into natural language and construct a response in a way that is easy for the end user to understand.

[0122] Providing a response

[0123] The server generates a response and sends it back to the end user's device. The end user can review the response through an interface. If the response is unsatisfactory, the end user can provide feedback. This feedback is processed by the server and used to improve future responses.

[0124] Specific examples

[0125] Example 1: Medical knowledge-based chatbot

[0126] 1. The user sends medical paper data from a communication device to the server.

[0127] 2. The server analyzes the data, generates metadata, and stores it in a storage device.

[0128] 3. The user configures the "medical chatbot" in the interface.

[0129] 4. When an end user asks, "Tell me about the latest cancer treatments," the server searches for relevant information from its storage device, and the generative AI model generates a response.

[0130] 5. The server generates a response and sends it back to the end user.

[0131] Example 2: Legal advice chatbot

[0132] 1. A user submits a legal document to the server.

[0133] 2. The server analyzes the data, generates metadata, and stores it in a storage device.

[0134] 3. Users customize the "Legal Advice Chatbot."

[0135] 4. When an end user asks, "What legal details are required for a new business contract?", the server retrieves information from relevant documents and a generative AI model generates the appropriate response.

[0136] 5. The server generates a response and sends it back to the end user.

[0137] Prompt Sentence Examples

[0138] "Tell me about the latest cancer treatments"

[0139] "I want to know the latest information on infectious disease prevention measures"

[0140] "Learn more about diabetes treatment"

[0141] As described above, the system of the present invention efficiently utilizes specialized knowledge, allowing users to easily create advanced chatbots that can appropriately respond to advanced questions related to specialized fields.

[0142] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0143] Step 1: Upload your expertise

[0144] Users access the web interface from their communication devices and select the data file containing their specialized knowledge. When the user presses the upload button, the selected file is sent to the server. The server temporarily stores the received file and checks the file format.

[0145] Input: Data file (e.g. PDF, DOCX) containing the expertise submitted by the user.

[0146] Output: The temporarily saved data file is stored on the server.

[0147] Step 2: Data analysis and metadata generation

[0148] The server extracts the contents of the temporarily saved files into text data using Apache Tika. It then analyzes the text data using the Natural Language Toolkit (NLTK) to identify keywords and important items. Based on the analyzed data, it generates metadata such as title, author, and category and stores it in a storage device.

[0149] Input: Data file temporarily stored on the server.

[0150] Output: The parsed text data and generated metadata are stored in a storage device.

[0151] Step 3: Create a chatbot

[0152] Users configure their chatbot through a web interface. They select the knowledge base data to use, the generative AI model (e.g., GPT-4), and interface settings, and generate the chatbot based on those settings. The server then builds the chatbot using the selected settings and connects the generative AI model.

[0153] Input: User-configured knowledge base data, generative AI models, and interface settings.

[0154] Output: The customized chatbot is built on the server and is ready for use.

[0155] Step 4: Enter your question and search

[0156] When an end user types a question into a chatbot, it is sent to a server that analyzes the question, extracts keywords, and then uses a search engine such as Elasticsearch to search for relevant information in storage.

[0157] Input: Questions submitted by end users.

[0158] Output: Keywords are extracted and related information is searched.

[0159] Step 5: Response Generation

[0160] The server generates a response based on the search results using a generative AI model (e.g., GPT-4), which generates a natural language response based on the search results and formats it in a format that is easy for the end user to understand.

[0161] Input: Relevant information identified by the search engine.

[0162] Output: The natural language response generated by the generative AI model.

[0163] Step 6: Providing a response

[0164] The server generates a response and sends it back to the end user's device. The end user receives the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the user can provide feedback. This feedback is processed by the server and used to improve future responses.

[0165] Input: The generated natural language response.

[0166] Output: The response and possible feedback provided to the end user.

[0167] (Application example 1)

[0168] 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."

[0169] Current factory work requires a wide variety of specialized knowledge for technical troubleshooting and maintenance. However, it is not easy for workers to quickly retrieve the appropriate information, which can lead to low work efficiency and mistakes. In addition, current systems do not provide sufficient means to obtain the necessary information in real time. To solve this problem, there is a need for a system that allows workers to easily receive appropriate advice based on specialized knowledge using smart devices.

[0170] 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.

[0171] In this invention, the server includes means for uploading expert knowledge from a user's terminal, means for analyzing the uploaded data to generate metadata and storing it in a database, means for a user to customize the chatbot through a web interface, means for searching for related information from a database in response to an input question, means for generating an appropriate response using a large-scale language model based on the search result, means for returning the generated response to the end user's terminal, and means for an end user to input a question in real time using a smart device and receive a response, thereby enabling workers in a factory to receive appropriate advice based on expert knowledge in real time.

[0172] "Specialized knowledge" refers to advanced information and knowledge in a particular field.

[0173] "User" refers to an individual or organization that uses the system of the present invention.

[0174] A "server" refers to a computer system that transfers, stores, analyzes, and generates responses to data.

[0175] "Terminal" refers to the device that a user uses to access the server, such as a PC, smartphone, or tablet.

[0176] "Data analysis" refers to the process of examining uploaded expert knowledge data from multiple angles to understand its meaning and structure.

[0177] "Metadata" refers to information about uploaded data, including accompanying information such as title, author, and category.

[0178] "Database" refers to a repository where analyzed and generated metadata is stored and can be searched and referenced as needed.

[0179] "Web interface" refers to the screen that allows users to access the system through a web browser and perform various operations.

[0180] A "chatbot" refers to a software program that automatically generates responses to user questions.

[0181] "Entered Question" refers to the inquiry sent by the end user to the system.

[0182] "Large-scale language model" refers to a machine learning model for processing natural language based on a wide range of datasets.

[0183] "Appropriate response" means the most relevant and accurate answer to the entered question.

[0184] "End user" refers to the ultimate user of the chatbot.

[0185] "Smart devices" refer to devices that are connected to the Internet and have advanced functionality, including smart glasses and smartphones.

[0186] "Real time" refers to a time range in which processing occurs immediately with little or no delay.

[0187] The system of the present invention is designed to efficiently perform technical troubleshooting and maintenance in factory operations, allowing users to receive real-time advice based on expert knowledge using their smart devices.

[0188] 1. Upload your expertise

[0189] First, a user uploads a data file containing specialized knowledge (such as a technical manual) from their own device to the server. The upload is done through a web interface, the file format is confirmed, and the file is saved as a temporary file on the server.

[0190] 2. Data analysis and metadata generation

[0191] The server analyzes the uploaded data files and generates metadata such as title, author, category, etc. Natural language processing technology is used for the analysis, and the metadata is stored in a database.

[0192] 3. Chatbot generation

[0193] Next, users customize the chatbot through a web interface, including the database to use, the large-scale language model (LLM), and interface settings. This customization results in a chatbot based on their expertise being built on the server.

[0194] 4. Question and Response Processing

[0195] End users input questions in real time using smart devices (such as smart glasses). These questions are sent to a server, which analyzes the received questions, extracts keywords, and searches for relevant information in a database. Based on the search results, the server uses a large-scale language model (LLM) to generate an appropriate response. This response is then sent back to the end user's smart device in real time.

[0196] 5. Specific Examples

[0197] Take a concrete example of troubleshooting a robotic arm in a factory: A worker puts on smart glasses and asks, "What caused the robotic arm to stop working? What can I do to fix it?" This question is sent to a server, which searches for relevant information in a database and uses a generative AI model to create a prompt like this:

[0198] text

[0199] Document name: maintenance_manual.txt

[0200] Q: What is the reason for the robot arm stopping and how can I fix it?

[0201] answer:

[0202] The server inputs this prompt into a large-scale language model to generate an appropriate response and provide it to the end user.

[0203] The hardware used includes smart devices (e.g., smart glasses) and servers. The software uses Python, the OpenAI API, and natural language processing algorithms. Text analysis algorithms are used for data analysis, tagging techniques are used for metadata generation, and large-scale language models (e.g., text-davinci-003) are used for response generation.

[0204] In this way, the system of the present invention provides advanced support for quickly and efficiently resolving technical problems in factory operations.

[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 uploads a data file containing expert knowledge from their device to the server. The user uses a web interface to select the data file and presses the send button. At this time, the file format is confirmed and the server temporarily stores the received file. The input is the expert knowledge data file, and the output is the data stored as a temporary file.

[0208] Step 2:

[0209] The server analyzes the uploaded data files and generates metadata. The server performs data analysis such as text extraction, keyword identification, and extraction of important items, and generates metadata such as title, author, and category. The input is data saved as a temporary file, and the output is the generated metadata.

[0210] Step 3:

[0211] The server stores the generated metadata in a database. The analyzed data and metadata are appropriately organized and saved in the database. The input is the generated metadata, and the output is the metadata stored in the database.

[0212] Step 4:

[0213] Users customize chatbots through a web interface. They select the database to use as a knowledge base, the large-scale language model (LLM), interface settings, and so on. This allows a chatbot based on expert knowledge to be built on the server. The input is customization setting information, and the output is the customized chatbot.

[0214] Step 5:

[0215] End users input questions in real time using smart devices. The questions are sent to the server from devices such as smart glasses or smartphones. The input is the question from the end user, and the output is the question sent to the server.

[0216] Step 6:

[0217] The server analyzes the received question, extracts keywords, and searches the database for related information. The server analyzes the question text and identifies keywords. It then searches for related entries in the database. The input is the received question, and the output is the related entries.

[0218] Step 7:

[0219] The server generates a response based on the search results using a large-scale language model (LLM). A prompt sentence is generated and input to a generative AI model (e.g., text-davinci-003) to get an appropriate response. The input is the relevant entry and the prompt sentence, and the output is the generated response. An example prompt sentence is:

[0220] text

[0221] Document name: maintenance_manual.txt

[0222] Q: What is the reason for the robot arm stopping and how can I fix it?

[0223] answer:

[0224] Step 8:

[0225] The server sends the generated response back to the end user's smart device, which receives the response through smart glasses or a smartphone. The input is the generated response, and the output is the response sent back to the smart device.

[0226] 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.

[0227] The system of the present invention combines an advanced chatbot that utilizes specialized knowledge with an emotion engine that recognizes the user's emotions. Specific embodiments of the system will be described below in natural language.

[0228] Upload your expertise

[0229] Users upload specialized document files from their own devices to the server. Examples include medical papers, technical manuals, and legal documents. Using a web interface, users select the files they want to upload and press the send button to begin the process. At this point, the file format is verified and saved as a temporary file on the server.

[0230] Data analysis and metadata generation

[0231] The server receives the uploaded data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items. As a result of the analysis, metadata such as the file title, author, and category is generated. The metadata is organized and stored in a database for search engine compatibility.

[0232] Chatbot generation

[0233] Users can customize their chatbot through a web interface, selecting the database they want to use as their knowledge base, the large-scale language model (LLM) they want to use, and interface settings, which then builds a chatbot on the server that generates responses based on their expert knowledge.

[0234] Emotion Engine Operation

[0235] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. This emotion engine uses natural language processing technology to analyze the emotional state of the user's input.

[0236] Enter a question and search

[0237] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the question is sent to the server.

[0238] The server extracts keywords from the received question and searches for related information in the database. Keyword searches and ID searches identify entries related to the question.

[0239] Response Generation

[0240] The server generates a response based on the identified search results using a large-scale language model (LLM). The LLM generates a human-like response based on expert knowledge and interpretation. It also references the user's emotional information, as recognized by the emotion engine, and adjusts the response accordingly. For example, if the user is expressing anger, the tone of the response can be toned down.

[0241] Providing a response

[0242] The server then sends the generated response back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a function to provide feedback, which is used to improve the accuracy of future responses.

[0243] Specific examples

[0244] Example 1: A chatbot using medical knowledge and emotion recognition

[0245] 1. The user uploads medical paper data from their device to the server.

[0246] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0247] 3. The user configures the "medical chatbot" through a web interface.

[0248] 4. When an end user asks, "Tell me about the latest cancer treatments," the comment also includes an emotional comment such as, "I'm worried."

[0249] 5. The server's emotion engine recognizes the user's "concern."

[0250] 6. The server searches the database for relevant information, and the LLM generates a response, softening the tone of the response based on the emotional information.

[0251] 7. The server generates a response and sends it back to the end user.

[0252] Example 2: Legal advice and emotion recognition chatbot

[0253] 1. A user uploads a legal document to the server.

[0254] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0255] 3. Users customize the "Legal Advice Chatbot."

[0256] 4. When an end user asks, "What legal details are required for a new business contract?", the comment also includes an emotional response: "I'm worried."

[0257] 5. The server's emotion engine recognizes the user's "anxiety."

[0258] 6. The server retrieves the information from relevant documents, and the LLM generates an appropriate response, adjusting the tone of the response to a reassuring one based on the emotional information.

[0259] 7. The server generates a response and sends it back to the end user.

[0260] As described above, the system of the present invention not only efficiently utilizes specialized knowledge, but also recognizes the user's emotions and adjusts the response content, thereby realizing more personalized dialogue.

[0261] The processing flow will be explained below.

[0262] Step 1:

[0263] Users upload document files containing their specialized knowledge from their own devices to the server. Using a web interface, users select the files to upload and press the send button. At this time, the file format is checked and saved as a temporary file on the server.

[0264] Step 2:

[0265] The server receives the received data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting key points. Specifically, the text is analyzed using natural language processing tools to extract key keywords and phrases.

[0266] Step 3:

[0267] The server generates metadata such as file title, author, and category from the analysis results, which are then organized and stored in a database for search engine compatibility.

[0268] Step 4:

[0269] Users customize their chatbot through a web interface, selecting the database they want to use as their knowledge base, the large-scale language model (LLM) they want to use, and interface settings—for example, choosing a medical database and the latest LLM.

[0270] Step 5:

[0271] The server integrates the configured knowledge base with the LLM model and builds a chatbot based on the specified settings, resulting in a chatbot that can respond with expert knowledge.

[0272] Step 6:

[0273] When a user or end-user types a question into the chatbot, they can include a comment related to their feelings along with the question, for example, "Tell me about this new treatment, but I'm a little nervous."

[0274] Step 7:

[0275] The server's emotion engine analyzes the user's input and identifies the emotion, for example, "anxiety," using natural language processing tools. This emotion information is then used in the response generation process.

[0276] Step 8:

[0277] The server extracts keywords from the received question and searches for related information in its database, performing keyword and ID searches to identify entries related to the question.

[0278] Step 9:

[0279] The server generates a response based on the identified search results using a large-scale language model (LLM). It also references the emotional information recognized by the emotion engine and adjusts the response accordingly. For example, it takes into account the emotional information of "anxiety" and makes the tone of the response gentler.

[0280] Step 10:

[0281] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application.

[0282] Step 11:

[0283] The user or end user can provide feedback on the returned responses, which the server collects and uses to improve future response accuracy and emotion recognition.

[0284] Specific examples

[0285] The user uploads medical paper data from their device to the server and sets up the chatbot. When the end user then asks, "Please tell me about the latest cancer treatments. I'm anxious," the server recognizes "anxiety" using its emotion engine, searches for information from the database, and uses LLM to generate a response in a gentle tone and send it back to the end user.

[0286] In this way, the system of the present invention leverages emotion recognition capabilities to deliver sophisticated, personalized responses based on expert knowledge.

[0287] Example 2

[0288] 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."

[0289] Conventional chatbot systems have difficulty efficiently utilizing expert knowledge and are unable to generate personalized responses that take into account the user's emotional state. Furthermore, they lack the ability to incorporate end-user feedback into the generated responses. Therefore, there is a need for a system that utilizes expert knowledge while providing appropriate responses that reflect the user's emotions.

[0290] 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.

[0291] In this invention, the server includes means for uploading expert knowledge from a user's terminal to the server, means for analyzing the uploaded data to generate metadata and storing it in a database, means for a user to customize the chatbot through a web interface, means for searching the database for related information in response to an input question, means for generating an appropriate response using a large-scale language model based on the search result, means for analyzing the user's emotions and adjusting the response content based on the emotional information, and means for returning the generated response to the end user's terminal. This makes it possible to generate personalized responses according to the user's emotional state and reflect the end user's feedback while efficiently utilizing expert knowledge.

[0292] "Expertise" refers to a high level of information or understanding gained in a particular field.

[0293] "User" refers to the entity that uses the system to upload specialized knowledge and customize the chatbot.

[0294] "Terminal" is a general term for computing devices and communication equipment used by users, including personal computers, smartphones, tablets, etc.

[0295] A "server" is a computing device that provides functions such as uploading specialized knowledge, analyzing data, generating metadata, building chatbots, and generating and returning responses.

[0296] The "emotion engine" is a module that uses natural language processing technology to analyze the emotional state of text input from users.

[0297] "Metadata" is data that describes information about the contents of a file, including the title, author, category, etc.

[0298] A "database" is a collection of data in which specialized knowledge and metadata are systematically stored and can be searched and used.

[0299] "Web interface" refers to the screens and input forms that allow users to interact with a system via a browser.

[0300] A "chatbot" is a program that uses large-scale language models to enable natural dialogue in response to user questions.

[0301] A "large-scale language model (LLM)" is a language model trained on large amounts of text data, and is capable of generating natural-sounding sentences like humans.

[0302] A "response" is the information or advice returned by the chatbot to the end user.

[0303] "Feedback" means comments or ratings provided by end users that are used to improve system performance and responsiveness.

[0304] "End user" refers to the final user who actually receives the response from the chatbot.

[0305] The system of the present invention provides an advanced chatbot that utilizes expert knowledge, recognizes user emotions, and adjusts its responses accordingly. The system includes a function for uploading expert knowledge from a user's device to a server, analyzing the uploaded data to generate metadata, and storing it in a database. The system then allows the user to customize the chatbot through a web interface, searches the database for relevant information in response to an input question, and generates an appropriate response using a large-scale language model (LLM) based on the search results. The system also includes a function for analyzing user emotions, adjusting the response based on the emotional information, and returning the generated response to the end user's device.

[0306] Specifically, the system works as follows:

[0307] Users upload document files containing specialized knowledge from their own devices to the server. For example, medical papers, technical manuals, legal documents, etc. The user uses a web interface to select the file to upload and press the send button. At this time, the file format is confirmed and saved as a temporary file on the server.

[0308] The server receives the uploaded data files and analyzes their contents. During this analysis process, text is extracted using Apache Tika, and natural language processing libraries such as NLTK and SpaCy are used to identify keywords and extract important information. The analysis results in metadata such as the file title, author, and category. The metadata is organized so that it can be used by search engines, and stored in a database (e.g., PostgreSQL).

[0309] Users customize their chatbot through a web interface, selecting the database they want to use as a knowledge base, a large-scale language model (e.g., OpenAI's GPT-4), and interface settings. This builds a chatbot on the server that generates responses based on their expert knowledge.

[0310] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. The emotion engine uses natural language processing techniques such as TextBlob and VADER to analyze the emotional state of the user's input.

[0311] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the question is sent to the server.

[0312] The server extracts keywords from the received question and searches for related information in the database. Keyword searches and ID searches identify entries related to the question.

[0313] The server generates a response based on the identified search results using a large-scale language model (LLM). The LLM generates a human-like response based on expert knowledge and interpretation. It also references the user's emotional information, as recognized by the emotion engine, and adjusts the response accordingly. For example, if the user is expressing anger, the tone of the response can be toned down.

[0314] The server then sends the generated response back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a function to provide feedback, which is used to improve the accuracy of future responses.

[0315] As a concrete example, consider the following scenario:

[0316] Example 1: A chatbot using medical knowledge and emotion recognition

[0317] 1. The user uploads medical paper data from their device to the server.

[0318] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0319] 3. The user configures the "medical chatbot" through a web interface.

[0320] 4. When an end user asks, "Tell me about the latest cancer treatments," the comment also includes an emotional comment such as, "I'm worried."

[0321] 5. The server's emotion engine recognizes the user's "concern."

[0322] 6. The server searches the database for relevant information, and the LLM generates a response, softening the tone of the response based on the emotional information.

[0323] 7. The server generates a response and sends it back to the end user.

[0324] Example 2: Legal advice and emotion recognition chatbot

[0325] 1. A user uploads a legal document to the server.

[0326] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0327] 3. Users customize the "Legal Advice Chatbot."

[0328] 4. When an end user asks, "What legal details are required for a new business contract?", the comment also includes an emotional response: "I'm worried."

[0329] 5. The server's emotion engine recognizes the user's "anxiety."

[0330] 6. The server retrieves the information from relevant documents, and the LLM generates an appropriate response, adjusting the tone of the response to a reassuring one based on the emotional information.

[0331] 7. The server generates a response and sends it back to the end user.

[0332] In this way, it is possible to efficiently utilize expert knowledge, recognize user emotions, and tailor responses to achieve more personalized interactions.

[0333] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0334] Step 1:

[0335] The user selects a document file containing the specialized knowledge and uploads it to the server from the terminal.

[0336] What happens: The user opens a file selection dialog in the web interface, selects a file to upload, and clicks the "Submit" button.

[0337] Input: A file containing expert knowledge stored on the user's device.

[0338] Output: The document file sent to the server.

[0339] Step 2:

[0340] The server checks the format of the received file and saves it as a temporary file.

[0341] Specific operation: The server verifies the format of the received file and, if it is a compatible format (e.g., PDF, DOCX), saves the file in a temporary folder.

[0342] Input: The document file submitted by the user.

[0343] Output: A temporary file saved on the server.

[0344] Step 3:

[0345] The server analyzes the saved file with a text extraction tool to extract text data.

[0346] What it does: The server uses Apache Tika to extract text from the file.

[0347] Input: A temporary file stored on the server.

[0348] Output: The extracted text data.

[0349] Step 4:

[0350] The server extracts keywords and important items from the analyzed text data.

[0351] How it works: The server uses natural language processing libraries such as NLTK and SpaCy to extract keywords and important items from text data.

[0352] Input: Extracted text data.

[0353] Output: Extracted keywords and important items.

[0354] Step 5:

[0355] The server generates metadata such as file title, author, category, etc. and stores it in a database.

[0356] Specific operation: The server constructs metadata based on the extracted information and stores it in a database (e.g., PostgreSQL).

[0357] Input: Extracted keywords and important items.

[0358] Output: Metadata stored in a database.

[0359] Step 6:

[0360] Users customize their chatbot through a web interface.

[0361] Specific behavior: The user selects the database to use as the knowledge base, the large language model (LLM), and the interface settings in the web interface.

[0362] Input: User customization settings.

[0363] Output: Customized chatbot configuration information.

[0364] Step 7:

[0365] The server builds a chatbot based on the selected configuration information.

[0366] Specific operation: The server uses the selected LLM (e.g., OpenAI's GPT-4) to build the chatbot's interface and response generation logic.

[0367] Input: Customized chatbot configuration information.

[0368] Output: The built chatbot.

[0369] Step 8:

[0370] The server uses an emotion engine to analyze emotions from the end user's input and records the results.

[0371] Specific operation: The server uses natural language processing libraries such as TextBlob and VADER to evaluate the emotional state of the input text and stores the results in internal data.

[0372] Input: End user input text.

[0373] Output: Parsed emotion data.

[0374] Step 9:

[0375] The server receives the end user's question and searches the question content in a database based on the analyzed emotional state.

[0376] How it works: The server uses NLTK or SpaCy to extract keywords from the question and search for relevant information in a database.

[0377] Input: End-user question and parsed sentiment data.

[0378] Output: The relevant information found.

[0379] Step 10:

[0380] The server generates responses based on the search results using a large-scale language model (LLM) and adjusts the tone based on sentiment data.

[0381] What it does: The server uses LLM to generate natural-sounding responses based on search results and sentiment analysis data, adjusting tone as needed.

[0382] Input: Retrieved relevant information and sentiment data.

[0383] Output: The generated response.

[0384] Step 11:

[0385] The server then sends the generated response back to the end user's terminal.

[0386] Specific behavior: The server sends the generated response as an HTTP response.

[0387] Input: The generated response.

[0388] Output: The response sent back to the end user.

[0389] Step 12:

[0390] End users receive responses and provide feedback through a web interface or dedicated application.

[0391] What happens: The end user reviews the response they receive and submits their rating and comments through the feedback form.

[0392] Input: Generated responses and end-user feedback.

[0393] Output: Feedback data stored on the server.

[0394] This enables the system to efficiently utilize specialized knowledge while creating an advanced chatbot that responds to the user's emotions.

[0395] (Application example 2)

[0396] 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."

[0397] Conventional chatbot systems simply provide factual information without considering the user's emotions. As a result, they are unable to respond appropriately even when the user is feeling anxious or fatigued, resulting in a poor user experience. The present invention aims to achieve more personalized dialogue by analyzing the user's emotions and providing responses that correspond to those emotions.

[0398] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for uploading specialized knowledge from a user's terminal to the server, means for analyzing the uploaded data to generate metadata and storing it in a database, means for the user to customize the chatbot through a web interface, means for searching for related information from a database for an input question, means for generating an appropriate response using a large-scale language model based on the search results, means for analyzing the user's emotions using an emotion engine, means for adjusting the tone of the response based on the emotion information, and means for returning the generated response to the end user's terminal. This makes it possible to provide an appropriate response according to the user's emotions.

[0399] "Expertise" is knowledge based on advanced information and experience in a particular field.

[0400] A "server" is a computer system that provides data and services to multiple client computers over a network.

[0401] "Upload" is the operation of transferring a data file from a client's computer to a server.

[0402] "Data analysis" is the process of organizing collected data and converting it into meaningful information.

[0403] "Metadata" is supplementary information that describes the content, structure, characteristics, etc. of data.

[0404] A "database" is an organized collection of data, a system that allows for efficient searching and manipulation.

[0405] "Web interface" is a collective term for web-based screens and functions that allow users to interact with a system over the Internet.

[0406] A "chatbot" is a program that uses natural language processing to automate interactions with users.

[0407] A "large-scale language model (LLM)" is a deep learning model trained on large amounts of text data that has the ability to understand and generate natural language.

[0408] An "emotion engine" is a technology that analyzes emotions from a user's statements and input and identifies their state.

[0409] "Tone of response" refers to the impression and nuance of emotion and expression contained in the generated reply.

[0410] The system embodying this invention uses user expertise to generate a chatbot, and also recognizes the user's emotions and adjusts responses. The system is composed of the following components:

[0411] Upload your expertise

[0412] The user uploads a document file containing specialized knowledge from their device to the server. The uploaded file is first checked for its format on the server, and if it is in the correct format, it is saved as a temporary file.

[0413] Data analysis and metadata generation

[0414] The server analyzes the uploaded data and generates metadata, including file title, author, category, etc. The generated metadata is stored in a database. Natural language processing (NLP) technology is used for data analysis.

[0415] Chatbot generation

[0416] Users can customize their chatbot through a web interface, selecting the database they want to use as a knowledge base and the large-scale language model (LLM) they want to use, such as GPT-3. Based on these settings, the server generates the chatbot.

[0417] Emotion Engine Operation

[0418] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. The emotion engine uses natural language processing technology to analyze the emotional state of the user's input. For example, it identifies emotions such as "worry."

[0419] Enter a question and search

[0420] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the server extracts keywords from the question and searches for relevant information in the database, using keyword or ID search.

[0421] Response Generation

[0422] The server uses a large-scale language model (LLM) based on the search results to generate an appropriate response. It also refers to the user's emotional information recognized by the emotion engine and adjusts the tone of the response accordingly. For example, if the user indicates "fatigue," the server adjusts the response to be more refreshing.

[0423] Providing a response

[0424] The server then sends the generated response back to the end user's device, who receives it through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a feature that allows the user to provide feedback, which is used to improve the accuracy of future responses.

[0425] Specific examples

[0426] As an example, consider the case where a user uploads a document about food to the server and configures a "food delivery helper" chatbot based on that information. The end user asks, "What's your lunch recommendation today?" and adds a comment containing the emotion, "I'm a little tired." In this case, the server's emotion engine recognizes the emotion and suggests a refreshing lunch menu.

[0427] Prompt Sentence Examples

[0428] "Question: What's your recommended lunch today?

[0429] Emotions: Tired

[0430] Metadata: Nutrition research papers, popular recipes

[0431] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0432] Step 1:

[0433] The user uploads a document file containing specialized knowledge from the terminal to the server. The file selected by the user is then sent to the server. The server checks the format of the uploaded file and saves it as a temporary file. The input is the file containing specialized knowledge, and the output is the saved temporary file.

[0434] Step 2:

[0435] The server analyzes the uploaded data and generates metadata. The analysis process involves extracting text, identifying keywords, and extracting important items. From the analysis results, metadata such as title, author, and category is generated and stored in a database. The input is the saved temporary file, and the output is a database entry containing the metadata.

[0436] Step 3:

[0437] Users customize their chatbots through a web interface. They configure the database to use, the large-scale language model (LLM) to use, and the chatbot's interface. The server generates the chatbot based on this configuration. The input is the user's customization settings, and the output is the customized chatbot.

[0438] Step 4:

[0439] An end user inputs questions and comments to a chatbot, which may include comments that indicate emotions. The input is the user's questions and comments.

[0440] Step 5:

[0441] The server uses an emotion engine to analyze emotions from end-user questions and comments, and identifies emotional states using natural language processing techniques. The input is the end-user's questions and comments, and the output is the identified emotion data.

[0442] Step 6:

[0443] The server searches for related information based on the question and emotion data. It identifies related information in a database using keyword or ID search. The input is the identified emotion data and the question, and the output is a database entry containing the related information.

[0444] Step 7:

[0445] The server generates an appropriate response using a large-scale language model (LLM) based on the information retrieved from the database, and adjusts the tone of the response by referring to the emotional data. The input is relevant information and emotional data, and the output is a response tailored according to the emotion.

[0446] Step 8:

[0447] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. The input is the adjusted response, and the output is the response displayed on the end user's device.

[0448] Step 9:

[0449] If the end user provides feedback on a response, the server receives that feedback and uses it to improve the accuracy of future responses. The input is the end user's feedback, and the output is an improved response generation algorithm.

[0450] 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.

[0451] 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.

[0452] 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.

[0453] [Second embodiment]

[0454] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0455] 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.

[0456] 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).

[0457] 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.

[0458] 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.

[0459] 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).

[0460] 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.

[0461] 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.

[0462] 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.

[0463] 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.

[0464] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0465] 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."

[0466] The system of the present invention provides a chatbot that generates intelligent responses using data with specialized knowledge. Specific embodiments of the system are described below in natural language.

[0467] Upload your expertise

[0468] Users first upload data files containing specialized knowledge from their own devices to the server. Examples include medical papers, technical manuals, and legal documents. Using a web interface, users select the files they want to upload and press the send button to begin the process. At this point, the file format is verified and saved as a temporary file on the server.

[0469] Data analysis and metadata generation

[0470] The server receives the uploaded data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items. As a result of the analysis, metadata such as the file title, author, and category is generated. This metadata is organized and stored in a database for use by search engines.

[0471] Chatbot generation

[0472] Users can customize their chatbot through a web interface, selecting the database they want to use as a knowledge base, the large-scale language model (LLM) they want to use, and interface settings, and then a chatbot is built on the server that generates responses based on their expert knowledge.

[0473] Enter a question and search

[0474] When an end user types a question into the chatbot, the question is sent to the server, which extracts keywords from the question and searches for related information in the database. Entries related to the question are identified through keyword and ID searches.

[0475] Response Generation

[0476] The server generates a response based on the search results using a large-scale language model (LLM). The LLM interprets the information based on expert knowledge and composes it in natural language, presenting it in a format that is easy for the end user to understand. For example, in response to a question such as "What are the latest cancer treatments?", the server searches the database for the latest medical papers, and the LLM generates an appropriate response based on their content.

[0477] Providing a response

[0478] Finally, the server sends the generated response back to the end user's device. The end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has the ability to provide feedback, which is used to improve the accuracy of future responses.

[0479] Specific examples

[0480] Example 1: Medical knowledge-based chatbot

[0481] 1. The user uploads medical paper data from their device to the server.

[0482] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0483] 3. The user configures the "medical chatbot" through a web interface.

[0484] 4. When an end user asks, "Tell me about the latest cancer treatments," the server searches the database for relevant information and the LLM generates a response.

[0485] 5. The server generates a response and sends it back to the end user.

[0486] Example 2: Legal advice chatbot

[0487] 1. A user uploads a legal document to the server.

[0488] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0489] 3. Users customize the "Legal Advice Chatbot."

[0490] 4. When an end user asks, "What legal details are required for a new business contract?", the server retrieves the information from the relevant documents and the LLM generates the appropriate response.

[0491] 5. The server generates a response and sends it back to the end user.

[0492] As a result, the system of the present invention efficiently utilizes specialized knowledge, enabling users to easily create advanced chatbots and appropriately respond to advanced questions related to specialized fields.

[0493] The processing flow will be explained below.

[0494] Step 1:

[0495] Users upload document files containing their specialized knowledge from their own devices to the server. Using a web interface, users select the files to upload and press the send button. At this time, the file format is confirmed and saved as a temporary file on the server.

[0496] Step 2:

[0497] The server receives the received data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items.

[0498] Step 3:

[0499] The server generates metadata such as file title, author, category, etc. from the analysis results, which are then organized and stored in a database for search engine compatibility.

[0500] Step 4:

[0501] Users configure the chatbot generation through a web interface, choosing the database they want to use as a knowledge base, the large-scale language model (LLM) they want to use, and interface settings to customize the chatbot.

[0502] Step 5:

[0503] The server receives the user's configuration and builds a chatbot that integrates the specified knowledge base and LLM model.

[0504] Step 6:

[0505] The user or end user types a question for the chatbot, which is then sent to the server.

[0506] Step 7:

[0507] The server analyzes the received question and extracts keywords. The extracted keywords are used to search for related information in the database. Entries related to the question are identified through keyword and ID searches.

[0508] Step 8:

[0509] The server generates a response based on the identified search results using a large-scale language model (LLM), which interprets the information based on expert knowledge and formats it in natural language in a format that is easy for the end user to understand.

[0510] Step 9:

[0511] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application.

[0512] Step 10:

[0513] If the response does not meet expectations, users can provide feedback, which is aggregated on the server and used to improve the accuracy of future responses.

[0514] Example 1

[0515] 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."

[0516] Current chatbot systems face the challenge of being unable to efficiently generate responses that incorporate specialized knowledge. Another problem is that it takes a lot of time and effort for users to upload and customize the knowledge base. In particular, when dealing with specialized knowledge, systems that can provide reliable responses instantly are required, but current technology is unable to adequately meet this demand.

[0517] 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.

[0518] In this invention, the server includes means for transmitting specialized knowledge from a user's communication device to the server, means for analyzing the transmitted data to generate metadata and storing it in a storage device, and means for the user to configure an information processing device through an interface, thereby enabling efficient construction of an advanced chatbot based on specialized knowledge and instantaneous generation of specialized responses.

[0519] "Specialized knowledge" is information that includes advanced knowledge and skills in a specific field.

[0520] "Users" are individuals or businesses that use the system to provide their expertise to the chatbot.

[0521] "Communication equipment" refers to a device for sending and receiving data via the Internet, including personal computers and smartphones.

[0522] A "server" is a computer system that receives, processes, and stores data submitted by users.

[0523] "Transmission means" means the method or process for transmitting data from the user's communication device to the server.

[0524] "Analysis means" refers to the method or technology by which the server analyzes the specialized knowledge and generates the metadata.

[0525] "Metadata" is information used to identify and organize the content of data, including title, author, category, etc.

[0526] "Storage device" refers to a device or system for storing analyzed data and metadata.

[0527] "Interface" means the user interface through which a user interacts with the system, including a web browser or dedicated application.

[0528] An "information processing device" is a program and hardware for processing data, such as a chatbot or search engine customized by the user.

[0529] An "enquiry" is a question or request that an end user types into a chatbot.

[0530] An "extraction means" is a method or technique for selecting relevant information from a storage device in response to a query.

[0531] A "generative AI model" is an artificial intelligence model that generates natural language responses based on input data.

[0532] A "response generation means" is a method or technique for generating a response based on the extracted information.

[0533] "Return mechanism" means the method or process for returning the generated response to the end user's communication device.

[0534] The system of the present invention provides a chatbot that generates sophisticated responses using data with specialized knowledge. Specific embodiments of the system are described in detail below.

[0535] Upload your expertise

[0536] Users use their communication devices to send data files containing specialized knowledge to the server. The data files can be, for example, medical papers, technical manuals, or legal documents. The user clicks the upload button through the interface, selects the file, and then presses the send button. The server temporarily stores the received file and checks whether the file format is appropriate.

[0537] Data analysis and metadata generation

[0538] The server analyzes the temporarily stored data files. Specifically, it uses Apache Tika to extract text data from the file contents. The extracted text data is then analyzed using a natural language processing tool (e.g., NLTK) to identify keywords and important items. Based on the results of this analysis, metadata such as title, author, and classification is generated and stored in a storage device. Generating and storing metadata allows for efficient subsequent data searches.

[0539] Chatbot generation

[0540] Users can customize their chatbot through the interface by selecting the data stored in the memory to use as a knowledge base, a large-scale language model (e.g., GPT-4), and interface settings. Based on these settings, the server builds a chatbot that generates responses based on expert knowledge. This chatbot utilizes generative AI models to generate responses to questions, allowing it to provide highly accurate responses even in situations requiring expert knowledge.

[0541] Enter a question and search

[0542] When an end user inputs a question into the chatbot, the question is sent to the server. The server analyzes the received question and extracts keywords. Natural language processing tools are used here as well. Based on the extracted keywords, related information is searched for in the storage device. A search engine such as Elasticsearch is used for this search.

[0543] Response Generation

[0544] The server generates a response based on the search results using a generative AI model (e.g., GPT-4), which leverages pre-trained models to translate the received data into natural language and construct a response in a way that is easy for the end user to understand.

[0545] Providing a response

[0546] The server generates a response and sends it back to the end user's device. The end user can review the response through an interface. If the response is unsatisfactory, the end user can provide feedback. This feedback is processed by the server and used to improve future responses.

[0547] Specific examples

[0548] Example 1: Medical knowledge-based chatbot

[0549] 1. The user sends medical paper data from a communication device to the server.

[0550] 2. The server analyzes the data, generates metadata, and stores it in a storage device.

[0551] 3. The user configures the "medical chatbot" in the interface.

[0552] 4. When an end user asks, "Tell me about the latest cancer treatments," the server searches for relevant information from its storage device, and the generative AI model generates a response.

[0553] 5. The server generates a response and sends it back to the end user.

[0554] Example 2: Legal advice chatbot

[0555] 1. A user submits a legal document to the server.

[0556] 2. The server analyzes the data, generates metadata, and stores it in a storage device.

[0557] 3. Users customize the "Legal Advice Chatbot."

[0558] 4. When an end user asks, "What legal details are required for a new business contract?", the server retrieves information from relevant documents and a generative AI model generates the appropriate response.

[0559] 5. The server generates a response and sends it back to the end user.

[0560] Prompt Sentence Examples

[0561] "Tell me about the latest cancer treatments"

[0562] "I want to know the latest information on infectious disease prevention measures"

[0563] "Learn more about diabetes treatment"

[0564] As described above, the system of the present invention efficiently utilizes specialized knowledge, allowing users to easily create advanced chatbots that can appropriately respond to advanced questions related to specialized fields.

[0565] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0566] Step 1: Upload your expertise

[0567] Users access the web interface from their communication devices and select the data file containing their specialized knowledge. When the user presses the upload button, the selected file is sent to the server. The server temporarily stores the received file and checks the file format.

[0568] Input: Data file (e.g. PDF, DOCX) containing the expertise submitted by the user.

[0569] Output: The temporarily saved data file is stored on the server.

[0570] Step 2: Data analysis and metadata generation

[0571] The server extracts the contents of the temporarily saved files into text data using Apache Tika. It then analyzes the text data using the Natural Language Toolkit (NLTK) to identify keywords and important items. Based on the analyzed data, it generates metadata such as title, author, and category and stores it in a storage device.

[0572] Input: Data file temporarily stored on the server.

[0573] Output: The parsed text data and generated metadata are stored in a storage device.

[0574] Step 3: Create a chatbot

[0575] Users configure their chatbot through a web interface. They select the knowledge base data to use, the generative AI model (e.g., GPT-4), and interface settings, and generate the chatbot based on those settings. The server then builds the chatbot using the selected settings and connects the generative AI model.

[0576] Input: User-configured knowledge base data, generative AI models, and interface settings.

[0577] Output: The customized chatbot is built on the server and is ready for use.

[0578] Step 4: Enter your question and search

[0579] When an end user types a question into a chatbot, it is sent to a server that analyzes the question, extracts keywords, and then uses a search engine such as Elasticsearch to search for relevant information in storage.

[0580] Input: Questions submitted by end users.

[0581] Output: Keywords are extracted and related information is searched.

[0582] Step 5: Response Generation

[0583] The server generates a response based on the search results using a generative AI model (e.g., GPT-4), which generates a natural language response based on the search results and formats it in a format that is easy for the end user to understand.

[0584] Input: Relevant information identified by the search engine.

[0585] Output: The natural language response generated by the generative AI model.

[0586] Step 6: Providing a response

[0587] The server generates a response and sends it back to the end user's device. The end user receives the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the user can provide feedback. This feedback is processed by the server and used to improve future responses.

[0588] Input: The generated natural language response.

[0589] Output: The response and possible feedback provided to the end user.

[0590] (Application example 1)

[0591] 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."

[0592] Current factory work requires a wide variety of specialized knowledge for technical troubleshooting and maintenance. However, it is not easy for workers to quickly retrieve the appropriate information, which can lead to low work efficiency and mistakes. In addition, current systems do not provide sufficient means to obtain the necessary information in real time. To solve this problem, there is a need for a system that allows workers to easily receive appropriate advice based on specialized knowledge using smart devices.

[0593] 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.

[0594] In this invention, the server includes means for uploading expert knowledge from a user's terminal, means for analyzing the uploaded data to generate metadata and storing it in a database, means for a user to customize the chatbot through a web interface, means for searching for related information from a database in response to an input question, means for generating an appropriate response using a large-scale language model based on the search result, means for returning the generated response to the end user's terminal, and means for an end user to input a question in real time using a smart device and receive a response, thereby enabling workers in a factory to receive appropriate advice based on expert knowledge in real time.

[0595] "Specialized knowledge" refers to advanced information and knowledge in a particular field.

[0596] "User" refers to an individual or organization that uses the system of the present invention.

[0597] A "server" refers to a computer system that transfers, stores, analyzes, and generates responses to data.

[0598] "Terminal" refers to the device that a user uses to access the server, such as a PC, smartphone, or tablet.

[0599] "Data analysis" refers to the process of examining uploaded expert knowledge data from multiple angles to understand its meaning and structure.

[0600] "Metadata" refers to information about uploaded data, including accompanying information such as title, author, and category.

[0601] "Database" refers to a repository where analyzed and generated metadata is stored and can be searched and referenced as needed.

[0602] "Web interface" refers to the screen that allows users to access the system through a web browser and perform various operations.

[0603] A "chatbot" refers to a software program that automatically generates responses to user questions.

[0604] "Entered Question" refers to the inquiry sent by the end user to the system.

[0605] "Large-scale language model" refers to a machine learning model for processing natural language based on a wide range of datasets.

[0606] "Appropriate response" means the most relevant and accurate answer to the entered question.

[0607] "End user" refers to the ultimate user of the chatbot.

[0608] "Smart devices" refer to devices that are connected to the Internet and have advanced functionality, including smart glasses and smartphones.

[0609] "Real time" refers to a time range in which processing occurs immediately with little or no delay.

[0610] The system of the present invention is designed to efficiently perform technical troubleshooting and maintenance in factory operations, allowing users to receive real-time advice based on expert knowledge using their smart devices.

[0611] 1. Upload your expertise

[0612] First, a user uploads a data file containing specialized knowledge (such as a technical manual) from their own device to the server. The upload is done through a web interface, the file format is confirmed, and the file is saved as a temporary file on the server.

[0613] 2. Data analysis and metadata generation

[0614] The server analyzes the uploaded data files and generates metadata such as title, author, category, etc. Natural language processing technology is used for the analysis, and the metadata is stored in a database.

[0615] 3. Chatbot generation

[0616] Next, users customize the chatbot through a web interface, including the database to use, the large-scale language model (LLM), and interface settings. This customization results in a chatbot based on their expertise being built on the server.

[0617] 4. Question and Response Processing

[0618] End users input questions in real time using smart devices (such as smart glasses). These questions are sent to a server, which analyzes the received questions, extracts keywords, and searches for relevant information in a database. Based on the search results, the server uses a large-scale language model (LLM) to generate an appropriate response. This response is then sent back to the end user's smart device in real time.

[0619] 5. Specific Examples

[0620] Take a concrete example of troubleshooting a robotic arm in a factory: A worker puts on smart glasses and asks, "What caused the robotic arm to stop working? What can I do to fix it?" This question is sent to a server, which searches for relevant information in a database and uses a generative AI model to create a prompt like this:

[0621] text

[0622] Document name: maintenance_manual.txt

[0623] Q: What is the reason for the robot arm stopping and how can I fix it?

[0624] answer:

[0625] The server inputs this prompt into a large-scale language model to generate an appropriate response and provide it to the end user.

[0626] The hardware used includes smart devices (e.g., smart glasses) and servers. The software uses Python, the OpenAI API, and natural language processing algorithms. Text analysis algorithms are used for data analysis, tagging techniques are used for metadata generation, and large-scale language models (e.g., text-davinci-003) are used for response generation.

[0627] In this way, the system of the present invention provides advanced support for quickly and efficiently resolving technical problems in factory operations.

[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 uploads a data file containing expert knowledge from their device to the server. The user uses a web interface to select the data file and presses the send button. At this time, the file format is confirmed and the server temporarily stores the received file. The input is the expert knowledge data file, and the output is the data stored as a temporary file.

[0631] Step 2:

[0632] The server analyzes the uploaded data files and generates metadata. The server performs data analysis such as text extraction, keyword identification, and extraction of important items, and generates metadata such as title, author, and category. The input is data saved as a temporary file, and the output is the generated metadata.

[0633] Step 3:

[0634] The server stores the generated metadata in a database. The analyzed data and metadata are appropriately organized and saved in the database. The input is the generated metadata, and the output is the metadata stored in the database.

[0635] Step 4:

[0636] Users customize chatbots through a web interface. They select the database to use as a knowledge base, the large-scale language model (LLM), interface settings, and so on. This allows a chatbot based on expert knowledge to be built on the server. The input is customization setting information, and the output is the customized chatbot.

[0637] Step 5:

[0638] End users input questions in real time using smart devices. The questions are sent to the server from devices such as smart glasses or smartphones. The input is the question from the end user, and the output is the question sent to the server.

[0639] Step 6:

[0640] The server analyzes the received question, extracts keywords, and searches the database for related information. The server analyzes the question text and identifies keywords. It then searches for related entries in the database. The input is the received question, and the output is the related entries.

[0641] Step 7:

[0642] The server generates a response based on the search results using a large-scale language model (LLM). A prompt sentence is generated and input to a generative AI model (e.g., text-davinci-003) to get an appropriate response. The input is the relevant entry and the prompt sentence, and the output is the generated response. An example prompt sentence is:

[0643] text

[0644] Document name: maintenance_manual.txt

[0645] Q: What is the reason for the robot arm stopping and how can I fix it?

[0646] answer:

[0647] Step 8:

[0648] The server sends the generated response back to the end user's smart device, which receives the response through smart glasses or a smartphone. The input is the generated response, and the output is the response sent back to the smart device.

[0649] 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.

[0650] The system of the present invention combines an advanced chatbot that utilizes specialized knowledge with an emotion engine that recognizes the user's emotions. Specific embodiments of the system will be described below in natural language.

[0651] Upload your expertise

[0652] Users upload specialized document files from their own devices to the server. Examples include medical papers, technical manuals, and legal documents. Using a web interface, users select the files they want to upload and press the send button to begin the process. At this point, the file format is verified and saved as a temporary file on the server.

[0653] Data analysis and metadata generation

[0654] The server receives the uploaded data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items. As a result of the analysis, metadata such as the file title, author, and category is generated. The metadata is organized and stored in a database for search engine compatibility.

[0655] Chatbot generation

[0656] Users can customize their chatbot through a web interface, selecting the database they want to use as their knowledge base, the large-scale language model (LLM) they want to use, and interface settings, which then builds a chatbot on the server that generates responses based on their expert knowledge.

[0657] Emotion Engine Operation

[0658] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. This emotion engine uses natural language processing technology to analyze the emotional state of the user's input.

[0659] Enter a question and search

[0660] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the question is sent to the server.

[0661] The server extracts keywords from the received question and searches for related information in the database. Keyword searches and ID searches identify entries related to the question.

[0662] Response Generation

[0663] The server generates a response based on the identified search results using a large-scale language model (LLM). The LLM generates a human-like response based on expert knowledge and interpretation. It also references the user's emotional information, as recognized by the emotion engine, and adjusts the response accordingly. For example, if the user is expressing anger, the tone of the response can be toned down.

[0664] Providing a response

[0665] The server then sends the generated response back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a function to provide feedback, which is used to improve the accuracy of future responses.

[0666] Specific examples

[0667] Example 1: A chatbot using medical knowledge and emotion recognition

[0668] 1. The user uploads medical paper data from their device to the server.

[0669] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0670] 3. The user configures the "medical chatbot" through a web interface.

[0671] 4. When an end user asks, "Tell me about the latest cancer treatments," the comment also includes an emotional comment such as, "I'm worried."

[0672] 5. The server's emotion engine recognizes the user's "concern."

[0673] 6. The server searches the database for relevant information, and the LLM generates a response, softening the tone of the response based on the emotional information.

[0674] 7. The server generates a response and sends it back to the end user.

[0675] Example 2: Legal advice and emotion recognition chatbot

[0676] 1. A user uploads a legal document to the server.

[0677] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0678] 3. Users customize the "Legal Advice Chatbot."

[0679] 4. When an end user asks, "What legal details are required for a new business contract?", the comment also includes an emotional response: "I'm worried."

[0680] 5. The server's emotion engine recognizes the user's "anxiety."

[0681] 6. The server retrieves the information from relevant documents, and the LLM generates an appropriate response, adjusting the tone of the response to a reassuring one based on the emotional information.

[0682] 7. The server generates a response and sends it back to the end user.

[0683] As described above, the system of the present invention not only efficiently utilizes specialized knowledge, but also recognizes the user's emotions and adjusts the response content, thereby realizing more personalized dialogue.

[0684] The processing flow will be explained below.

[0685] Step 1:

[0686] Users upload document files containing their specialized knowledge from their own devices to the server. Using a web interface, users select the files to upload and press the send button. At this time, the file format is checked and saved as a temporary file on the server.

[0687] Step 2:

[0688] The server receives the received data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting key points. Specifically, the text is analyzed using natural language processing tools to extract key keywords and phrases.

[0689] Step 3:

[0690] The server generates metadata such as file title, author, and category from the analysis results, which are then organized and stored in a database for search engine compatibility.

[0691] Step 4:

[0692] Users customize their chatbot through a web interface, selecting the database they want to use as their knowledge base, the large-scale language model (LLM) they want to use, and interface settings—for example, choosing a medical database and the latest LLM.

[0693] Step 5:

[0694] The server integrates the configured knowledge base with the LLM model and builds a chatbot based on the specified settings, resulting in a chatbot that can respond with expert knowledge.

[0695] Step 6:

[0696] When a user or end-user types a question into the chatbot, they can include a comment related to their feelings along with the question, for example, "Tell me about this new treatment, but I'm a little nervous."

[0697] Step 7:

[0698] The server's emotion engine analyzes the user's input and identifies the emotion, for example, "anxiety," using natural language processing tools. This emotion information is then used in the response generation process.

[0699] Step 8:

[0700] The server extracts keywords from the received question and searches for related information in its database, performing keyword and ID searches to identify entries related to the question.

[0701] Step 9:

[0702] The server generates a response based on the identified search results using a large-scale language model (LLM). It also references the emotional information recognized by the emotion engine and adjusts the response accordingly. For example, it takes into account the emotional information of "anxiety" and makes the tone of the response gentler.

[0703] Step 10:

[0704] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application.

[0705] Step 11:

[0706] The user or end user can provide feedback on the returned responses, which the server collects and uses to improve future response accuracy and emotion recognition.

[0707] Specific examples

[0708] The user uploads medical paper data from their device to the server and sets up the chatbot. When the end user then asks, "Please tell me about the latest cancer treatments. I'm anxious," the server recognizes "anxiety" using its emotion engine, searches for information from the database, and uses LLM to generate a response in a gentle tone and send it back to the end user.

[0709] In this way, the system of the present invention leverages emotion recognition capabilities to deliver sophisticated, personalized responses based on expert knowledge.

[0710] Example 2

[0711] 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."

[0712] Conventional chatbot systems have difficulty efficiently utilizing expert knowledge and are unable to generate personalized responses that take into account the user's emotional state. Furthermore, they lack the ability to incorporate end-user feedback into the generated responses. Therefore, there is a need for a system that utilizes expert knowledge while providing appropriate responses that reflect the user's emotions.

[0713] 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.

[0714] In this invention, the server includes means for uploading expert knowledge from a user's terminal to the server, means for analyzing the uploaded data to generate metadata and storing it in a database, means for a user to customize the chatbot through a web interface, means for searching the database for related information in response to an input question, means for generating an appropriate response using a large-scale language model based on the search result, means for analyzing the user's emotions and adjusting the response content based on the emotional information, and means for returning the generated response to the end user's terminal. This makes it possible to generate personalized responses according to the user's emotional state and reflect the end user's feedback while efficiently utilizing expert knowledge.

[0715] "Expertise" refers to a high level of information or understanding gained in a particular field.

[0716] "User" refers to the entity that uses the system to upload specialized knowledge and customize the chatbot.

[0717] "Terminal" is a general term for computing devices and communication equipment used by users, including personal computers, smartphones, tablets, etc.

[0718] A "server" is a computing device that provides functions such as uploading specialized knowledge, analyzing data, generating metadata, building chatbots, and generating and returning responses.

[0719] The "emotion engine" is a module that uses natural language processing technology to analyze the emotional state of text input from users.

[0720] "Metadata" is data that describes information about the contents of a file, including the title, author, category, etc.

[0721] A "database" is a collection of data in which specialized knowledge and metadata are systematically stored and can be searched and used.

[0722] "Web interface" refers to the screens and input forms that allow users to interact with a system via a browser.

[0723] A "chatbot" is a program that uses large-scale language models to enable natural dialogue in response to user questions.

[0724] A "large-scale language model (LLM)" is a language model trained on large amounts of text data, and is capable of generating natural-sounding sentences like humans.

[0725] A "response" is the information or advice returned by the chatbot to the end user.

[0726] "Feedback" means comments or ratings provided by end users that are used to improve system performance and responsiveness.

[0727] "End user" refers to the final user who actually receives the response from the chatbot.

[0728] The system of the present invention provides an advanced chatbot that utilizes expert knowledge, recognizes user emotions, and adjusts its responses accordingly. The system includes a function for uploading expert knowledge from a user's device to a server, analyzing the uploaded data to generate metadata, and storing it in a database. The system then allows the user to customize the chatbot through a web interface, searches the database for relevant information in response to an input question, and generates an appropriate response using a large-scale language model (LLM) based on the search results. The system also includes a function for analyzing user emotions, adjusting the response based on the emotional information, and returning the generated response to the end user's device.

[0729] Specifically, the system works as follows:

[0730] Users upload document files containing specialized knowledge from their own devices to the server. For example, medical papers, technical manuals, legal documents, etc. The user uses a web interface to select the file to upload and press the send button. At this time, the file format is confirmed and saved as a temporary file on the server.

[0731] The server receives the uploaded data files and analyzes their contents. During this analysis process, text is extracted using Apache Tika, and natural language processing libraries such as NLTK and SpaCy are used to identify keywords and extract important information. The analysis results in metadata such as the file title, author, and category. The metadata is organized so that it can be used by search engines, and stored in a database (e.g., PostgreSQL).

[0732] Users customize their chatbot through a web interface, selecting the database they want to use as a knowledge base, a large-scale language model (e.g., OpenAI's GPT-4), and interface settings. This builds a chatbot on the server that generates responses based on their expert knowledge.

[0733] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. The emotion engine uses natural language processing techniques such as TextBlob and VADER to analyze the emotional state of the user's input.

[0734] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the question is sent to the server.

[0735] The server extracts keywords from the received question and searches for related information in the database. Keyword searches and ID searches identify entries related to the question.

[0736] The server generates a response based on the identified search results using a large-scale language model (LLM). The LLM generates a human-like response based on expert knowledge and interpretation. It also references the user's emotional information, as recognized by the emotion engine, and adjusts the response accordingly. For example, if the user is expressing anger, the tone of the response can be toned down.

[0737] The server then sends the generated response back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a function to provide feedback, which is used to improve the accuracy of future responses.

[0738] As a concrete example, consider the following scenario:

[0739] Example 1: A chatbot using medical knowledge and emotion recognition

[0740] 1. The user uploads medical paper data from their device to the server.

[0741] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0742] 3. The user configures the "medical chatbot" through a web interface.

[0743] 4. When an end user asks, "Tell me about the latest cancer treatments," the comment also includes an emotional comment such as, "I'm worried."

[0744] 5. The server's emotion engine recognizes the user's "concern."

[0745] 6. The server searches the database for relevant information, and the LLM generates a response, softening the tone of the response based on the emotional information.

[0746] 7. The server generates a response and sends it back to the end user.

[0747] Example 2: Legal advice and emotion recognition chatbot

[0748] 1. A user uploads a legal document to the server.

[0749] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0750] 3. Users customize the "Legal Advice Chatbot."

[0751] 4. When an end user asks, "What legal details are required for a new business contract?", the comment also includes an emotional response: "I'm worried."

[0752] 5. The server's emotion engine recognizes the user's "anxiety."

[0753] 6. The server retrieves the information from relevant documents, and the LLM generates an appropriate response, adjusting the tone of the response to a reassuring one based on the emotional information.

[0754] 7. The server generates a response and sends it back to the end user.

[0755] In this way, it is possible to efficiently utilize expert knowledge, recognize user emotions, and tailor responses to achieve more personalized interactions.

[0756] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0757] Step 1:

[0758] The user selects a document file containing the specialized knowledge and uploads it to the server from the terminal.

[0759] What happens: The user opens a file selection dialog in the web interface, selects a file to upload, and clicks the "Submit" button.

[0760] Input: A file containing expert knowledge stored on the user's device.

[0761] Output: The document file sent to the server.

[0762] Step 2:

[0763] The server checks the format of the received file and saves it as a temporary file.

[0764] Specific operation: The server verifies the format of the received file and, if it is a compatible format (e.g., PDF, DOCX), saves the file in a temporary folder.

[0765] Input: The document file submitted by the user.

[0766] Output: A temporary file saved on the server.

[0767] Step 3:

[0768] The server analyzes the saved file with a text extraction tool to extract text data.

[0769] What it does: The server uses Apache Tika to extract text from the file.

[0770] Input: A temporary file stored on the server.

[0771] Output: The extracted text data.

[0772] Step 4:

[0773] The server extracts keywords and important items from the analyzed text data.

[0774] How it works: The server uses natural language processing libraries such as NLTK and SpaCy to extract keywords and important items from text data.

[0775] Input: Extracted text data.

[0776] Output: Extracted keywords and important items.

[0777] Step 5:

[0778] The server generates metadata such as file title, author, category, etc. and stores it in a database.

[0779] Specific operation: The server constructs metadata based on the extracted information and stores it in a database (e.g., PostgreSQL).

[0780] Input: Extracted keywords and important items.

[0781] Output: Metadata stored in a database.

[0782] Step 6:

[0783] Users customize their chatbot through a web interface.

[0784] Specific behavior: The user selects the database to use as the knowledge base, the large language model (LLM), and the interface settings in the web interface.

[0785] Input: User customization settings.

[0786] Output: Customized chatbot configuration information.

[0787] Step 7:

[0788] The server builds a chatbot based on the selected configuration information.

[0789] Specific operation: The server uses the selected LLM (e.g., OpenAI's GPT-4) to build the chatbot's interface and response generation logic.

[0790] Input: Customized chatbot configuration information.

[0791] Output: The built chatbot.

[0792] Step 8:

[0793] The server uses an emotion engine to analyze emotions from the end user's input and records the results.

[0794] Specific operation: The server uses natural language processing libraries such as TextBlob and VADER to evaluate the emotional state of the input text and stores the results in internal data.

[0795] Input: End user input text.

[0796] Output: Parsed emotion data.

[0797] Step 9:

[0798] The server receives the end user's question and searches the question content in a database based on the analyzed emotional state.

[0799] How it works: The server uses NLTK or SpaCy to extract keywords from the question and search for relevant information in a database.

[0800] Input: End-user question and parsed sentiment data.

[0801] Output: The relevant information found.

[0802] Step 10:

[0803] The server generates responses based on the search results using a large-scale language model (LLM) and adjusts the tone based on sentiment data.

[0804] What it does: The server uses LLM to generate natural-sounding responses based on search results and sentiment analysis data, adjusting tone as needed.

[0805] Input: Retrieved relevant information and sentiment data.

[0806] Output: The generated response.

[0807] Step 11:

[0808] The server then sends the generated response back to the end user's terminal.

[0809] Specific behavior: The server sends the generated response as an HTTP response.

[0810] Input: The generated response.

[0811] Output: The response sent back to the end user.

[0812] Step 12:

[0813] End users receive responses and provide feedback through a web interface or dedicated application.

[0814] What happens: The end user reviews the response they receive and submits their rating and comments through the feedback form.

[0815] Input: Generated responses and end-user feedback.

[0816] Output: Feedback data stored on the server.

[0817] This enables the system to efficiently utilize specialized knowledge while creating an advanced chatbot that responds to the user's emotions.

[0818] (Application example 2)

[0819] 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."

[0820] Conventional chatbot systems simply provide factual information without considering the user's emotions. As a result, they are unable to respond appropriately even when the user is feeling anxious or fatigued, resulting in a poor user experience. The present invention aims to achieve more personalized dialogue by analyzing the user's emotions and providing responses that correspond to those emotions.

[0821] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for uploading specialized knowledge from a user's terminal to the server, means for analyzing the uploaded data to generate metadata and storing it in a database, means for the user to customize the chatbot through a web interface, means for searching for related information from a database for an input question, means for generating an appropriate response using a large-scale language model based on the search results, means for analyzing the user's emotions using an emotion engine, means for adjusting the tone of the response based on the emotion information, and means for returning the generated response to the end user's terminal. This makes it possible to provide an appropriate response according to the user's emotions.

[0822] "Expertise" is knowledge based on advanced information and experience in a particular field.

[0823] A "server" is a computer system that provides data and services to multiple client computers over a network.

[0824] "Upload" is the operation of transferring a data file from a client's computer to a server.

[0825] "Data analysis" is the process of organizing collected data and converting it into meaningful information.

[0826] "Metadata" is supplementary information that describes the content, structure, characteristics, etc. of data.

[0827] A "database" is an organized collection of data, a system that allows for efficient searching and manipulation.

[0828] "Web interface" is a collective term for web-based screens and functions that allow users to interact with a system over the Internet.

[0829] A "chatbot" is a program that uses natural language processing to automate interactions with users.

[0830] A "large-scale language model (LLM)" is a deep learning model trained on large amounts of text data that has the ability to understand and generate natural language.

[0831] An "emotion engine" is a technology that analyzes emotions from a user's statements and input and identifies their state.

[0832] "Tone of response" refers to the impression and nuance of emotion and expression contained in the generated reply.

[0833] The system embodying this invention uses user expertise to generate a chatbot, and also recognizes the user's emotions and adjusts responses. The system is composed of the following components:

[0834] Upload your expertise

[0835] The user uploads a document file containing specialized knowledge from their device to the server. The uploaded file is first checked for its format on the server, and if it is in the correct format, it is saved as a temporary file.

[0836] Data analysis and metadata generation

[0837] The server analyzes the uploaded data and generates metadata, including file title, author, category, etc. The generated metadata is stored in a database. Natural language processing (NLP) technology is used for data analysis.

[0838] Chatbot generation

[0839] Users can customize their chatbot through a web interface, selecting the database they want to use as a knowledge base and the large-scale language model (LLM) they want to use, such as GPT-3. Based on these settings, the server generates the chatbot.

[0840] Emotion Engine Operation

[0841] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. The emotion engine uses natural language processing technology to analyze the emotional state of the user's input. For example, it identifies emotions such as "worry."

[0842] Enter a question and search

[0843] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the server extracts keywords from the question and searches for relevant information in the database, using keyword or ID search.

[0844] Response Generation

[0845] The server uses a large-scale language model (LLM) based on the search results to generate an appropriate response. It also refers to the user's emotional information recognized by the emotion engine and adjusts the tone of the response accordingly. For example, if the user indicates "fatigue," the server adjusts the response to be more refreshing.

[0846] Providing a response

[0847] The server then sends the generated response back to the end user's device, who receives it through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a feature that allows the user to provide feedback, which is used to improve the accuracy of future responses.

[0848] Specific examples

[0849] As an example, consider the case where a user uploads a document about food to the server and configures a "food delivery helper" chatbot based on that information. The end user asks, "What's your lunch recommendation today?" and adds a comment containing the emotion, "I'm a little tired." In this case, the server's emotion engine recognizes the emotion and suggests a refreshing lunch menu.

[0850] Prompt Sentence Examples

[0851] "Question: What's your recommended lunch today?

[0852] Emotions: Tired

[0853] Metadata: Nutrition research papers, popular recipes

[0854] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0855] Step 1:

[0856] The user uploads a document file containing specialized knowledge from the terminal to the server. The file selected by the user is then sent to the server. The server checks the format of the uploaded file and saves it as a temporary file. The input is the file containing specialized knowledge, and the output is the saved temporary file.

[0857] Step 2:

[0858] The server analyzes the uploaded data and generates metadata. The analysis process involves extracting text, identifying keywords, and extracting important items. From the analysis results, metadata such as title, author, and category is generated and stored in a database. The input is the saved temporary file, and the output is a database entry containing the metadata.

[0859] Step 3:

[0860] Users customize their chatbots through a web interface. They configure the database to use, the large-scale language model (LLM) to use, and the chatbot's interface. The server generates the chatbot based on this configuration. The input is the user's customization settings, and the output is the customized chatbot.

[0861] Step 4:

[0862] An end user inputs questions and comments to a chatbot, which may include comments that indicate emotions. The input is the user's questions and comments.

[0863] Step 5:

[0864] The server uses an emotion engine to analyze emotions from end-user questions and comments, and identifies emotional states using natural language processing techniques. The input is the end-user's questions and comments, and the output is the identified emotion data.

[0865] Step 6:

[0866] The server searches for related information based on the question and emotion data. It identifies related information in a database using keyword or ID search. The input is the identified emotion data and the question, and the output is a database entry containing the related information.

[0867] Step 7:

[0868] The server generates an appropriate response using a large-scale language model (LLM) based on the information retrieved from the database, and adjusts the tone of the response by referring to the emotional data. The input is relevant information and emotional data, and the output is a response tailored according to the emotion.

[0869] Step 8:

[0870] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. The input is the adjusted response, and the output is the response displayed on the end user's device.

[0871] Step 9:

[0872] If the end user provides feedback on a response, the server receives that feedback and uses it to improve the accuracy of future responses. The input is the end user's feedback, and the output is an improved response generation algorithm.

[0873] 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.

[0874] 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.

[0875] 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.

[0876] [Third embodiment]

[0877] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0878] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0879] 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).

[0880] 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.

[0881] 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.

[0882] 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).

[0883] 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.

[0884] 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.

[0885] 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.

[0886] 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.

[0887] 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.

[0888] 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."

[0889] The system of the present invention provides a chatbot that generates intelligent responses using data with specialized knowledge. Specific embodiments of the system are described below in natural language.

[0890] Upload your expertise

[0891] Users first upload data files containing specialized knowledge from their own devices to the server. Examples include medical papers, technical manuals, and legal documents. Using a web interface, users select the files they want to upload and press the send button to begin the process. At this point, the file format is verified and saved as a temporary file on the server.

[0892] Data analysis and metadata generation

[0893] The server receives the uploaded data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items. As a result of the analysis, metadata such as the file title, author, and category is generated. This metadata is organized and stored in a database for use by search engines.

[0894] Chatbot generation

[0895] Users can customize their chatbot through a web interface, selecting the database they want to use as a knowledge base, the large-scale language model (LLM) they want to use, and interface settings, and then a chatbot is built on the server that generates responses based on their expert knowledge.

[0896] Enter a question and search

[0897] When an end user types a question into the chatbot, the question is sent to the server, which extracts keywords from the question and searches for related information in the database. Entries related to the question are identified through keyword and ID searches.

[0898] Response Generation

[0899] The server generates a response based on the search results using a large-scale language model (LLM). The LLM interprets the information based on expert knowledge and composes it in natural language, presenting it in a format that is easy for the end user to understand. For example, in response to a question such as "What are the latest cancer treatments?", the server searches the database for the latest medical papers, and the LLM generates an appropriate response based on their content.

[0900] Providing a response

[0901] Finally, the server sends the generated response back to the end user's device. The end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has the ability to provide feedback, which is used to improve the accuracy of future responses.

[0902] Specific examples

[0903] Example 1: Medical knowledge-based chatbot

[0904] 1. The user uploads medical paper data from their device to the server.

[0905] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0906] 3. The user configures the "medical chatbot" through a web interface.

[0907] 4. When an end user asks, "Tell me about the latest cancer treatments," the server searches the database for relevant information and the LLM generates a response.

[0908] 5. The server generates a response and sends it back to the end user.

[0909] Example 2: Legal advice chatbot

[0910] 1. A user uploads a legal document to the server.

[0911] 2. The server analyzes the data, generates metadata, and stores it in a database.

[0912] 3. Users customize the "Legal Advice Chatbot."

[0913] 4. When an end user asks, "What legal details are required for a new business contract?", the server retrieves the information from the relevant documents and the LLM generates the appropriate response.

[0914] 5. The server generates a response and sends it back to the end user.

[0915] As a result, the system of the present invention efficiently utilizes specialized knowledge, enabling users to easily create advanced chatbots and appropriately respond to advanced questions related to specialized fields.

[0916] The processing flow will be explained below.

[0917] Step 1:

[0918] Users upload document files containing their specialized knowledge from their own devices to the server. Using a web interface, users select the files to upload and press the send button. At this time, the file format is confirmed and saved as a temporary file on the server.

[0919] Step 2:

[0920] The server receives the received data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items.

[0921] Step 3:

[0922] The server generates metadata such as file title, author, category, etc. from the analysis results, which are then organized and stored in a database for search engine compatibility.

[0923] Step 4:

[0924] Users configure the chatbot generation through a web interface, choosing the database they want to use as a knowledge base, the large-scale language model (LLM) they want to use, and interface settings to customize the chatbot.

[0925] Step 5:

[0926] The server receives the user's configuration and builds a chatbot that integrates the specified knowledge base and LLM model.

[0927] Step 6:

[0928] The user or end user types a question for the chatbot, which is then sent to the server.

[0929] Step 7:

[0930] The server analyzes the received question and extracts keywords. The extracted keywords are used to search for related information in the database. Entries related to the question are identified through keyword and ID searches.

[0931] Step 8:

[0932] The server generates a response based on the identified search results using a large-scale language model (LLM), which interprets the information based on expert knowledge and formats it in natural language in a format that is easy for the end user to understand.

[0933] Step 9:

[0934] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application.

[0935] Step 10:

[0936] If the response does not meet expectations, users can provide feedback, which is aggregated on the server and used to improve the accuracy of future responses.

[0937] Example 1

[0938] 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."

[0939] Current chatbot systems face the challenge of being unable to efficiently generate responses that incorporate specialized knowledge. Another problem is that it takes a lot of time and effort for users to upload and customize the knowledge base. In particular, when dealing with specialized knowledge, systems that can provide reliable responses instantly are required, but current technology is unable to adequately meet this demand.

[0940] 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.

[0941] In this invention, the server includes means for transmitting specialized knowledge from a user's communication device to the server, means for analyzing the transmitted data to generate metadata and storing it in a storage device, and means for the user to configure an information processing device through an interface, thereby enabling efficient construction of an advanced chatbot based on specialized knowledge and instantaneous generation of specialized responses.

[0942] "Specialized knowledge" is information that includes advanced knowledge and skills in a specific field.

[0943] "Users" are individuals or businesses that use the system to provide their expertise to the chatbot.

[0944] "Communication equipment" refers to a device for sending and receiving data via the Internet, including personal computers and smartphones.

[0945] A "server" is a computer system that receives, processes, and stores data submitted by users.

[0946] "Transmission means" means the method or process for transmitting data from the user's communication device to the server.

[0947] "Analysis means" refers to the method or technology by which the server analyzes the specialized knowledge and generates the metadata.

[0948] "Metadata" is information used to identify and organize the content of data, including title, author, category, etc.

[0949] "Storage device" refers to a device or system for storing analyzed data and metadata.

[0950] "Interface" means the user interface through which a user interacts with the system, including a web browser or dedicated application.

[0951] An "information processing device" is a program and hardware for processing data, such as a chatbot or search engine customized by the user.

[0952] An "enquiry" is a question or request that an end user types into a chatbot.

[0953] An "extraction means" is a method or technique for selecting relevant information from a storage device in response to a query.

[0954] A "generative AI model" is an artificial intelligence model that generates natural language responses based on input data.

[0955] A "response generation means" is a method or technique for generating a response based on the extracted information.

[0956] "Return mechanism" means the method or process for returning the generated response to the end user's communication device.

[0957] The system of the present invention provides a chatbot that generates sophisticated responses using data with specialized knowledge. Specific embodiments of the system are described in detail below.

[0958] Upload your expertise

[0959] Users use their communication devices to send data files containing specialized knowledge to the server. The data files can be, for example, medical papers, technical manuals, or legal documents. The user clicks the upload button through the interface, selects the file, and then presses the send button. The server temporarily stores the received file and checks whether the file format is appropriate.

[0960] Data analysis and metadata generation

[0961] The server analyzes the temporarily stored data files. Specifically, it uses Apache Tika to extract text data from the file contents. The extracted text data is then analyzed using a natural language processing tool (e.g., NLTK) to identify keywords and important items. Based on the results of this analysis, metadata such as title, author, and classification is generated and stored in a storage device. Generating and storing metadata allows for efficient subsequent data searches.

[0962] Chatbot generation

[0963] Users can customize their chatbot through the interface by selecting the data stored in the memory to use as a knowledge base, a large-scale language model (e.g., GPT-4), and interface settings. Based on these settings, the server builds a chatbot that generates responses based on expert knowledge. This chatbot utilizes generative AI models to generate responses to questions, allowing it to provide highly accurate responses even in situations requiring expert knowledge.

[0964] Enter a question and search

[0965] When an end user inputs a question into the chatbot, the question is sent to the server. The server analyzes the received question and extracts keywords. Natural language processing tools are used here as well. Based on the extracted keywords, related information is searched for in the storage device. A search engine such as Elasticsearch is used for this search.

[0966] Response Generation

[0967] The server generates a response based on the search results using a generative AI model (e.g., GPT-4), which leverages pre-trained models to translate the received data into natural language and construct a response in a way that is easy for the end user to understand.

[0968] Providing a response

[0969] The server generates a response and sends it back to the end user's device. The end user can review the response through an interface. If the response is unsatisfactory, the end user can provide feedback. This feedback is processed by the server and used to improve future responses.

[0970] Specific examples

[0971] Example 1: Medical knowledge-based chatbot

[0972] 1. The user sends medical paper data from a communication device to the server.

[0973] 2. The server analyzes the data, generates metadata, and stores it in a storage device.

[0974] 3. The user configures the "medical chatbot" in the interface.

[0975] 4. When an end user asks, "Tell me about the latest cancer treatments," the server searches for relevant information from its storage device, and the generative AI model generates a response.

[0976] 5. The server generates a response and sends it back to the end user.

[0977] Example 2: Legal advice chatbot

[0978] 1. A user submits a legal document to the server.

[0979] 2. The server analyzes the data, generates metadata, and stores it in a storage device.

[0980] 3. Users customize the "Legal Advice Chatbot."

[0981] 4. When an end user asks, "What legal details are required for a new business contract?", the server retrieves information from relevant documents and a generative AI model generates the appropriate response.

[0982] 5. The server generates a response and sends it back to the end user.

[0983] Prompt Sentence Examples

[0984] "Tell me about the latest cancer treatments"

[0985] "I want to know the latest information on infectious disease prevention measures"

[0986] "Learn more about diabetes treatment"

[0987] As described above, the system of the present invention efficiently utilizes specialized knowledge, allowing users to easily create advanced chatbots that can appropriately respond to advanced questions related to specialized fields.

[0988] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0989] Step 1: Upload your expertise

[0990] Users access the web interface from their communication devices and select the data file containing their specialized knowledge. When the user presses the upload button, the selected file is sent to the server. The server temporarily stores the received file and checks the file format.

[0991] Input: Data file (e.g. PDF, DOCX) containing the expertise submitted by the user.

[0992] Output: The temporarily saved data file is stored on the server.

[0993] Step 2: Data analysis and metadata generation

[0994] The server extracts the contents of the temporarily saved files into text data using Apache Tika. It then analyzes the text data using the Natural Language Toolkit (NLTK) to identify keywords and important items. Based on the analyzed data, it generates metadata such as title, author, and category and stores it in a storage device.

[0995] Input: Data file temporarily stored on the server.

[0996] Output: The parsed text data and generated metadata are stored in a storage device.

[0997] Step 3: Create a chatbot

[0998] Users configure their chatbot through a web interface. They select the knowledge base data to use, the generative AI model (e.g., GPT-4), and interface settings, and generate the chatbot based on those settings. The server then builds the chatbot using the selected settings and connects the generative AI model.

[0999] Input: User-configured knowledge base data, generative AI models, and interface settings.

[1000] Output: The customized chatbot is built on the server and is ready for use.

[1001] Step 4: Enter your question and search

[1002] When an end user types a question into a chatbot, it is sent to a server that analyzes the question, extracts keywords, and then uses a search engine such as Elasticsearch to search for relevant information in storage.

[1003] Input: Questions submitted by end users.

[1004] Output: Keywords are extracted and related information is searched.

[1005] Step 5: Response Generation

[1006] The server generates a response based on the search results using a generative AI model (e.g., GPT-4), which generates a natural language response based on the search results and formats it in a format that is easy for the end user to understand.

[1007] Input: Relevant information identified by the search engine.

[1008] Output: The natural language response generated by the generative AI model.

[1009] Step 6: Providing a response

[1010] The server generates a response and sends it back to the end user's device. The end user receives the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the user can provide feedback. This feedback is processed by the server and used to improve future responses.

[1011] Input: The generated natural language response.

[1012] Output: The response and possible feedback provided to the end user.

[1013] (Application example 1)

[1014] 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."

[1015] Current factory work requires a wide variety of specialized knowledge for technical troubleshooting and maintenance. However, it is not easy for workers to quickly retrieve the appropriate information, which can lead to low work efficiency and mistakes. In addition, current systems do not provide sufficient means to obtain the necessary information in real time. To solve this problem, there is a need for a system that allows workers to easily receive appropriate advice based on specialized knowledge using smart devices.

[1016] 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.

[1017] In this invention, the server includes means for uploading expert knowledge from a user's terminal, means for analyzing the uploaded data to generate metadata and storing it in a database, means for a user to customize the chatbot through a web interface, means for searching for related information from a database in response to an input question, means for generating an appropriate response using a large-scale language model based on the search result, means for returning the generated response to the end user's terminal, and means for an end user to input a question in real time using a smart device and receive a response, thereby enabling workers in a factory to receive appropriate advice based on expert knowledge in real time.

[1018] "Specialized knowledge" refers to advanced information and knowledge in a particular field.

[1019] "User" refers to an individual or organization that uses the system of the present invention.

[1020] A "server" refers to a computer system that transfers, stores, analyzes, and generates responses to data.

[1021] "Terminal" refers to the device that a user uses to access the server, such as a PC, smartphone, or tablet.

[1022] "Data analysis" refers to the process of examining uploaded expert knowledge data from multiple angles to understand its meaning and structure.

[1023] "Metadata" refers to information about uploaded data, including accompanying information such as title, author, and category.

[1024] "Database" refers to a repository where analyzed and generated metadata is stored and can be searched and referenced as needed.

[1025] "Web interface" refers to the screen that allows users to access the system through a web browser and perform various operations.

[1026] A "chatbot" refers to a software program that automatically generates responses to user questions.

[1027] "Entered Question" refers to the inquiry sent by the end user to the system.

[1028] "Large-scale language model" refers to a machine learning model for processing natural language based on a wide range of datasets.

[1029] "Appropriate response" means the most relevant and accurate answer to the entered question.

[1030] "End user" refers to the ultimate user of the chatbot.

[1031] "Smart devices" refer to devices that are connected to the Internet and have advanced functionality, including smart glasses and smartphones.

[1032] "Real time" refers to a time range in which processing occurs immediately with little or no delay.

[1033] The system of the present invention is designed to efficiently perform technical troubleshooting and maintenance in factory operations, allowing users to receive real-time advice based on expert knowledge using their smart devices.

[1034] 1. Upload your expertise

[1035] First, a user uploads a data file containing specialized knowledge (such as a technical manual) from their own device to the server. The upload is done through a web interface, the file format is confirmed, and the file is saved as a temporary file on the server.

[1036] 2. Data analysis and metadata generation

[1037] The server analyzes the uploaded data files and generates metadata such as title, author, category, etc. Natural language processing technology is used for the analysis, and the metadata is stored in a database.

[1038] 3. Chatbot generation

[1039] Next, users customize the chatbot through a web interface, including the database to use, the large-scale language model (LLM), and interface settings. This customization results in a chatbot based on their expertise being built on the server.

[1040] 4. Question and Response Processing

[1041] End users input questions in real time using smart devices (such as smart glasses). These questions are sent to a server, which analyzes the received questions, extracts keywords, and searches for relevant information in a database. Based on the search results, the server uses a large-scale language model (LLM) to generate an appropriate response. This response is then sent back to the end user's smart device in real time.

[1042] 5. Specific Examples

[1043] Take a concrete example of troubleshooting a robotic arm in a factory: A worker puts on smart glasses and asks, "What caused the robotic arm to stop working? What can I do to fix it?" This question is sent to a server, which searches for relevant information in a database and uses a generative AI model to create a prompt like this:

[1044] text

[1045] Document name: maintenance_manual.txt

[1046] Q: What is the reason for the robot arm stopping and how can I fix it?

[1047] answer:

[1048] The server inputs this prompt into a large-scale language model to generate an appropriate response and provide it to the end user.

[1049] The hardware used includes smart devices (e.g., smart glasses) and servers. The software uses Python, the OpenAI API, and natural language processing algorithms. Text analysis algorithms are used for data analysis, tagging techniques are used for metadata generation, and large-scale language models (e.g., text-davinci-003) are used for response generation.

[1050] In this way, the system of the present invention provides advanced support for quickly and efficiently resolving technical problems in factory operations.

[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 uploads a data file containing expert knowledge from their device to the server. The user uses a web interface to select the data file and presses the send button. At this time, the file format is confirmed and the server temporarily stores the received file. The input is the expert knowledge data file, and the output is the data stored as a temporary file.

[1054] Step 2:

[1055] The server analyzes the uploaded data files and generates metadata. The server performs data analysis such as text extraction, keyword identification, and extraction of important items, and generates metadata such as title, author, and category. The input is data saved as a temporary file, and the output is the generated metadata.

[1056] Step 3:

[1057] The server stores the generated metadata in a database. The analyzed data and metadata are appropriately organized and saved in the database. The input is the generated metadata, and the output is the metadata stored in the database.

[1058] Step 4:

[1059] Users customize chatbots through a web interface. They select the database to use as a knowledge base, the large-scale language model (LLM), interface settings, and so on. This allows a chatbot based on expert knowledge to be built on the server. The input is customization setting information, and the output is the customized chatbot.

[1060] Step 5:

[1061] End users input questions in real time using smart devices. The questions are sent to the server from devices such as smart glasses or smartphones. The input is the question from the end user, and the output is the question sent to the server.

[1062] Step 6:

[1063] The server analyzes the received question, extracts keywords, and searches the database for related information. The server analyzes the question text and identifies keywords. It then searches for related entries in the database. The input is the received question, and the output is the related entries.

[1064] Step 7:

[1065] The server generates a response based on the search results using a large-scale language model (LLM). A prompt sentence is generated and input to a generative AI model (e.g., text-davinci-003) to get an appropriate response. The input is the relevant entry and the prompt sentence, and the output is the generated response. An example prompt sentence is:

[1066] text

[1067] Document name: maintenance_manual.txt

[1068] Q: What is the reason for the robot arm stopping and how can I fix it?

[1069] answer:

[1070] Step 8:

[1071] The server sends the generated response back to the end user's smart device, which receives the response through smart glasses or a smartphone. The input is the generated response, and the output is the response sent back to the smart device.

[1072] 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.

[1073] The system of the present invention combines an advanced chatbot that utilizes specialized knowledge with an emotion engine that recognizes the user's emotions. Specific embodiments of the system will be described below in natural language.

[1074] Upload your expertise

[1075] Users upload specialized document files from their own devices to the server. Examples include medical papers, technical manuals, and legal documents. Using a web interface, users select the files they want to upload and press the send button to begin the process. At this point, the file format is verified and saved as a temporary file on the server.

[1076] Data analysis and metadata generation

[1077] The server receives the uploaded data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items. As a result of the analysis, metadata such as the file title, author, and category is generated. The metadata is organized and stored in a database for search engine compatibility.

[1078] Chatbot generation

[1079] Users can customize their chatbot through a web interface, selecting the database they want to use as their knowledge base, the large-scale language model (LLM) they want to use, and interface settings, which then builds a chatbot on the server that generates responses based on their expert knowledge.

[1080] Emotion Engine Operation

[1081] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. This emotion engine uses natural language processing technology to analyze the emotional state of the user's input.

[1082] Enter a question and search

[1083] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the question is sent to the server.

[1084] The server extracts keywords from the received question and searches for related information in the database. Keyword searches and ID searches identify entries related to the question.

[1085] Response Generation

[1086] The server generates a response based on the identified search results using a large-scale language model (LLM). The LLM generates a human-like response based on expert knowledge and interpretation. It also references the user's emotional information, as recognized by the emotion engine, and adjusts the response accordingly. For example, if the user is expressing anger, the tone of the response can be toned down.

[1087] Providing a response

[1088] The server then sends the generated response back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a function to provide feedback, which is used to improve the accuracy of future responses.

[1089] Specific examples

[1090] Example 1: A chatbot using medical knowledge and emotion recognition

[1091] 1. The user uploads medical paper data from their device to the server.

[1092] 2. The server analyzes the data, generates metadata, and stores it in a database.

[1093] 3. The user configures the "medical chatbot" through a web interface.

[1094] 4. When an end user asks, "Tell me about the latest cancer treatments," the comment also includes an emotional comment such as, "I'm worried."

[1095] 5. The server's emotion engine recognizes the user's "concern."

[1096] 6. The server searches the database for relevant information, and the LLM generates a response, softening the tone of the response based on the emotional information.

[1097] 7. The server generates a response and sends it back to the end user.

[1098] Example 2: Legal advice and emotion recognition chatbot

[1099] 1. A user uploads a legal document to the server.

[1100] 2. The server analyzes the data, generates metadata, and stores it in a database.

[1101] 3. Users customize the "Legal Advice Chatbot."

[1102] 4. When an end user asks, "What legal details are required for a new business contract?", the comment also includes an emotional response: "I'm worried."

[1103] 5. The server's emotion engine recognizes the user's "anxiety."

[1104] 6. The server retrieves the information from relevant documents, and the LLM generates an appropriate response, adjusting the tone of the response to a reassuring one based on the emotional information.

[1105] 7. The server generates a response and sends it back to the end user.

[1106] As described above, the system of the present invention not only efficiently utilizes specialized knowledge, but also recognizes the user's emotions and adjusts the response content, thereby realizing more personalized dialogue.

[1107] The processing flow will be explained below.

[1108] Step 1:

[1109] Users upload document files containing their specialized knowledge from their own devices to the server. Using a web interface, users select the files to upload and press the send button. At this time, the file format is checked and saved as a temporary file on the server.

[1110] Step 2:

[1111] The server receives the received data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting key points. Specifically, the text is analyzed using natural language processing tools to extract key keywords and phrases.

[1112] Step 3:

[1113] The server generates metadata such as file title, author, and category from the analysis results, which are then organized and stored in a database for search engine compatibility.

[1114] Step 4:

[1115] Users customize their chatbot through a web interface, selecting the database they want to use as their knowledge base, the large-scale language model (LLM) they want to use, and interface settings—for example, choosing a medical database and the latest LLM.

[1116] Step 5:

[1117] The server integrates the configured knowledge base with the LLM model and builds a chatbot based on the specified settings, resulting in a chatbot that can respond with expert knowledge.

[1118] Step 6:

[1119] When a user or end-user types a question into the chatbot, they can include a comment related to their feelings along with the question, for example, "Tell me about this new treatment, but I'm a little nervous."

[1120] Step 7:

[1121] The server's emotion engine analyzes the user's input and identifies the emotion, for example, "anxiety," using natural language processing tools. This emotion information is then used in the response generation process.

[1122] Step 8:

[1123] The server extracts keywords from the received question and searches for related information in its database, performing keyword and ID searches to identify entries related to the question.

[1124] Step 9:

[1125] The server generates a response based on the identified search results using a large-scale language model (LLM). It also references the emotional information recognized by the emotion engine and adjusts the response accordingly. For example, it takes into account the emotional information of "anxiety" and makes the tone of the response gentler.

[1126] Step 10:

[1127] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application.

[1128] Step 11:

[1129] The user or end user can provide feedback on the returned responses, which the server collects and uses to improve future response accuracy and emotion recognition.

[1130] Specific examples

[1131] The user uploads medical paper data from their device to the server and sets up the chatbot. When the end user then asks, "Please tell me about the latest cancer treatments. I'm anxious," the server recognizes "anxiety" using its emotion engine, searches for information from the database, and uses LLM to generate a response in a gentle tone and send it back to the end user.

[1132] In this way, the system of the present invention leverages emotion recognition capabilities to deliver sophisticated, personalized responses based on expert knowledge.

[1133] Example 2

[1134] 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."

[1135] Conventional chatbot systems have difficulty efficiently utilizing expert knowledge and are unable to generate personalized responses that take into account the user's emotional state. Furthermore, they lack the ability to incorporate end-user feedback into the generated responses. Therefore, there is a need for a system that utilizes expert knowledge while providing appropriate responses that reflect the user's emotions.

[1136] 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.

[1137] In this invention, the server includes means for uploading expert knowledge from a user's terminal to the server, means for analyzing the uploaded data to generate metadata and storing it in a database, means for a user to customize the chatbot through a web interface, means for searching the database for related information in response to an input question, means for generating an appropriate response using a large-scale language model based on the search result, means for analyzing the user's emotions and adjusting the response content based on the emotional information, and means for returning the generated response to the end user's terminal. This makes it possible to generate personalized responses according to the user's emotional state and reflect the end user's feedback while efficiently utilizing expert knowledge.

[1138] "Expertise" refers to a high level of information or understanding gained in a particular field.

[1139] "User" refers to the entity that uses the system to upload specialized knowledge and customize the chatbot.

[1140] "Terminal" is a general term for computing devices and communication equipment used by users, including personal computers, smartphones, tablets, etc.

[1141] A "server" is a computing device that provides functions such as uploading specialized knowledge, analyzing data, generating metadata, building chatbots, and generating and returning responses.

[1142] The "emotion engine" is a module that uses natural language processing technology to analyze the emotional state of text input from users.

[1143] "Metadata" is data that describes information about the contents of a file, including the title, author, category, etc.

[1144] A "database" is a collection of data in which specialized knowledge and metadata are systematically stored and can be searched and used.

[1145] "Web interface" refers to the screens and input forms that allow users to interact with a system via a browser.

[1146] A "chatbot" is a program that uses large-scale language models to enable natural dialogue in response to user questions.

[1147] A "large-scale language model (LLM)" is a language model trained on large amounts of text data, and is capable of generating natural-sounding sentences like humans.

[1148] A "response" is the information or advice returned by the chatbot to the end user.

[1149] "Feedback" means comments or ratings provided by end users that are used to improve system performance and responsiveness.

[1150] "End user" refers to the final user who actually receives the response from the chatbot.

[1151] The system of the present invention provides an advanced chatbot that utilizes expert knowledge, recognizes user emotions, and adjusts its responses accordingly. The system includes a function for uploading expert knowledge from a user's device to a server, analyzing the uploaded data to generate metadata, and storing it in a database. The system then allows the user to customize the chatbot through a web interface, searches the database for relevant information in response to an input question, and generates an appropriate response using a large-scale language model (LLM) based on the search results. The system also includes a function for analyzing user emotions, adjusting the response based on the emotional information, and returning the generated response to the end user's device.

[1152] Specifically, the system works as follows:

[1153] Users upload document files containing specialized knowledge from their own devices to the server. For example, medical papers, technical manuals, legal documents, etc. The user uses a web interface to select the file to upload and press the send button. At this time, the file format is confirmed and saved as a temporary file on the server.

[1154] The server receives the uploaded data files and analyzes their contents. During this analysis process, text is extracted using Apache Tika, and natural language processing libraries such as NLTK and SpaCy are used to identify keywords and extract important information. The analysis results in metadata such as the file title, author, and category. The metadata is organized so that it can be used by search engines, and stored in a database (e.g., PostgreSQL).

[1155] Users customize their chatbot through a web interface, selecting the database they want to use as a knowledge base, a large-scale language model (e.g., OpenAI's GPT-4), and interface settings. This builds a chatbot on the server that generates responses based on their expert knowledge.

[1156] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. The emotion engine uses natural language processing techniques such as TextBlob and VADER to analyze the emotional state of the user's input.

[1157] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the question is sent to the server.

[1158] The server extracts keywords from the received question and searches for related information in the database. Keyword searches and ID searches identify entries related to the question.

[1159] The server generates a response based on the identified search results using a large-scale language model (LLM). The LLM generates a human-like response based on expert knowledge and interpretation. It also references the user's emotional information, as recognized by the emotion engine, and adjusts the response accordingly. For example, if the user is expressing anger, the tone of the response can be toned down.

[1160] The server then sends the generated response back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a function to provide feedback, which is used to improve the accuracy of future responses.

[1161] As a concrete example, consider the following scenario:

[1162] Example 1: A chatbot using medical knowledge and emotion recognition

[1163] 1. The user uploads medical paper data from their device to the server.

[1164] 2. The server analyzes the data, generates metadata, and stores it in a database.

[1165] 3. The user configures the "medical chatbot" through a web interface.

[1166] 4. When an end user asks, "Tell me about the latest cancer treatments," the comment also includes an emotional comment such as, "I'm worried."

[1167] 5. The server's emotion engine recognizes the user's "concern."

[1168] 6. The server searches the database for relevant information, and the LLM generates a response, softening the tone of the response based on the emotional information.

[1169] 7. The server generates a response and sends it back to the end user.

[1170] Example 2: Legal advice and emotion recognition chatbot

[1171] 1. A user uploads a legal document to the server.

[1172] 2. The server analyzes the data, generates metadata, and stores it in a database.

[1173] 3. Users customize the "Legal Advice Chatbot."

[1174] 4. When an end user asks, "What legal details are required for a new business contract?", the comment also includes an emotional response: "I'm worried."

[1175] 5. The server's emotion engine recognizes the user's "anxiety."

[1176] 6. The server retrieves the information from relevant documents, and the LLM generates an appropriate response, adjusting the tone of the response to a reassuring one based on the emotional information.

[1177] 7. The server generates a response and sends it back to the end user.

[1178] In this way, it is possible to efficiently utilize expert knowledge, recognize user emotions, and tailor responses to achieve more personalized interactions.

[1179] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1180] Step 1:

[1181] The user selects a document file containing the specialized knowledge and uploads it to the server from the terminal.

[1182] What happens: The user opens a file selection dialog in the web interface, selects a file to upload, and clicks the "Submit" button.

[1183] Input: A file containing expert knowledge stored on the user's device.

[1184] Output: The document file sent to the server.

[1185] Step 2:

[1186] The server checks the format of the received file and saves it as a temporary file.

[1187] Specific operation: The server verifies the format of the received file and, if it is a compatible format (e.g., PDF, DOCX), saves the file in a temporary folder.

[1188] Input: The document file submitted by the user.

[1189] Output: A temporary file saved on the server.

[1190] Step 3:

[1191] The server analyzes the saved file with a text extraction tool to extract text data.

[1192] What it does: The server uses Apache Tika to extract text from the file.

[1193] Input: A temporary file stored on the server.

[1194] Output: The extracted text data.

[1195] Step 4:

[1196] The server extracts keywords and important items from the analyzed text data.

[1197] How it works: The server uses natural language processing libraries such as NLTK and SpaCy to extract keywords and important items from text data.

[1198] Input: Extracted text data.

[1199] Output: Extracted keywords and important items.

[1200] Step 5:

[1201] The server generates metadata such as file title, author, category, etc. and stores it in a database.

[1202] Specific operation: The server constructs metadata based on the extracted information and stores it in a database (e.g., PostgreSQL).

[1203] Input: Extracted keywords and important items.

[1204] Output: Metadata stored in a database.

[1205] Step 6:

[1206] Users customize their chatbot through a web interface.

[1207] Specific behavior: The user selects the database to use as the knowledge base, the large language model (LLM), and the interface settings in the web interface.

[1208] Input: User customization settings.

[1209] Output: Customized chatbot configuration information.

[1210] Step 7:

[1211] The server builds a chatbot based on the selected configuration information.

[1212] Specific operation: The server uses the selected LLM (e.g., OpenAI's GPT-4) to build the chatbot's interface and response generation logic.

[1213] Input: Customized chatbot configuration information.

[1214] Output: The built chatbot.

[1215] Step 8:

[1216] The server uses an emotion engine to analyze emotions from the end user's input and records the results.

[1217] Specific operation: The server uses natural language processing libraries such as TextBlob and VADER to evaluate the emotional state of the input text and stores the results in internal data.

[1218] Input: End user input text.

[1219] Output: Parsed emotion data.

[1220] Step 9:

[1221] The server receives the end user's question and searches the question content in a database based on the analyzed emotional state.

[1222] How it works: The server uses NLTK or SpaCy to extract keywords from the question and search for relevant information in a database.

[1223] Input: End-user question and parsed sentiment data.

[1224] Output: The relevant information found.

[1225] Step 10:

[1226] The server generates responses based on the search results using a large-scale language model (LLM) and adjusts the tone based on sentiment data.

[1227] What it does: The server uses LLM to generate natural-sounding responses based on search results and sentiment analysis data, adjusting tone as needed.

[1228] Input: Retrieved relevant information and sentiment data.

[1229] Output: The generated response.

[1230] Step 11:

[1231] The server then sends the generated response back to the end user's terminal.

[1232] Specific behavior: The server sends the generated response as an HTTP response.

[1233] Input: The generated response.

[1234] Output: The response sent back to the end user.

[1235] Step 12:

[1236] End users receive responses and provide feedback through a web interface or dedicated application.

[1237] What happens: The end user reviews the response they receive and submits their rating and comments through the feedback form.

[1238] Input: Generated responses and end-user feedback.

[1239] Output: Feedback data stored on the server.

[1240] This enables the system to efficiently utilize specialized knowledge while creating an advanced chatbot that responds to the user's emotions.

[1241] (Application example 2)

[1242] 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."

[1243] Conventional chatbot systems simply provide factual information without considering the user's emotions. As a result, they are unable to respond appropriately even when the user is feeling anxious or fatigued, resulting in a poor user experience. The present invention aims to achieve more personalized dialogue by analyzing the user's emotions and providing responses that correspond to those emotions.

[1244] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for uploading specialized knowledge from a user's terminal to the server, means for analyzing the uploaded data to generate metadata and storing it in a database, means for the user to customize the chatbot through a web interface, means for searching for related information from a database for an input question, means for generating an appropriate response using a large-scale language model based on the search results, means for analyzing the user's emotions using an emotion engine, means for adjusting the tone of the response based on the emotion information, and means for returning the generated response to the end user's terminal. This makes it possible to provide an appropriate response according to the user's emotions.

[1245] "Expertise" is knowledge based on advanced information and experience in a particular field.

[1246] A "server" is a computer system that provides data and services to multiple client computers over a network.

[1247] "Upload" is the operation of transferring a data file from a client's computer to a server.

[1248] "Data analysis" is the process of organizing collected data and converting it into meaningful information.

[1249] "Metadata" is supplementary information that describes the content, structure, characteristics, etc. of data.

[1250] A "database" is an organized collection of data, a system that allows for efficient searching and manipulation.

[1251] "Web interface" is a collective term for web-based screens and functions that allow users to interact with a system over the Internet.

[1252] A "chatbot" is a program that uses natural language processing to automate interactions with users.

[1253] A "large-scale language model (LLM)" is a deep learning model trained on large amounts of text data that has the ability to understand and generate natural language.

[1254] An "emotion engine" is a technology that analyzes emotions from a user's statements and input and identifies their state.

[1255] "Tone of response" refers to the impression and nuance of emotion and expression contained in the generated reply.

[1256] The system embodying this invention uses user expertise to generate a chatbot, and also recognizes the user's emotions and adjusts responses. The system is composed of the following components:

[1257] Upload your expertise

[1258] The user uploads a document file containing specialized knowledge from their device to the server. The uploaded file is first checked for its format on the server, and if it is in the correct format, it is saved as a temporary file.

[1259] Data analysis and metadata generation

[1260] The server analyzes the uploaded data and generates metadata, including file title, author, category, etc. The generated metadata is stored in a database. Natural language processing (NLP) technology is used for data analysis.

[1261] Chatbot generation

[1262] Users can customize their chatbot through a web interface, selecting the database they want to use as a knowledge base and the large-scale language model (LLM) they want to use, such as GPT-3. Based on these settings, the server generates the chatbot.

[1263] Emotion Engine Operation

[1264] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. The emotion engine uses natural language processing technology to analyze the emotional state of the user's input. For example, it identifies emotions such as "worry."

[1265] Enter a question and search

[1266] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the server extracts keywords from the question and searches for relevant information in the database, using keyword or ID search.

[1267] Response Generation

[1268] The server uses a large-scale language model (LLM) based on the search results to generate an appropriate response. It also refers to the user's emotional information recognized by the emotion engine and adjusts the tone of the response accordingly. For example, if the user indicates "fatigue," the server adjusts the response to be more refreshing.

[1269] Providing a response

[1270] The server then sends the generated response back to the end user's device, who receives it through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a feature that allows the user to provide feedback, which is used to improve the accuracy of future responses.

[1271] Specific examples

[1272] As an example, consider the case where a user uploads a document about food to the server and configures a "food delivery helper" chatbot based on that information. The end user asks, "What's your lunch recommendation today?" and adds a comment containing the emotion, "I'm a little tired." In this case, the server's emotion engine recognizes the emotion and suggests a refreshing lunch menu.

[1273] Prompt Sentence Examples

[1274] "Question: What's your recommended lunch today?

[1275] Emotions: Tired

[1276] Metadata: Nutrition research papers, popular recipes

[1277] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1278] Step 1:

[1279] The user uploads a document file containing specialized knowledge from the terminal to the server. The file selected by the user is then sent to the server. The server checks the format of the uploaded file and saves it as a temporary file. The input is the file containing specialized knowledge, and the output is the saved temporary file.

[1280] Step 2:

[1281] The server analyzes the uploaded data and generates metadata. The analysis process involves extracting text, identifying keywords, and extracting important items. From the analysis results, metadata such as title, author, and category is generated and stored in a database. The input is the saved temporary file, and the output is a database entry containing the metadata.

[1282] Step 3:

[1283] Users customize their chatbots through a web interface. They configure the database to use, the large-scale language model (LLM) to use, and the chatbot's interface. The server generates the chatbot based on this configuration. The input is the user's customization settings, and the output is the customized chatbot.

[1284] Step 4:

[1285] An end user inputs questions and comments to a chatbot, which may include comments that indicate emotions. The input is the user's questions and comments.

[1286] Step 5:

[1287] The server uses an emotion engine to analyze emotions from end-user questions and comments, and identifies emotional states using natural language processing techniques. The input is the end-user's questions and comments, and the output is the identified emotion data.

[1288] Step 6:

[1289] The server searches for related information based on the question and emotion data. It identifies related information in a database using keyword or ID search. The input is the identified emotion data and the question, and the output is a database entry containing the related information.

[1290] Step 7:

[1291] The server generates an appropriate response using a large-scale language model (LLM) based on the information retrieved from the database, and adjusts the tone of the response by referring to the emotional data. The input is relevant information and emotional data, and the output is a response tailored according to the emotion.

[1292] Step 8:

[1293] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. The input is the adjusted response, and the output is the response displayed on the end user's device.

[1294] Step 9:

[1295] If the end user provides feedback on a response, the server receives that feedback and uses it to improve the accuracy of future responses. The input is the end user's feedback, and the output is an improved response generation algorithm.

[1296] 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.

[1297] 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.

[1298] 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.

[1299] [Fourth embodiment]

[1300] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1301] 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.

[1302] 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).

[1303] 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.

[1304] 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.

[1305] 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).

[1306] 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.

[1307] 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.

[1308] 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.

[1309] 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.

[1310] 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.

[1311] 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.

[1312] 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."

[1313] The system of the present invention provides a chatbot that generates intelligent responses using data with specialized knowledge. Specific embodiments of the system are described below in natural language.

[1314] Upload your expertise

[1315] Users first upload data files containing specialized knowledge from their own devices to the server. Examples include medical papers, technical manuals, and legal documents. Using a web interface, users select the files they want to upload and press the send button to begin the process. At this point, the file format is verified and saved as a temporary file on the server.

[1316] Data analysis and metadata generation

[1317] The server receives the uploaded data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items. As a result of the analysis, metadata such as the file title, author, and category is generated. This metadata is organized and stored in a database for use by search engines.

[1318] Chatbot generation

[1319] Users can customize their chatbot through a web interface, selecting the database they want to use as a knowledge base, the large-scale language model (LLM) they want to use, and interface settings, and then a chatbot is built on the server that generates responses based on their expert knowledge.

[1320] Enter a question and search

[1321] When an end user types a question into the chatbot, the question is sent to the server, which extracts keywords from the question and searches for related information in the database. Entries related to the question are identified through keyword and ID searches.

[1322] Response Generation

[1323] The server generates a response based on the search results using a large-scale language model (LLM). The LLM interprets the information based on expert knowledge and composes it in natural language, presenting it in a format that is easy for the end user to understand. For example, in response to a question such as "What are the latest cancer treatments?", the server searches the database for the latest medical papers, and the LLM generates an appropriate response based on their content.

[1324] Providing a response

[1325] Finally, the server sends the generated response back to the end user's device. The end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has the ability to provide feedback, which is used to improve the accuracy of future responses.

[1326] Specific examples

[1327] Example 1: Medical knowledge-based chatbot

[1328] 1. The user uploads medical paper data from their device to the server.

[1329] 2. The server analyzes the data, generates metadata, and stores it in a database.

[1330] 3. The user configures the "medical chatbot" through a web interface.

[1331] 4. When an end user asks, "Tell me about the latest cancer treatments," the server searches the database for relevant information and the LLM generates a response.

[1332] 5. The server generates a response and sends it back to the end user.

[1333] Example 2: Legal advice chatbot

[1334] 1. A user uploads a legal document to the server.

[1335] 2. The server analyzes the data, generates metadata, and stores it in a database.

[1336] 3. Users customize the "Legal Advice Chatbot."

[1337] 4. When an end user asks, "What legal details are required for a new business contract?", the server retrieves the information from the relevant documents and the LLM generates the appropriate response.

[1338] 5. The server generates a response and sends it back to the end user.

[1339] As a result, the system of the present invention efficiently utilizes specialized knowledge, enabling users to easily create advanced chatbots and appropriately respond to advanced questions related to specialized fields.

[1340] The processing flow will be explained below.

[1341] Step 1:

[1342] Users upload document files containing their specialized knowledge from their own devices to the server. Using a web interface, users select the files to upload and press the send button. At this time, the file format is confirmed and saved as a temporary file on the server.

[1343] Step 2:

[1344] The server receives the received data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items.

[1345] Step 3:

[1346] The server generates metadata such as file title, author, category, etc. from the analysis results, which are then organized and stored in a database for search engine compatibility.

[1347] Step 4:

[1348] Users configure the chatbot generation through a web interface, choosing the database they want to use as a knowledge base, the large-scale language model (LLM) they want to use, and interface settings to customize the chatbot.

[1349] Step 5:

[1350] The server receives the user's configuration and builds a chatbot that integrates the specified knowledge base and LLM model.

[1351] Step 6:

[1352] The user or end user types a question for the chatbot, which is then sent to the server.

[1353] Step 7:

[1354] The server analyzes the received question and extracts keywords. The extracted keywords are used to search for related information in the database. Entries related to the question are identified through keyword and ID searches.

[1355] Step 8:

[1356] The server generates a response based on the identified search results using a large-scale language model (LLM), which interprets the information based on expert knowledge and formats it in natural language in a format that is easy for the end user to understand.

[1357] Step 9:

[1358] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application.

[1359] Step 10:

[1360] If the response does not meet expectations, users can provide feedback, which is aggregated on the server and used to improve the accuracy of future responses.

[1361] Example 1

[1362] 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."

[1363] Current chatbot systems face the challenge of being unable to efficiently generate responses that incorporate specialized knowledge. Another problem is that it takes a lot of time and effort for users to upload and customize the knowledge base. In particular, when dealing with specialized knowledge, systems that can provide reliable responses instantly are required, but current technology is unable to adequately meet this demand.

[1364] 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.

[1365] In this invention, the server includes means for transmitting specialized knowledge from a user's communication device to the server, means for analyzing the transmitted data to generate metadata and storing it in a storage device, and means for the user to configure an information processing device through an interface, thereby enabling efficient construction of an advanced chatbot based on specialized knowledge and instantaneous generation of specialized responses.

[1366] "Specialized knowledge" is information that includes advanced knowledge and skills in a specific field.

[1367] "Users" are individuals or businesses that use the system to provide their expertise to the chatbot.

[1368] "Communication equipment" refers to a device for sending and receiving data via the Internet, including personal computers and smartphones.

[1369] A "server" is a computer system that receives, processes, and stores data submitted by users.

[1370] "Transmission means" means the method or process for transmitting data from the user's communication device to the server.

[1371] "Analysis means" refers to the method or technology by which the server analyzes the specialized knowledge and generates the metadata.

[1372] "Metadata" is information used to identify and organize the content of data, including title, author, category, etc.

[1373] "Storage device" refers to a device or system for storing analyzed data and metadata.

[1374] "Interface" means the user interface through which a user interacts with the system, including a web browser or dedicated application.

[1375] An "information processing device" is a program and hardware for processing data, such as a chatbot or search engine customized by the user.

[1376] An "enquiry" is a question or request that an end user types into a chatbot.

[1377] An "extraction means" is a method or technique for selecting relevant information from a storage device in response to a query.

[1378] A "generative AI model" is an artificial intelligence model that generates natural language responses based on input data.

[1379] A "response generation means" is a method or technique for generating a response based on the extracted information.

[1380] "Return mechanism" means the method or process for returning the generated response to the end user's communication device.

[1381] The system of the present invention provides a chatbot that generates sophisticated responses using data with specialized knowledge. Specific embodiments of the system are described in detail below.

[1382] Upload your expertise

[1383] Users use their communication devices to send data files containing specialized knowledge to the server. The data files can be, for example, medical papers, technical manuals, or legal documents. The user clicks the upload button through the interface, selects the file, and then presses the send button. The server temporarily stores the received file and checks whether the file format is appropriate.

[1384] Data analysis and metadata generation

[1385] The server analyzes the temporarily stored data files. Specifically, it uses Apache Tika to extract text data from the file contents. The extracted text data is then analyzed using a natural language processing tool (e.g., NLTK) to identify keywords and important items. Based on the results of this analysis, metadata such as title, author, and classification is generated and stored in a storage device. Generating and storing metadata allows for efficient subsequent data searches.

[1386] Chatbot generation

[1387] Users can customize their chatbot through the interface by selecting the data stored in the memory to use as a knowledge base, a large-scale language model (e.g., GPT-4), and interface settings. Based on these settings, the server builds a chatbot that generates responses based on expert knowledge. This chatbot utilizes generative AI models to generate responses to questions, allowing it to provide highly accurate responses even in situations requiring expert knowledge.

[1388] Enter a question and search

[1389] When an end user inputs a question into the chatbot, the question is sent to the server. The server analyzes the received question and extracts keywords. Natural language processing tools are used here as well. Based on the extracted keywords, related information is searched for in the storage device. A search engine such as Elasticsearch is used for this search.

[1390] Response Generation

[1391] The server generates a response based on the search results using a generative AI model (e.g., GPT-4), which leverages pre-trained models to translate the received data into natural language and construct a response in a way that is easy for the end user to understand.

[1392] Providing a response

[1393] The server generates a response and sends it back to the end user's device. The end user can review the response through an interface. If the response is unsatisfactory, the end user can provide feedback. This feedback is processed by the server and used to improve future responses.

[1394] Specific examples

[1395] Example 1: Medical knowledge-based chatbot

[1396] 1. The user sends medical paper data from a communication device to the server.

[1397] 2. The server analyzes the data, generates metadata, and stores it in a storage device.

[1398] 3. The user configures the "medical chatbot" in the interface.

[1399] 4. When an end user asks, "Tell me about the latest cancer treatments," the server searches for relevant information from its storage device, and the generative AI model generates a response.

[1400] 5. The server generates a response and sends it back to the end user.

[1401] Example 2: Legal advice chatbot

[1402] 1. A user submits a legal document to the server.

[1403] 2. The server analyzes the data, generates metadata, and stores it in a storage device.

[1404] 3. Users customize the "Legal Advice Chatbot."

[1405] 4. When an end user asks, "What legal details are required for a new business contract?", the server retrieves information from relevant documents and a generative AI model generates the appropriate response.

[1406] 5. The server generates a response and sends it back to the end user.

[1407] Prompt Sentence Examples

[1408] "Tell me about the latest cancer treatments"

[1409] "I want to know the latest information on infectious disease prevention measures"

[1410] "Learn more about diabetes treatment"

[1411] As described above, the system of the present invention efficiently utilizes specialized knowledge, allowing users to easily create advanced chatbots that can appropriately respond to advanced questions related to specialized fields.

[1412] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1413] Step 1: Upload your expertise

[1414] Users access the web interface from their communication devices and select the data file containing their specialized knowledge. When the user presses the upload button, the selected file is sent to the server. The server temporarily stores the received file and checks the file format.

[1415] Input: Data file (e.g. PDF, DOCX) containing the expertise submitted by the user.

[1416] Output: The temporarily saved data file is stored on the server.

[1417] Step 2: Data analysis and metadata generation

[1418] The server extracts the contents of the temporarily saved files into text data using Apache Tika. It then analyzes the text data using the Natural Language Toolkit (NLTK) to identify keywords and important items. Based on the analyzed data, it generates metadata such as title, author, and category and stores it in a storage device.

[1419] Input: Data file temporarily stored on the server.

[1420] Output: The parsed text data and generated metadata are stored in a storage device.

[1421] Step 3: Create a chatbot

[1422] Users configure their chatbot through a web interface. They select the knowledge base data to use, the generative AI model (e.g., GPT-4), and interface settings, and generate the chatbot based on those settings. The server then builds the chatbot using the selected settings and connects the generative AI model.

[1423] Input: User-configured knowledge base data, generative AI models, and interface settings.

[1424] Output: The customized chatbot is built on the server and is ready for use.

[1425] Step 4: Enter your question and search

[1426] When an end user types a question into a chatbot, it is sent to a server that analyzes the question, extracts keywords, and then uses a search engine such as Elasticsearch to search for relevant information in storage.

[1427] Input: Questions submitted by end users.

[1428] Output: Keywords are extracted and related information is searched.

[1429] Step 5: Response Generation

[1430] The server generates a response based on the search results using a generative AI model (e.g., GPT-4), which generates a natural language response based on the search results and formats it in a format that is easy for the end user to understand.

[1431] Input: Relevant information identified by the search engine.

[1432] Output: The natural language response generated by the generative AI model.

[1433] Step 6: Providing a response

[1434] The server generates a response and sends it back to the end user's device. The end user receives the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the user can provide feedback. This feedback is processed by the server and used to improve future responses.

[1435] Input: The generated natural language response.

[1436] Output: The response and possible feedback provided to the end user.

[1437] (Application example 1)

[1438] 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."

[1439] Current factory work requires a wide variety of specialized knowledge for technical troubleshooting and maintenance. However, it is not easy for workers to quickly retrieve the appropriate information, which can lead to low work efficiency and mistakes. In addition, current systems do not provide sufficient means to obtain the necessary information in real time. To solve this problem, there is a need for a system that allows workers to easily receive appropriate advice based on specialized knowledge using smart devices.

[1440] 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.

[1441] In this invention, the server includes means for uploading expert knowledge from a user's terminal, means for analyzing the uploaded data to generate metadata and storing it in a database, means for a user to customize the chatbot through a web interface, means for searching for related information from a database in response to an input question, means for generating an appropriate response using a large-scale language model based on the search result, means for returning the generated response to the end user's terminal, and means for an end user to input a question in real time using a smart device and receive a response, thereby enabling workers in a factory to receive appropriate advice based on expert knowledge in real time.

[1442] "Specialized knowledge" refers to advanced information and knowledge in a particular field.

[1443] "User" refers to an individual or organization that uses the system of the present invention.

[1444] A "server" refers to a computer system that transfers, stores, analyzes, and generates responses to data.

[1445] "Terminal" refers to the device that a user uses to access the server, such as a PC, smartphone, or tablet.

[1446] "Data analysis" refers to the process of examining uploaded expert knowledge data from multiple angles to understand its meaning and structure.

[1447] "Metadata" refers to information about uploaded data, including accompanying information such as title, author, and category.

[1448] "Database" refers to a repository where analyzed and generated metadata is stored and can be searched and referenced as needed.

[1449] "Web interface" refers to the screen that allows users to access the system through a web browser and perform various operations.

[1450] A "chatbot" refers to a software program that automatically generates responses to user questions.

[1451] "Entered Question" refers to the inquiry sent by the end user to the system.

[1452] "Large-scale language model" refers to a machine learning model for processing natural language based on a wide range of datasets.

[1453] "Appropriate response" means the most relevant and accurate answer to the entered question.

[1454] "End user" refers to the ultimate user of the chatbot.

[1455] "Smart devices" refer to devices that are connected to the Internet and have advanced functionality, including smart glasses and smartphones.

[1456] "Real time" refers to a time range in which processing occurs immediately with little or no delay.

[1457] The system of the present invention is designed to efficiently perform technical troubleshooting and maintenance in factory operations, allowing users to receive real-time advice based on expert knowledge using their smart devices.

[1458] 1. Upload your expertise

[1459] First, a user uploads a data file containing specialized knowledge (such as a technical manual) from their own device to the server. The upload is done through a web interface, the file format is confirmed, and the file is saved as a temporary file on the server.

[1460] 2. Data analysis and metadata generation

[1461] The server analyzes the uploaded data files and generates metadata such as title, author, category, etc. Natural language processing technology is used for the analysis, and the metadata is stored in a database.

[1462] 3. Chatbot generation

[1463] Next, users customize the chatbot through a web interface, including the database to use, the large-scale language model (LLM), and interface settings. This customization results in a chatbot based on their expertise being built on the server.

[1464] 4. Question and Response Processing

[1465] End users input questions in real time using smart devices (such as smart glasses). These questions are sent to a server, which analyzes the received questions, extracts keywords, and searches for relevant information in a database. Based on the search results, the server uses a large-scale language model (LLM) to generate an appropriate response. This response is then sent back to the end user's smart device in real time.

[1466] 5. Specific Examples

[1467] Take a concrete example of troubleshooting a robotic arm in a factory: A worker puts on smart glasses and asks, "What caused the robotic arm to stop working? What can I do to fix it?" This question is sent to a server, which searches for relevant information in a database and uses a generative AI model to create a prompt like this:

[1468] text

[1469] Document name: maintenance_manual.txt

[1470] Q: What is the reason for the robot arm stopping and how can I fix it?

[1471] answer:

[1472] The server inputs this prompt into a large-scale language model to generate an appropriate response and provide it to the end user.

[1473] The hardware used includes smart devices (e.g., smart glasses) and servers. The software uses Python, the OpenAI API, and natural language processing algorithms. Text analysis algorithms are used for data analysis, tagging techniques are used for metadata generation, and large-scale language models (e.g., text-davinci-003) are used for response generation.

[1474] In this way, the system of the present invention provides advanced support for quickly and efficiently resolving technical problems in factory operations.

[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 uploads a data file containing expert knowledge from their device to the server. The user uses a web interface to select the data file and presses the send button. At this time, the file format is confirmed and the server temporarily stores the received file. The input is the expert knowledge data file, and the output is the data stored as a temporary file.

[1478] Step 2:

[1479] The server analyzes the uploaded data files and generates metadata. The server performs data analysis such as text extraction, keyword identification, and extraction of important items, and generates metadata such as title, author, and category. The input is data saved as a temporary file, and the output is the generated metadata.

[1480] Step 3:

[1481] The server stores the generated metadata in a database. The analyzed data and metadata are appropriately organized and saved in the database. The input is the generated metadata, and the output is the metadata stored in the database.

[1482] Step 4:

[1483] Users customize chatbots through a web interface. They select the database to use as a knowledge base, the large-scale language model (LLM), interface settings, and so on. This allows a chatbot based on expert knowledge to be built on the server. The input is customization setting information, and the output is the customized chatbot.

[1484] Step 5:

[1485] End users input questions in real time using smart devices. The questions are sent to the server from devices such as smart glasses or smartphones. The input is the question from the end user, and the output is the question sent to the server.

[1486] Step 6:

[1487] The server analyzes the received question, extracts keywords, and searches the database for related information. The server analyzes the question text and identifies keywords. It then searches for related entries in the database. The input is the received question, and the output is the related entries.

[1488] Step 7:

[1489] The server generates a response based on the search results using a large-scale language model (LLM). A prompt sentence is generated and input to a generative AI model (e.g., text-davinci-003) to get an appropriate response. The input is the relevant entry and the prompt sentence, and the output is the generated response. An example prompt sentence is:

[1490] text

[1491] Document name: maintenance_manual.txt

[1492] Q: What is the reason for the robot arm stopping and how can I fix it?

[1493] answer:

[1494] Step 8:

[1495] The server sends the generated response back to the end user's smart device, which receives the response through smart glasses or a smartphone. The input is the generated response, and the output is the response sent back to the smart device.

[1496] 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.

[1497] The system of the present invention combines an advanced chatbot that utilizes specialized knowledge with an emotion engine that recognizes the user's emotions. Specific embodiments of the system will be described below in natural language.

[1498] Upload your expertise

[1499] Users upload specialized document files from their own devices to the server. Examples include medical papers, technical manuals, and legal documents. Using a web interface, users select the files they want to upload and press the send button to begin the process. At this point, the file format is verified and saved as a temporary file on the server.

[1500] Data analysis and metadata generation

[1501] The server receives the uploaded data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting important items. As a result of the analysis, metadata such as the file title, author, and category is generated. The metadata is organized and stored in a database for search engine compatibility.

[1502] Chatbot generation

[1503] Users can customize their chatbot through a web interface, selecting the database they want to use as their knowledge base, the large-scale language model (LLM) they want to use, and interface settings, which then builds a chatbot on the server that generates responses based on their expert knowledge.

[1504] Emotion Engine Operation

[1505] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. This emotion engine uses natural language processing technology to analyze the emotional state of the user's input.

[1506] Enter a question and search

[1507] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the question is sent to the server.

[1508] The server extracts keywords from the received question and searches for related information in the database. Keyword searches and ID searches identify entries related to the question.

[1509] Response Generation

[1510] The server generates a response based on the identified search results using a large-scale language model (LLM). The LLM generates a human-like response based on expert knowledge and interpretation. It also references the user's emotional information, as recognized by the emotion engine, and adjusts the response accordingly. For example, if the user is expressing anger, the tone of the response can be toned down.

[1511] Providing a response

[1512] The server then sends the generated response back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a function to provide feedback, which is used to improve the accuracy of future responses.

[1513] Specific examples

[1514] Example 1: A chatbot using medical knowledge and emotion recognition

[1515] 1. The user uploads medical paper data from their device to the server.

[1516] 2. The server analyzes the data, generates metadata, and stores it in a database.

[1517] 3. The user configures the "medical chatbot" through a web interface.

[1518] 4. When an end user asks, "Tell me about the latest cancer treatments," the comment also includes an emotional comment such as, "I'm worried."

[1519] 5. The server's emotion engine recognizes the user's "concern."

[1520] 6. The server searches the database for relevant information, and the LLM generates a response, softening the tone of the response based on the emotional information.

[1521] 7. The server generates a response and sends it back to the end user.

[1522] Example 2: Legal advice and emotion recognition chatbot

[1523] 1. A user uploads a legal document to the server.

[1524] 2. The server analyzes the data, generates metadata, and stores it in a database.

[1525] 3. Users customize the "Legal Advice Chatbot."

[1526] 4. When an end user asks, "What legal details are required for a new business contract?", the comment also includes an emotional response: "I'm worried."

[1527] 5. The server's emotion engine recognizes the user's "anxiety."

[1528] 6. The server retrieves the information from relevant documents, and the LLM generates an appropriate response, adjusting the tone of the response to a reassuring one based on the emotional information.

[1529] 7. The server generates a response and sends it back to the end user.

[1530] As described above, the system of the present invention not only efficiently utilizes specialized knowledge, but also recognizes the user's emotions and adjusts the response content, thereby realizing more personalized dialogue.

[1531] The processing flow will be explained below.

[1532] Step 1:

[1533] Users upload document files containing their specialized knowledge from their own devices to the server. Using a web interface, users select the files to upload and press the send button. At this time, the file format is checked and saved as a temporary file on the server.

[1534] Step 2:

[1535] The server receives the received data file and analyzes its contents. This analysis process involves extracting text, identifying keywords, and extracting key points. Specifically, the text is analyzed using natural language processing tools to extract key keywords and phrases.

[1536] Step 3:

[1537] The server generates metadata such as file title, author, and category from the analysis results, which are then organized and stored in a database for search engine compatibility.

[1538] Step 4:

[1539] Users customize their chatbot through a web interface, selecting the database they want to use as their knowledge base, the large-scale language model (LLM) they want to use, and interface settings—for example, choosing a medical database and the latest LLM.

[1540] Step 5:

[1541] The server integrates the configured knowledge base with the LLM model and builds a chatbot based on the specified settings, resulting in a chatbot that can respond with expert knowledge.

[1542] Step 6:

[1543] When a user or end-user types a question into the chatbot, they can include a comment related to their feelings along with the question, for example, "Tell me about this new treatment, but I'm a little nervous."

[1544] Step 7:

[1545] The server's emotion engine analyzes the user's input and identifies the emotion, for example, "anxiety," using natural language processing tools. This emotion information is then used in the response generation process.

[1546] Step 8:

[1547] The server extracts keywords from the received question and searches for related information in its database, performing keyword and ID searches to identify entries related to the question.

[1548] Step 9:

[1549] The server generates a response based on the identified search results using a large-scale language model (LLM). It also references the emotional information recognized by the emotion engine and adjusts the response accordingly. For example, it takes into account the emotional information of "anxiety" and makes the tone of the response gentler.

[1550] Step 10:

[1551] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application.

[1552] Step 11:

[1553] The user or end user can provide feedback on the returned responses, which the server collects and uses to improve future response accuracy and emotion recognition.

[1554] Specific examples

[1555] The user uploads medical paper data from their device to the server and sets up the chatbot. When the end user then asks, "Please tell me about the latest cancer treatments. I'm anxious," the server recognizes "anxiety" using its emotion engine, searches for information from the database, and uses LLM to generate a response in a gentle tone and send it back to the end user.

[1556] In this way, the system of the present invention leverages emotion recognition capabilities to deliver sophisticated, personalized responses based on expert knowledge.

[1557] Example 2

[1558] 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."

[1559] Conventional chatbot systems have difficulty efficiently utilizing expert knowledge and are unable to generate personalized responses that take into account the user's emotional state. Furthermore, they lack the ability to incorporate end-user feedback into the generated responses. Therefore, there is a need for a system that utilizes expert knowledge while providing appropriate responses that reflect the user's emotions.

[1560] 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.

[1561] In this invention, the server includes means for uploading expert knowledge from a user's terminal to the server, means for analyzing the uploaded data to generate metadata and storing it in a database, means for a user to customize the chatbot through a web interface, means for searching the database for related information in response to an input question, means for generating an appropriate response using a large-scale language model based on the search result, means for analyzing the user's emotions and adjusting the response content based on the emotional information, and means for returning the generated response to the end user's terminal. This makes it possible to generate personalized responses according to the user's emotional state and reflect the end user's feedback while efficiently utilizing expert knowledge.

[1562] "Expertise" refers to a high level of information or understanding gained in a particular field.

[1563] "User" refers to the entity that uses the system to upload specialized knowledge and customize the chatbot.

[1564] "Terminal" is a general term for computing devices and communication equipment used by users, including personal computers, smartphones, tablets, etc.

[1565] A "server" is a computing device that provides functions such as uploading specialized knowledge, analyzing data, generating metadata, building chatbots, and generating and returning responses.

[1566] The "emotion engine" is a module that uses natural language processing technology to analyze the emotional state of text input from users.

[1567] "Metadata" is data that describes information about the contents of a file, including the title, author, category, etc.

[1568] A "database" is a collection of data in which specialized knowledge and metadata are systematically stored and can be searched and used.

[1569] "Web interface" refers to the screens and input forms that allow users to interact with a system via a browser.

[1570] A "chatbot" is a program that uses large-scale language models to enable natural dialogue in response to user questions.

[1571] A "large-scale language model (LLM)" is a language model trained on large amounts of text data, and is capable of generating natural-sounding sentences like humans.

[1572] A "response" is the information or advice returned by the chatbot to the end user.

[1573] "Feedback" means comments or ratings provided by end users that are used to improve system performance and responsiveness.

[1574] "End user" refers to the final user who actually receives the response from the chatbot.

[1575] The system of the present invention provides an advanced chatbot that utilizes expert knowledge, recognizes user emotions, and adjusts its responses accordingly. The system includes a function for uploading expert knowledge from a user's device to a server, analyzing the uploaded data to generate metadata, and storing it in a database. The system then allows the user to customize the chatbot through a web interface, searches the database for relevant information in response to an input question, and generates an appropriate response using a large-scale language model (LLM) based on the search results. The system also includes a function for analyzing user emotions, adjusting the response based on the emotional information, and returning the generated response to the end user's device.

[1576] Specifically, the system works as follows:

[1577] Users upload document files containing specialized knowledge from their own devices to the server. For example, medical papers, technical manuals, legal documents, etc. The user uses a web interface to select the file to upload and press the send button. At this time, the file format is confirmed and saved as a temporary file on the server.

[1578] The server receives the uploaded data files and analyzes their contents. During this analysis process, text is extracted using Apache Tika, and natural language processing libraries such as NLTK and SpaCy are used to identify keywords and extract important information. The analysis results in metadata such as the file title, author, and category. The metadata is organized so that it can be used by search engines, and stored in a database (e.g., PostgreSQL).

[1579] Users customize their chatbot through a web interface, selecting the database they want to use as a knowledge base, a large-scale language model (e.g., OpenAI's GPT-4), and interface settings. This builds a chatbot on the server that generates responses based on their expert knowledge.

[1580] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. The emotion engine uses natural language processing techniques such as TextBlob and VADER to analyze the emotional state of the user's input.

[1581] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the question is sent to the server.

[1582] The server extracts keywords from the received question and searches for related information in the database. Keyword searches and ID searches identify entries related to the question.

[1583] The server generates a response based on the identified search results using a large-scale language model (LLM). The LLM generates a human-like response based on expert knowledge and interpretation. It also references the user's emotional information, as recognized by the emotion engine, and adjusts the response accordingly. For example, if the user is expressing anger, the tone of the response can be toned down.

[1584] The server then sends the generated response back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a function to provide feedback, which is used to improve the accuracy of future responses.

[1585] As a concrete example, consider the following scenario:

[1586] Example 1: A chatbot using medical knowledge and emotion recognition

[1587] 1. The user uploads medical paper data from their device to the server.

[1588] 2. The server analyzes the data, generates metadata, and stores it in a database.

[1589] 3. The user configures the "medical chatbot" through a web interface.

[1590] 4. When an end user asks, "Tell me about the latest cancer treatments," the comment also includes an emotional comment such as, "I'm worried."

[1591] 5. The server's emotion engine recognizes the user's "concern."

[1592] 6. The server searches the database for relevant information, and the LLM generates a response, softening the tone of the response based on the emotional information.

[1593] 7. The server generates a response and sends it back to the end user.

[1594] Example 2: Legal advice and emotion recognition chatbot

[1595] 1. A user uploads a legal document to the server.

[1596] 2. The server analyzes the data, generates metadata, and stores it in a database.

[1597] 3. Users customize the "Legal Advice Chatbot."

[1598] 4. When an end user asks, "What legal details are required for a new business contract?", the comment also includes an emotional response: "I'm worried."

[1599] 5. The server's emotion engine recognizes the user's "anxiety."

[1600] 6. The server retrieves the information from relevant documents, and the LLM generates an appropriate response, adjusting the tone of the response to a reassuring one based on the emotional information.

[1601] 7. The server generates a response and sends it back to the end user.

[1602] In this way, it is possible to efficiently utilize expert knowledge, recognize user emotions, and tailor responses to achieve more personalized interactions.

[1603] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1604] Step 1:

[1605] The user selects a document file containing the specialized knowledge and uploads it to the server from the terminal.

[1606] What happens: The user opens a file selection dialog in the web interface, selects a file to upload, and clicks the "Submit" button.

[1607] Input: A file containing expert knowledge stored on the user's device.

[1608] Output: The document file sent to the server.

[1609] Step 2:

[1610] The server checks the format of the received file and saves it as a temporary file.

[1611] Specific operation: The server verifies the format of the received file and, if it is a compatible format (e.g., PDF, DOCX), saves the file in a temporary folder.

[1612] Input: The document file submitted by the user.

[1613] Output: A temporary file saved on the server.

[1614] Step 3:

[1615] The server analyzes the saved file with a text extraction tool to extract text data.

[1616] What it does: The server uses Apache Tika to extract text from the file.

[1617] Input: A temporary file stored on the server.

[1618] Output: The extracted text data.

[1619] Step 4:

[1620] The server extracts keywords and important items from the analyzed text data.

[1621] How it works: The server uses natural language processing libraries such as NLTK and SpaCy to extract keywords and important items from text data.

[1622] Input: Extracted text data.

[1623] Output: Extracted keywords and important items.

[1624] Step 5:

[1625] The server generates metadata such as file title, author, category, etc. and stores it in a database.

[1626] Specific operation: The server constructs metadata based on the extracted information and stores it in a database (e.g., PostgreSQL).

[1627] Input: Extracted keywords and important items.

[1628] Output: Metadata stored in a database.

[1629] Step 6:

[1630] Users customize their chatbot through a web interface.

[1631] Specific behavior: The user selects the database to use as the knowledge base, the large language model (LLM), and the interface settings in the web interface.

[1632] Input: User customization settings.

[1633] Output: Customized chatbot configuration information.

[1634] Step 7:

[1635] The server builds a chatbot based on the selected configuration information.

[1636] Specific operation: The server uses the selected LLM (e.g., OpenAI's GPT-4) to build the chatbot's interface and response generation logic.

[1637] Input: Customized chatbot configuration information.

[1638] Output: The built chatbot.

[1639] Step 8:

[1640] The server uses an emotion engine to analyze emotions from the end user's input and records the results.

[1641] Specific operation: The server uses natural language processing libraries such as TextBlob and VADER to evaluate the emotional state of the input text and stores the results in internal data.

[1642] Input: End user input text.

[1643] Output: Parsed emotion data.

[1644] Step 9:

[1645] The server receives the end user's question and searches the question content in a database based on the analyzed emotional state.

[1646] How it works: The server uses NLTK or SpaCy to extract keywords from the question and search for relevant information in a database.

[1647] Input: End-user question and parsed sentiment data.

[1648] Output: The relevant information found.

[1649] Step 10:

[1650] The server generates responses based on the search results using a large-scale language model (LLM) and adjusts the tone based on sentiment data.

[1651] What it does: The server uses LLM to generate natural-sounding responses based on search results and sentiment analysis data, adjusting tone as needed.

[1652] Input: Retrieved relevant information and sentiment data.

[1653] Output: The generated response.

[1654] Step 11:

[1655] The server then sends the generated response back to the end user's terminal.

[1656] Specific behavior: The server sends the generated response as an HTTP response.

[1657] Input: The generated response.

[1658] Output: The response sent back to the end user.

[1659] Step 12:

[1660] End users receive responses and provide feedback through a web interface or dedicated application.

[1661] What happens: The end user reviews the response they receive and submits their rating and comments through the feedback form.

[1662] Input: Generated responses and end-user feedback.

[1663] Output: Feedback data stored on the server.

[1664] This enables the system to efficiently utilize specialized knowledge while creating an advanced chatbot that responds to the user's emotions.

[1665] (Application example 2)

[1666] 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."

[1667] Conventional chatbot systems simply provide factual information without considering the user's emotions. As a result, they are unable to respond appropriately even when the user is feeling anxious or fatigued, resulting in a poor user experience. The present invention aims to achieve more personalized dialogue by analyzing the user's emotions and providing responses that correspond to those emotions.

[1668] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for uploading specialized knowledge from a user's terminal to the server, means for analyzing the uploaded data to generate metadata and storing it in a database, means for the user to customize the chatbot through a web interface, means for searching for related information from a database for an input question, means for generating an appropriate response using a large-scale language model based on the search results, means for analyzing the user's emotions using an emotion engine, means for adjusting the tone of the response based on the emotion information, and means for returning the generated response to the end user's terminal. This makes it possible to provide an appropriate response according to the user's emotions.

[1669] "Expertise" is knowledge based on advanced information and experience in a particular field.

[1670] A "server" is a computer system that provides data and services to multiple client computers over a network.

[1671] "Upload" is the operation of transferring a data file from a client's computer to a server.

[1672] "Data analysis" is the process of organizing collected data and converting it into meaningful information.

[1673] "Metadata" is supplementary information that describes the content, structure, characteristics, etc. of data.

[1674] A "database" is an organized collection of data, a system that allows for efficient searching and manipulation.

[1675] "Web interface" is a collective term for web-based screens and functions that allow users to interact with a system over the Internet.

[1676] A "chatbot" is a program that uses natural language processing to automate interactions with users.

[1677] A "large-scale language model (LLM)" is a deep learning model trained on large amounts of text data that has the ability to understand and generate natural language.

[1678] An "emotion engine" is a technology that analyzes emotions from a user's statements and input and identifies their state.

[1679] "Tone of response" refers to the impression and nuance of emotion and expression contained in the generated reply.

[1680] The system embodying this invention uses user expertise to generate a chatbot, and also recognizes the user's emotions and adjusts responses. The system is composed of the following components:

[1681] Upload your expertise

[1682] The user uploads a document file containing specialized knowledge from their device to the server. The uploaded file is first checked for its format on the server, and if it is in the correct format, it is saved as a temporary file.

[1683] Data analysis and metadata generation

[1684] The server analyzes the uploaded data and generates metadata, including file title, author, category, etc. The generated metadata is stored in a database. Natural language processing (NLP) technology is used for data analysis.

[1685] Chatbot generation

[1686] Users can customize their chatbot through a web interface, selecting the database they want to use as a knowledge base and the large-scale language model (LLM) they want to use, such as GPT-3. Based on these settings, the server generates the chatbot.

[1687] Emotion Engine Operation

[1688] The server is equipped with an emotion engine that recognizes emotions from questions and comments entered by users or end users to the chatbot. The emotion engine uses natural language processing technology to analyze the emotional state of the user's input. For example, it identifies emotions such as "worry."

[1689] Enter a question and search

[1690] When an end user types a question into the chatbot, the question is first analyzed by the emotion engine to determine the user's emotional state, and then the server extracts keywords from the question and searches for relevant information in the database, using keyword or ID search.

[1691] Response Generation

[1692] The server uses a large-scale language model (LLM) based on the search results to generate an appropriate response. It also refers to the user's emotional information recognized by the emotion engine and adjusts the tone of the response accordingly. For example, if the user indicates "fatigue," the server adjusts the response to be more refreshing.

[1693] Providing a response

[1694] The server then sends the generated response back to the end user's device, who receives it through a web interface or a dedicated application. If the response does not meet the user's expectations, the system also has a feature that allows the user to provide feedback, which is used to improve the accuracy of future responses.

[1695] Specific examples

[1696] As an example, consider the case where a user uploads a document about food to the server and configures a "food delivery helper" chatbot based on that information. The end user asks, "What's your lunch recommendation today?" and adds a comment containing the emotion, "I'm a little tired." In this case, the server's emotion engine recognizes the emotion and suggests a refreshing lunch menu.

[1697] Prompt Sentence Examples

[1698] "Question: What's your recommended lunch today?

[1699] Emotions: Tired

[1700] Metadata: Nutrition research papers, popular recipes

[1701] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1702] Step 1:

[1703] The user uploads a document file containing specialized knowledge from the terminal to the server. The file selected by the user is then sent to the server. The server checks the format of the uploaded file and saves it as a temporary file. The input is the file containing specialized knowledge, and the output is the saved temporary file.

[1704] Step 2:

[1705] The server analyzes the uploaded data and generates metadata. The analysis process involves extracting text, identifying keywords, and extracting important items. From the analysis results, metadata such as title, author, and category is generated and stored in a database. The input is the saved temporary file, and the output is a database entry containing the metadata.

[1706] Step 3:

[1707] Users customize their chatbots through a web interface. They configure the database to use, the large-scale language model (LLM) to use, and the chatbot's interface. The server generates the chatbot based on this configuration. The input is the user's customization settings, and the output is the customized chatbot.

[1708] Step 4:

[1709] An end user inputs questions and comments to a chatbot, which may include comments that indicate emotions. The input is the user's questions and comments.

[1710] Step 5:

[1711] The server uses an emotion engine to analyze emotions from end-user questions and comments, and identifies emotional states using natural language processing techniques. The input is the end-user's questions and comments, and the output is the identified emotion data.

[1712] Step 6:

[1713] The server searches for related information based on the question and emotion data. It identifies related information in a database using keyword or ID search. The input is the identified emotion data and the question, and the output is a database entry containing the related information.

[1714] Step 7:

[1715] The server generates an appropriate response using a large-scale language model (LLM) based on the information retrieved from the database, and adjusts the tone of the response by referring to the emotional data. The input is relevant information and emotional data, and the output is a response tailored according to the emotion.

[1716] Step 8:

[1717] The server generates a response and sends it back to the end user's device, where the end user can receive the response through a web interface or a dedicated application. The input is the adjusted response, and the output is the response displayed on the end user's device.

[1718] Step 9:

[1719] If the end user provides feedback on a response, the server receives that feedback and uses it to improve the accuracy of future responses. The input is the end user's feedback, and the output is an improved response generation algorithm.

[1720] 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.

[1721] 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.

[1722] 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.

[1723] 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.

[1724] 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.

[1725] 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.

[1726] 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).

[1727] 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.

[1728] 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."

[1729] 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.

[1730] 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).

[1731] 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.

[1732] 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.

[1733] 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.

[1734] 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.

[1735] 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.

[1736] 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.

[1737] 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.

[1738] 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.

[1739] 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.

[1740] 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.

[1741] The following is further disclosed regarding the above embodiment.

[1742] (Claim 1)

[1743] A means for uploading expertise from the user's device to the server;

[1744] a means for analyzing the uploaded data to generate metadata and store it in a database;

[1745] A way for users to customize the chatbot through a web interface;

[1746] a means for searching a database for relevant information in response to an input question;

[1747] means for generating an appropriate response using a large-scale language model based on the search results;

[1748] means for transmitting the generated response back to the end user's terminal;

[1749] A system including:

[1750] (Claim 2)

[1751] 2. The system according to claim 1, wherein when uploading the expert knowledge data, the file format is checked and the data is saved as a temporary file.

[1752] (Claim 3)

[1753] 2. The system according to claim 1, wherein in analyzing the data, metadata such as title, author, category, etc. are generated and stored in the database.

[1754] "Example 1"

[1755] (Claim 1)

[1756] means for transmitting the expertise from the user's communications device to a server;

[1757] means for analyzing the transmitted data to generate metadata and storing the metadata in a storage device;

[1758] a means for a user to configure an information processing device through an interface;

[1759] means for extracting relevant information from a storage device in response to an input query;

[1760] a means for generating an appropriate response using a generative AI model based on the extracted results;

[1761] means for transmitting the generated response back to the end user's communications device;

[1762] A system including:

[1763] (Claim 2)

[1764] 2. The system according to claim 1, wherein, when transmitting the expert knowledge data, the format of the document is confirmed and saved as a temporary file.

[1765] (Claim 3)

[1766] 2. The system according to claim 1, wherein in analyzing the data, metadata such as title, creator, classification, etc. are generated and stored in a storage device.

[1767] "Application Example 1"

[1768] (Claim 1)

[1769] A means for uploading expertise from the user's device to the server;

[1770] a means for analyzing the uploaded data to generate metadata and store it in a database;

[1771] A way for users to customize the chatbot through a web interface;

[1772] a means for searching a database for relevant information in response to an input question;

[1773] means for generating an appropriate response using a large-scale language model based on the search results;

[1774] means for transmitting the generated response back to the end user's terminal;

[1775] A means for end users to enter questions and receive responses in real time using their smart devices;

[1776] A system including:

[1777] (Claim 2)

[1778] 2. The system according to claim 1, wherein when uploading the expert knowledge data, the file format is checked and the data is saved as a temporary file.

[1779] (Claim 3)

[1780] 2. The system according to claim 1, wherein in analyzing the data, metadata such as title, author, category, etc. are generated and stored in the database.

[1781] "Example 2: Combining Emotion Engines"

[1782] (Claim 1)

[1783] A means for uploading expertise from the user's device to the server;

[1784] a means for analyzing the uploaded data to generate metadata and store it in a database;

[1785] A way for users to customize the chatbot through a web interface;

[1786] a means for searching a database for relevant information in response to an input question;

[1787] means for generating an appropriate response using a large-scale language model based on the search results;

[1788] A means for analyzing a user's emotions and adjusting response content based on the emotional information;

[1789] means for transmitting the generated response back to the end user's terminal;

[1790] A system including:

[1791] (Claim 2)

[1792] 2. The system according to claim 1, wherein when uploading the expert knowledge data, the file format is checked and the data is saved as a temporary file.

[1793] (Claim 3)

[1794] 2. The system according to claim 1, wherein in analyzing the data, metadata such as title, author, category, etc. are generated and stored in the database.

[1795] (Claim 4)

[1796] 2. The system of claim 1, wherein the user's emotions are analyzed using natural language processing techniques.

[1797] (Claim 5)

[1798] 10. The system of claim 1, wherein end-user feedback on the generated responses is collected and used to improve the accuracy of future responses.

[1799] "Application example 2 when combining emotion engines"

[1800] (Claim 1)

[1801] A means for uploading expertise from the user's device to the server;

[1802] a means for analyzing the uploaded data to generate metadata and store it in a database;

[1803] A way for users to customize the chatbot through a web interface;

[1804] a means for searching a database for relevant information in response to an input question;

[1805] means for generating an appropriate response using a large-scale language model based on the search results;

[1806] a means for analyzing user sentiment using a sentiment engine;

[1807] a means of adjusting the tone of responses based on emotional information;

[1808] means for transmitting the generated response back to the end user's terminal;

[1809] A system including:

[1810] (Claim 2)

[1811] 2. The system according to claim 1, wherein when uploading the expert knowledge data, the file format is checked and the data is saved as a temporary file.

[1812] (Claim 3)

[1813] 2. The system according to claim 1, wherein in analyzing the data, metadata such as title, author, category, etc. are generated and stored in the database. [Explanation of symbols]

[1814] 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 uploading expertise from the user's device to the server; a means for analyzing the uploaded data to generate metadata and store it in a database; A way for users to customize the chatbot through a web interface; a means for searching a database for relevant information in response to an input question; means for generating an appropriate response using a large-scale language model based on the search results; means for transmitting the generated response back to the end user's terminal; A system including:

2. 2. The system according to claim 1, wherein, when uploading the expert knowledge data, the file format is confirmed and the data is saved as a temporary file.

3. 2. The system according to claim 1, wherein in analyzing the data, metadata such as title, author, category, etc. are generated and stored in the database.

Images