System

Abstract

A system is provided.SOLUTION: A system comprising: means for obtaining a question input by a user; means for processing the question and selecting an appropriate expert domain model; means for generating an answer using the selected expert domain model; and means for returning the generated answer to the user.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] It indicates the "problem that the invention aims to solve" and the "means for solving the problem."

[0005] Modern small and medium-sized enterprises often face situations that require specialized knowledge in areas such as finance, law, patents, and IT. However, the time and cost constraints of hiring experts present a major hurdle. Furthermore, the difficulty of quickly obtaining appropriate advice hinders the launch of new businesses and the improvement of business operations. Given this current situation, there is a need for a system that allows companies to easily obtain specialized knowledge at low cost. [Means for solving the problem]

[0006] In order to solve the above problems, the present invention provides a system that acquires a question input by a user, appropriately processes the question to select a domain model, generates an answer using the selected model, and returns the answer to the user. Specifically, the system includes the following means:

[0007] 1. A means of capturing questions entered by the user.

[0008] 2. A means of processing the question and selecting the appropriate disciplinary model.

[0009] 3. A means of generating answers using selected domain models.

[0010] 4. A means of returning the generated answer to the user.

[0011] The system also includes a means for analyzing the intent of questions using natural language processing technology, and a means for users to input questions via a chat interface and display the generated answers, allowing users to more smoothly acquire specialized knowledge.

[0012] Ok, below are some definitions of important words:

[0013] ---

[0014] A "user" is a person or organization that utilizes the system to enter questions and receive answers.

[0015] The "means for obtaining a question" refers to an interface and a function for receiving and processing question data entered by a user.

[0016] The "means for processing a question" refers to algorithms and techniques for analyzing an input question, understanding its content, and selecting an appropriate domain model.

[0017] An "expertise model" is an artificial intelligence model that generates answers to questions based on knowledge of a specific area of ​​expertise, such as finance, law, patents, or IT.

[0018] An "answer generation means" is an algorithm or technique for generating a specific answer to a user's question using a selected domain model.

[0019] The "means for returning the generated answer to the user" refers to the infrastructure and technology for transmitting the generated answer to the user's terminal and displaying it on a chat interface or the like.

[0020] "Natural language processing technology" is a computer technology that analyzes input text data and understands its intent and content.

[0021] A "chat interface" is a user interface that allows users to enter text-based questions and view responses from the system.

[0022] "User-readable format" refers to text or other presentation formatted to be easily accessible to users. [Brief explanation of the drawings]

[0023] [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

[0024] 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.

[0025] First, the terms used in the following description will be explained.

[0026] 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).

[0027] 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.

[0028] 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.

[0029] 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.

[0030] 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."

[0031] [First embodiment]

[0032] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0033] 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.

[0034] 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).

[0035] 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.

[0036] 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.

[0037] 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.

[0038] 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.

[0039] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0040] 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.

[0041] 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.

[0042] 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.

[0043] 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."

[0044] Understood. Below is the "Form for Carrying Out the Invention" from the patent specification.

[0045] ---

[0046] The present invention relates to a system that provides generative AI models specialized in four specialized fields, namely finance, legal affairs, patents, and IT, for small and medium-sized enterprises. Specific embodiments for implementing the present invention are described below.

[0047] System Overview

[0048] This system involves a series of processes: acquiring a question entered by a user, analyzing the question, selecting an appropriate domain model, generating an answer using the selected model, and returning it to the user. The system mainly involves three entities: a server, a terminal, and a user.

[0049] Program processing overview

[0050] 1. User inputs a question

[0051] User: Enters a question using the chat interface. For example, the user enters, "What are the legal procedures required to start a new business?"

[0052] 2. Submit your question

[0053] Terminal: The terminal formats the entered question appropriately and sends it over the Internet to a server.

[0054] 3. Question Analysis and Model Selection

[0055] Server: Receives the question and analyzes its intent using natural language processing technology. Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT) for the question.

[0056] 4. Answer Generation

[0057] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[0058] 5. Replying to Answers

[0059] Server: Formats the generated answer and sends it to the user's device.

[0060] Terminal: Displays the received response in the chat interface for the user to review.

[0061] Specific example explanation

[0062] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[0063] 1. User: Type a question into the chat interface: "What are the legal procedures required to start a new business?"

[0064] 2. Terminal: Receives the query and sends it to the server over the Internet.

[0065] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[0066] 4. Server: Selects a legal-focused generative AI model and passes the question to it.

[0067] 5. Generative AI model: Based on the question, it generates an answer such as, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[0068] 6. Server: Formats the generated answer and sends it to the user's device.

[0069] 7. Terminal: The received response is displayed in the chat interface, allowing the user to check the response and quickly understand the necessary steps.

[0070] In this way, this system provides an environment in which users can quickly acquire specialized knowledge at low cost, and supports the efficient operation of small and medium-sized enterprises.

[0071] The processing flow will be explained below.

[0072] Understood. Below I will explain the program process in concrete steps.

[0073] ---

[0074] Step 1:

[0075] User: Type a question into the chat interface. For example, "What are the legal steps required to start a new business?"

[0076] Step 2:

[0077] Terminal: Receives the questions entered by the user, formats them as text data, and sends the formatted data to a server over the Internet.

[0078] Step 3:

[0079] Server: Receives queries from the terminal via the receiving port. Decodes and extracts the received text data.

[0080] Step 4:

[0081] Server: Passes the decoded question data to a natural language processing (NLP) module, which analyzes the question and extracts key keywords and context.

[0082] Step 5:

[0083] NLP module: Analyzes the intent of the question, removes unnecessary information, and determines the appropriate area of ​​expertise. In this example, it determines that the question is related to "legal affairs."

[0084] Step 6:

[0085] Server: Based on the analysis results of the NLP module, select a legal-specialized generative AI model and pass the question data to the selected model.

[0086] Step 7:

[0087] Generative AI model: Generates appropriate answers based on question data. For example, it generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[0088] Step 8:

[0089] Server: Receives the generated answer, formats it in a format that is easy for the user to read, and sends the formatted answer data to the terminal.

[0090] Step 9:

[0091] Terminal: The received response data is decoded and displayed in the chat interface. The user can check the displayed response and obtain the necessary information.

[0092] ---

[0093] Through these steps, the system provides appropriate, expert answers to questions entered by users. The specific operations performed at each processing step allow users to quickly acquire expert knowledge and support the efficient operation of small and medium-sized businesses.

[0094] Example 1

[0095] 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."

[0096] Small and medium-sized enterprises (SMEs) face difficulties in obtaining prompt and appropriate answers to financial, legal, patent, and IT-related issues that require specialized knowledge. Obtaining expert advice often requires time and money, posing a significant challenge, especially for SMEs with limited resources. Furthermore, there is a lack of systems that can properly analyze questions and provide information in the most appropriate specialized field. Therefore, there is a need for a system that can provide specialized knowledge efficiently and economically.

[0097] 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.

[0098] In this invention, the server includes means for acquiring a question input by a user, means for formatting the question and sending it to the server, means for analyzing the question and selecting an appropriate domain model in the server, means for sending a prompt to the selected domain model and receiving a generated answer, and means for formatting the generated answer and sending it to the user's terminal, thereby enabling users to quickly acquire specialized knowledge at low cost.

[0099] Ok, below are some definitions of important words:

[0100] A "user" is an entity that utilizes the system to input and confirm questions.

[0101] A "question" is information that a user enters into the system using the chat interface.

[0102] A "terminal" is a hardware device that a user uses to enter questions and receive answers from a server.

[0103] A "server" is a computer system that processes questions submitted by users, selects appropriate domain expertise models, and returns generated answers to the users.

[0104] A "specialized domain model" is a generative AI model that specializes in a specific field, such as finance, law, patents, or IT.

[0105] A "generative AI model" is an artificial intelligence model that generates answers in natural language based on input prompts.

[0106] A "prompt" is an input sentence used to generate an answer for a generative AI model.

[0107] "Formatting" refers to shaping questions entered by users and answers generated by generative AI models into a form suitable for communication or display.

[0108] "Natural language processing technology" is a technology that enables computers to understand, interpret, and generate human language.

[0109] A "chat interface" is an interactive user interface that allows a user to enter questions and receive answers from a generative AI model.

[0110] An "answer" is information generated by a generative AI model based on a prompt.

[0111] This invention relates to a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, legal affairs, patents, and IT. This system efficiently acquires, analyzes, and answers user questions, and involves three entities: the user, the terminal, and the server.

[0112] System Overview

[0113] The system executes a series of processes: a user inputs a question using a chat interface, the question is received by a server, an answer is generated based on an appropriate generative AI model, and the answer is returned to the user. Details for specifically implementing the system are described below.

[0114] Hardware and software used

[0115] Hardware:

[0116] Device: A device used by a user, such as a PC, smartphone, or tablet.

[0117] Server: A server (e.g., a cloud service server) for processing questions and managing generative AI models.

[0118] software:

[0119] Chat interface: An interface for users to enter questions and view generated answers.

[0120] Natural language processing technology: Software to analyze the intent of the question (e.g., Google Cloud NLP API, IBM Watson).

[0121] Generative AI models: Artificial intelligence models for generating answers based on domain expertise (e.g., OpenAI GPT-3).

[0122] Example of a system

[0123] User Action:

[0124] Users access the chat interface from a PC or smartphone browser and enter questions such as, "Please tell me what legal procedures are required to start a new business."

[0125] Terminal behavior:

[0126] The terminal receives the user's question, formats it appropriately, and sends it to the server over the Internet using an HTTP request (for example, the POST method).

[0127] Server behavior:

[0128] The server analyzes the received question using natural language processing technology (e.g., Google Cloud NLP API or IBM Watson) and determines that the question is legal. It then selects a generative AI model (e.g., OpenAI GPT-3) specialized in the legal domain, generates a prompt (e.g., "What legal procedures are required to start a new business?"), and sends it to the AI ​​model.

[0129] Generative AI model answers:

[0130] Based on the prompt, the generative AI model generates an answer such as, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office," and sends this answer back to the server.

[0131] Sending the response from the server to the device:

[0132] The server then formats the answers received from the generative AI model and sends them to the user's device, which then displays the answers in the chat interface for the user to review.

[0133] Examples of prompt statements

[0134] Prompt: "What are the legal steps required to start a new business?"

[0135] The present invention allows small and medium-sized enterprises to acquire specialized knowledge quickly and at low cost, enabling efficient business operations. This system can handle everything from processing inquiries to providing specialized information.

[0136] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0137] Understood. Below is a detailed explanation of the system program processing flow.

[0138] Step 1: User enters question

[0139] User: Accesses the chat interface in a browser on a PC or smartphone and types in a question.

[0140] Input: A user types a question in text format, such as "What are the legal steps required to start a new business?"

[0141] Output: The question text is sent to the chat interface.

[0142] What it does: The user types a question into the chat interface using a keyboard or touchscreen.

[0143] Step 2: Format and submit your question

[0144] Terminal: Converts the user's question into an appropriate format (e.g., JSON format) and sends it to the server.

[0145] Input: The question text entered by the user.

[0146] Data processing: Format the question text into JSON format.

[0147] Output: The formatted question data is sent to the server as an HTTP request (e.g., a POST request).

[0148] What happens: The device's browser runs JavaScript code, formats the question data appropriately, and sends it to the server.

[0149] Step 3: Receiving and parsing the question

[0150] Server: Analyzes the received question using natural language processing technology and understands its intent.

[0151] Input: Formatted question data sent from the terminal.

[0152] Data Computing: Analyze the intent of the question using natural language processing technology (e.g., Google Cloud NLP API, IBM Watson).

[0153] Output: The analysis result will be the domain of expertise the question is related to. For example, it will be determined that the question is about "legal affairs."

[0154] Specific operation: The server calls the natural language processing API, analyzes the question data, and identifies the area of ​​expertise.

[0155] Step 4: Model selection and prompt generation

[0156] Server: Selects the appropriate domain model based on the analysis results and generates prompts.

[0157] Input: Analysis results (e.g., a question about "legal").

[0158] Data processing: Select a specialized domain model (e.g., a generative AI model specialized in legal matters) and generate a prompt. Create a prompt such as, "Please tell me the legal procedures required to start a new business."

[0159] Output: The selected generative AI model and the prompt.

[0160] Specific operation: The server automatically generates a prompt sentence and prepares to send it to the selected generative AI model.

[0161] Step 5: Submitting an Answer Generation Request

[0162] Server: Sends prompts to the generative AI model and requests an answer.

[0163] Input: The generated prompt statement.

[0164] Data computation: Send prompts to a generative AI model (e.g., OpenAI GPT-3) to generate an answer.

[0165] Output: The generated answer.

[0166] How it works: The server uses HTTP requests to send prompts to the generative AI model and receive response data.

[0167] Step 6: Receiving and formatting responses

[0168] Server: Formats the answers received from the generative AI model.

[0169] Input: The answer data received from the generative AI model.

[0170] Data processing: Format the received response data into an easy-to-read format (e.g., HTML format).

[0171] Output: Formatted response data.

[0172] What happens: The server organizes the response data and converts it into a format suitable for display.

[0173] Step 7: Submit and view your responses

[0174] Server: Sends the formatted response data to the user's device.

[0175] Input: Formatted response data.

[0176] Output: The answer data sent to the user's device.

[0177] Specific operation: The server sends the answer data as an HTTP response.

[0178] Terminal: Displays the response received from the server in the chat interface.

[0179] Input: The response data received from the server.

[0180] Output: The response displayed in the chat interface.

[0181] What happens: The device's browser runs the JavaScript code and displays the response in the chat interface, where the user can view the response on their screen.

[0182] The above is the specific processing flow of the system program.

[0183] (Application example 1)

[0184] 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."

[0185] In today's small and medium-sized enterprises, it is difficult to secure personnel with specialized knowledge in legal affairs, finance, patents, and IT security. As a result, strengthening corporate security and legal responses are often delayed. Furthermore, there are limitations to independently obtaining information in these specialized fields, making it difficult to take prompt and accurate measures. The present invention aims to solve these problems by providing specialized knowledge in real time and supporting the efficient and safe operation of small and medium-sized enterprises.

[0186] 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.

[0187] In this invention, the server includes means for acquiring a question input by a user, means for processing the question and selecting an appropriate specialty domain model, means for generating an answer using the selected specialty domain model, means for returning the generated answer to the user, means for inputting a question via a smart device, and means for providing advice on security measures in real time. This enables a user to easily input specialized questions via a smart device and quickly receive necessary security measures and legal advice in real time.

[0188] The "means for acquiring a question entered by a user" is an interface for receiving a question entered by a user through a smart device and transmitting the content of the question to the system.

[0189] The "means for processing the query and selecting an appropriate domain model" is a processing device for analyzing the received query and selecting the most appropriate domain model based on the content of the query.

[0190] The "means for generating an answer using a selected domain model" is an algorithm or software that uses a selected domain model to generate an answer to a user's question.

[0191] The "means for returning the generated answer to the user" refers to a communication means and interface for transmitting the generated answer to the user's smart device and displaying it.

[0192] "Means for inputting questions via a smart device" refers to a function that allows users to input questions by voice or text using a device such as a smartphone or smart glasses.

[0193] "Means for providing security advice in real time" refers to a system-wide function that provides appropriate measures and advice immediately in response to security questions.

[0194] This invention relates to a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, legal affairs, patents, and IT security. The entire system mainly involves three entities: a server, a terminal, and a user, and operates according to the following specific processing steps.

[0195] System Overview

[0196] Program processing overview

[0197] 1. User enters question:

[0198] User: Using a smart device (such as a smartphone or smart glasses), the user types a question, for example, "How can I improve the security of my company's network?"

[0199] 2. Submit your question:

[0200] Terminal: Receives the query and sends it to the server via the Internet.

[0201] 3. Question analysis and model selection:

[0202] Server: Receives the question and analyzes its intent using natural language processing technology (e.g., BERT, GPT-4). Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT security) for the question.

[0203] 4. Generate answers:

[0204] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about security, a generative AI model specializing in IT security will generate an answer such as, "To strengthen network security, it is important to first review your firewall settings and configure access control lists appropriately."

[0205] 5. Answer Response:

[0206] Server: Formats the generated answer and sends it to the user's device.

[0207] Terminal: The received answers are displayed on the interface of the smart device for the user to review.

[0208] Hardware and Software Configuration

[0209] Natural language processing engines: BERT, GPT-4

[0210] Communication API: HTTPS

[0211] Devices: Smartphones (iOS / Android), smart glasses

[0212] Specific processing explanation

[0213] 1. Data acceptance and transmission:

[0214] The device receives the question, formats it appropriately, and sends it to the server via HTTPS.

[0215] 2. Question Analysis:

[0216] The server analyzes the received question using natural language processing engines such as BERT and GPT-4, which identifies the intent of the question and selects the most appropriate domain expertise model.

[0217] 3. Answer generation:

[0218] The domain expertise model generates the best answer based on the question, for example, a security question will be answered with firewall and access control list settings.

[0219] 4. Submitting and Viewing Answers:

[0220] The server formats the generated response in JSON format and sends it to the device, which receives it and displays it on the device's interface. Furthermore, by using a voice output function, the system can be used in environments with limited visual capabilities.

[0221] Examples of concrete examples and prompts

[0222] Specific examples

[0223] A user asks a question on their smartphone: "How can I strengthen our company's network security?" The device receives the question and sends it to the server. The server analyzes the question and selects an AI model specialized in IT security. The AI ​​model generates an answer: "To strengthen network security, it is important to first review your firewall settings and configure the access control list appropriately." The generated answer is sent to the device, where the user confirms it.

[0224] Prompt Sentence Examples

[0225] "What settings or techniques would you recommend to enhance our network security?"

[0226] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0227] Step 1:

[0228] The user uses a smartphone or smart glasses to ask a question by voice or by entering text, such as "How can I improve the security of my company's network?" The entered data is converted into text on the fly and passed to the device's application.

[0229] Step 2:

[0230] The device then formats the received question appropriately and sends it to the server using the HTTPS protocol, where the question is transferred to the server in JSON format.

[0231] Step 3:

[0232] The server analyzes the received question data. First, it uses a natural language processing engine (BERT or GPT-4) to analyze the intent of the question. Specifically, it tokenizes the question text, understands the context, and identifies which domain the question relates to. For example, it extracts keywords such as "network security," "cloud," and "access control."

[0233] Step 4:

[0234] Based on the analysis results, the server selects the most appropriate specialized domain model. The server selects an appropriate model from multiple generative AI models for finance, legal affairs, patents, IT security, etc. In this case, it selects a generative AI model specialized in IT security.

[0235] Step 5:

[0236] The selected generative AI model generates an answer based on the question. The model references past data and knowledge bases to generate the best answer to the question. For example, it might generate an answer like, "To strengthen network security, it is important to first review your firewall settings and configure access control lists appropriately."

[0237] Step 6:

[0238] The server then formats the generated answer appropriately and sends it to the device in JSON format, which may include the answer as well as related references and links.

[0239] Step 7:

[0240] The device displays the received answers on the user interface. The display method can be selected as text display or voice output. The user can check the answer content on the screen and obtain the necessary information. For example, by using voice output, the content can be understood even when there are visual constraints.

[0241] Step 8:

[0242] The user can then take specific action based on the displayed answers. If necessary, they can enter further questions to obtain new information. For example, they can ask again, "Please tell me how to specifically configure the access control list."

[0243] Through this series of processing steps, small and medium-sized business users can receive expert advice in real time, enabling them to take security measures quickly and accurately.

[0244] 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.

[0245] Understood. Below is a description of the "Mode for Carrying Out the Invention" based on the claims.

[0246] ---

[0247] The present invention combines a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, law, patents, and IT, with an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described below.

[0248] System Overview

[0249] This system involves a series of processes: it acquires a question entered by a user, analyzes the question, selects an appropriate domain model, generates an answer using the selected model, and returns it to the user. Furthermore, by combining it with an emotion engine that recognizes the user's emotions, it provides a more personalized response.

[0250] The system mainly involves three entities: a server, a terminal, and a user.

[0251] Program processing overview

[0252] 1. User inputs a question

[0253] User: Enters a question using the chat interface, for example, "What are the legal steps required to start a new business?"

[0254] 2. Submit your question

[0255] Terminal: The terminal formats the entered question appropriately and sends it over the Internet to a server.

[0256] 3. Question Analysis and Model Selection

[0257] Server: Receives the question and analyzes its intent using natural language processing technology. Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT) for the question.

[0258] 4. User Emotion Recognition

[0259] Emotion engine: Analyzes the user's emotions from the text they input and other interactions, for example, to determine whether they are stressed or excited.

[0260] 5. Answer Generation

[0261] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[0262] 6. Adjust the tone of your response

[0263] Server: Adjust the tone and wording of the generated answer based on the user's emotions as recognized by the emotion engine. For example, if the user is feeling stressed, the answer will be more friendly and soothing.

[0264] 7. Replying to Answers

[0265] Server: Formats the generated answer and sends it to the user's device.

[0266] Terminal: Displays the received response in the chat interface for the user to review.

[0267] Specific example explanation

[0268] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[0269] 1. User: Type a question into the chat interface: "What are the legal procedures required to start a new business?"

[0270] 2. Terminal: Receives the query and sends it to the server over the Internet.

[0271] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[0272] 4. Server: Selects a legal-focused generative AI model and passes the question to it.

[0273] 5. Emotion engine: Analyzes the user's emotions when entering a question and determines whether the user is feeling stressed.

[0274] 6. Generative AI model: Generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[0275] 7. Server: Based on the analysis results of the emotion engine, adjust the response to a friendly tone, such as, "I know you're having a lot of trouble, but don't worry. To start a new business, you first need to create and submit a certificate of registered matters. Next, you'll need to file a notification with the tax office."

[0276] 8. Server: Formats the generated answer and sends it to the user's device.

[0277] 9. Terminal: The received response is displayed in the chat interface, allowing the user to check the response and obtain the necessary information.

[0278] In this way, the system recognizes users' emotions and provides responses tailored to each individual user's situation, enabling more personalized support, further supporting the efficient operation of small and medium-sized businesses.

[0279] The processing flow will be explained below.

[0280] Understood. Below I will explain the process in concrete steps.

[0281] ---

[0282] Step 1:

[0283] User: Type a question into the chat interface, for example, "What are the legal steps I need to take to start a new business?"

[0284] Step 2:

[0285] Terminal: Receives the question, formats it as text data, and sends the formatted data to a server over the Internet.

[0286] Step 3:

[0287] Server: Receives questions from the terminal via the receiving port. Decodes the received text data and extracts the question content.

[0288] Step 4:

[0289] Server: Passes the decoded question data to a natural language processing (NLP) module, which analyzes the question and extracts key keywords and context.

[0290] Step 5:

[0291] NLP module: Analyzes the intent of the question, removes unnecessary information, and determines the appropriate area of ​​expertise. In this example, it determines that the question is related to "legal affairs."

[0292] Step 6:

[0293] Server: Based on the analysis results of the NLP module, select a legal-specialized generative AI model and pass the question data to the selected model.

[0294] Step 7:

[0295] Emotion engine: Analyzes the user's emotions from the text they input and other interactions, for example, to determine whether they are stressed or excited.

[0296] Step 8:

[0297] Generative AI model: Generates appropriate answers based on question data. For example, it generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[0298] Step 9:

[0299] Server: Adjust the tone and wording of the generated answer based on the user's emotions as recognized by the emotion engine. For example, if the user is feeling stressed, the answer will be more friendly and soothing.

[0300] Step 10:

[0301] Server: Formats the generated answer and sends it to the user's device.

[0302] Step 11:

[0303] Terminal: The received response data is decoded and displayed in the chat interface. The user can check the displayed response and obtain the necessary information.

[0304] ---

[0305] Through these steps, the system can provide appropriate, expert answers to questions entered by users, and can also provide personalized responses that take the user's feelings into consideration. The specific operations performed at each processing step allow users to quickly acquire expert knowledge, thereby supporting the efficient operation of small and medium-sized businesses.

[0306] Example 2

[0307] 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."

[0308] Conventional technologies select domain expertise models to provide appropriate answers to questions entered by users, but they lack personalized responses based on the user's emotions and do not fully consider the stressful situations faced by small and medium-sized business owners, particularly those faced with complex problems. This reduces user satisfaction when obtaining information and hinders efficient operations.

[0309] 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.

[0310] In this invention, the server includes means for acquiring a question input by a user, means for processing the question and selecting an appropriate domain model, means for generating an answer using the selected domain model, means for adjusting the tone of the generated answer, and means for returning the generated answer to the user, thereby enabling a personalized answer that takes into account the user's emotions.

[0311] The "means for acquiring a question entered by a user" refers to a device or software for receiving a question entered by a user using a chat interface or the like and capturing it as data.

[0312] The "means for processing questions and selecting an appropriate specialized field model" refers to a device or software that analyzes the received question using natural language processing technology and selects a model corresponding to a specialized field such as finance, law, patents, or IT based on the content of the question.

[0313] A "means for generating an answer using a selected domain-specific model" is a device or software that executes a domain-specific model to generate an appropriate answer using a pre-trained dataset based on the analysis of the question.

[0314] The "means for adjusting the tone of the generated response" refers to a device or software that modifies the expression and style of the generated response content to match the user's emotional state, and presents it in a form that is familiar and easy for the user to understand.

[0315] The "means for returning the generated response to the user" is a device or software for transmitting the tone-adjusted response to the user's terminal and ultimately displaying it on the user's chat interface.

[0316] "Natural language processing technology" is a technology that enables computers to understand, interpret, and generate human language, and specifically, it is a technology that includes algorithms and models that analyze and generate text.

[0317] A "chat interface" is a user interface for users to input text, and is a platform for entering questions and displaying answers.

[0318] An "emotion engine" is a software module that analyzes the emotions a user feels from the text entered by the user and other interaction data.

[0319] A "data packet" is a small unit of data used when transmitting data over a network such as the Internet.

[0320] The present invention is a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, law, patents, and IT. It also combines an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described below.

[0321] Overall system configuration

[0322] The system involves a series of processes: it acquires a question entered by a user, analyzes the question, selects an appropriate domain model, generates an answer using the selected model, and returns it to the user. Furthermore, it combines an emotion engine that recognizes the user's emotions to provide a more personalized response.

[0323] The system mainly uses the following hardware and software:

[0324] Server: Responsible for question analysis, domain model selection, answer generation, and tone adjustment.

[0325] Terminal: Enter questions and display answers.

[0326] Emotion engine: A software module that analyzes the user's emotions.

[0327] Natural language processing technology: Techniques for analyzing the intent of questions (e.g., BERT, GPT-3).

[0328] Processing flow, data processing and data calculation

[0329] 1. User inputs a question

[0330] The user uses the chat interface to input a question, for example, "Please tell me the legal procedures required to start a new business." This question is captured as data on the device.

[0331] 2. Submit your question

[0332] The terminal converts the question into an appropriate format (e.g., JSON format) and sends it to a server over the Internet.

[0333] 3. Question Analysis and Domain Model Selection

[0334] The server analyzes the received question data, using natural language processing technology (such as BERT or GPT-3) to analyze the content and intent of the question, and based on the results, selects the domain model (finance, legal, patent, or IT) that best suits the question.

[0335] 4. Recognition of user emotions using an emotion engine

[0336] The emotion engine installed on the server recognizes the user's emotions from the text they input and other interaction data, analyzing their emotional state such as whether they are stressed, nervous, excited, etc.

[0337] 5. Answer Generation

[0338] The selected domain-specific model generates an appropriate answer to the question. For example, if the question is about legal matters, a legal-specific generative AI model will generate the answer. The generation process uses a pre-trained dataset to provide a specific answer based on the user's question.

[0339] 6. Adjust the tone of your response

[0340] Instead of generating responses as is, the emotion engine adjusts the tone and expression based on the user's emotional state. For example, if the user is feeling stressed, the response will be modified to be more friendly and soothing.

[0341] 7. Replying to Answers

[0342] The server converts the tone-adjusted response into an appropriate format and sends it to the user's terminal, which displays the received response in a chat interface for the user to review.

[0343] Specific example explanation

[0344] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[0345] User: Type a question into the chat interface: "What are the legal steps required to start a new business?"

[0346] Terminal: Receives the query and sends it to the server over the Internet.

[0347] Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[0348] Server: Selects a legal-focused generative AI model and passes the question to it.

[0349] Emotion engine: Analyzes the user's emotions and determines whether the user is feeling stressed.

[0350] Generative AI model: Generates the answer, "To start a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[0351] Server: Based on the analysis results of the emotion engine, adjust the response to a more friendly tone: "We understand that you are facing many challenges, but please rest assured. To start a new business, you must first create and submit a certificate of registered matters. Next, you will need to file a notification with the tax office."

[0352] Server: Formats the generated answer and sends it to the user's device.

[0353] Terminal: The received response is displayed in the chat interface, allowing the user to check the response and obtain the necessary information.

[0354] In this way, the system recognizes users' emotions and provides responses tailored to each individual user's situation, enabling more personalized support, which can help small and medium-sized businesses operate more efficiently.

[0355] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0356] Step 1:

[0357] User question input

[0358] User: Uses the chat interface to type a question, for example, "What are the legal steps required to start a new business?"

[0359] Input: Text data entered by the user.

[0360] Output: The retrieved question text.

[0361] Specific operation: When the user enters text in the chat box and presses the "Send" button, the text data is sent to the terminal.

[0362] Step 2:

[0363] Submit a Question

[0364] Terminal: Converts the question into an appropriate format, such as JSON, and sends it to a server over the Internet.

[0365] Input: The retrieved question text.

[0366] Output: Formatted question data.

[0367] Specific operation: Converts text data into JSON format and sends it to the server using an HTTP request.

[0368] Step 3:

[0369] Question analysis and domain model selection

[0370] Server: Analyzes the received question data. It uses natural language processing technology (e.g., BERT, GPT-3, etc.) to analyze the content and intent of the question, and based on the results, selects the most appropriate domain model (finance, legal, patent, or IT).

[0371] Input: Formatted question data.

[0372] Output: Information about the question intent and the selected domain model.

[0373] What it does: The natural language processing model analyzes the question text and extracts its intent. Based on the results, it selects the appropriate domain model.

[0374] Step 4:

[0375] Recognizing user emotions with an emotion engine

[0376] Server: The emotion engine analyzes the user's input text and determines the user's emotional state (stress, tension, excitement, etc.).

[0377] Input: The text entered by the user.

[0378] Output: User emotion data.

[0379] What it does: Sentiment analysis algorithms analyze text data and assign emotional tags (e.g., stress, relief, etc.).

[0380] Step 5:

[0381] Generate answers

[0382] Server: The selected domain model generates an appropriate answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "To start a new business, you must first create and submit a certificate of registered matters. Next, you will need to file a notification with the tax office."

[0383] Input: Question intent and selected domain model information.

[0384] Output: The generated answer.

[0385] What it does: The domain-specific model generates answer text based on a pre-trained dataset.

[0386] Step 6:

[0387] Adjusting the tone of your response

[0388] Server: Adjusts the tone and expression of the generated answer based on the user's emotional state as recognized by the emotion engine. If the user is feeling stressed, the answer will be modified to be more friendly and soothing.

[0389] Input: Generated answers and user sentiment data.

[0390] Output: Tone-adjusted answer.

[0391] What it does: Text processing algorithms analyze the generated answers and modify them to suit the user's sentiment and style.

[0392] Step 7:

[0393] Response to the answer

[0394] Server: Sends the formatted answer to the user's device.

[0395] Terminal: Displays the received response in the chat interface for the user to review.

[0396] Input: Tone-adjusted answer.

[0397] Output: The response displayed in the chat interface.

[0398] Specific operation: The server sends the answer data to the device as an HTTP response, and the device displays the received data in the chat box.

[0399] (Application example 2)

[0400] 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."

[0401] With conventional systems, it is difficult for small and medium-sized enterprises (SMEs) to quickly obtain appropriate information when seeking specialized knowledge in finance, law, patents, or IT. Furthermore, because the system responds uniformly without considering the user's feelings, it is difficult to provide appropriate support tailored to the user's situation. As a result, user satisfaction declines and effective communication becomes difficult.

[0402] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring a question entered by a user, means for processing the question and selecting an appropriate expertise domain model, means for generating an answer using the selected expertise domain model, means for recognizing the user's emotion, means for adjusting the tone of the generated answer based on the recognized emotion, and means for returning the generated answer to the user. This makes it possible to quickly provide the expert knowledge desired by the user and to provide a personalized response according to the user's emotion.

[0403] The "means for acquiring a user-input question" is an interface or device for acquiring text information entered by a user.

[0404] The "means for processing the question and selecting the appropriate domain-specific model" refers to an algorithm or device that analyzes the received question and selects the most appropriate domain-specific generative AI model based on the question's content.

[0405] A "means for generating an answer using a selected domain model" is a system or device that uses a selected generative AI model to create an appropriate answer to a user's question.

[0406] A "means for recognizing user emotions" is an algorithm or device that extracts and analyzes a user's emotional state from text entered by the user and other interactions.

[0407] "Means for adjusting the tone of the generated response based on the recognized emotion" refers to an algorithm or device that appropriately changes the expression or nuance of the generated response depending on the user's emotional state.

[0408] The "means for returning the generated answer to the user" is an interface or device for displaying or communicating the generated answer to the user.

[0409] This invention is a system that combines a system that provides generative AI models specialized in the fields of finance, law, patents, and IT for small and medium-sized enterprises with an emotion engine that recognizes user emotions.

[0410] System components and hardware / software

[0411] Hardware:

[0412] Server: A server with high-performance computing power

[0413] User devices: smartphones, tablets, PCs

[0414] software:

[0415] Natural language processing technology: Generative AI models such as GPT-4

[0416] Emotion recognition technology: EmotionAPI

[0417] System Operation Overview

[0418] A way to capture the question the user enters

[0419] Users can input questions through a chat interface on their smartphone or PC, such as, "Have you obtained a patent for your new product?"

[0420] A means of processing questions and selecting appropriate subject-matter models

[0421] The server uses natural language processing technology to analyze questions sent from the device to the server, and based on this analysis, selects an appropriate domain model, such as legal, financial, patent, or IT.

[0422] A means of generating answers using selected domain models

[0423] The selected generative AI model generates a detailed answer to the question, for example, for a patent-related question, it might generate an answer such as, "The patent for the new product has already been obtained and granted by the United States Patent and Trademark Office (USPTO)."

[0424] A means of recognizing user emotions

[0425] The emotion engine recognizes the user's emotions from the questions entered by the user and the context of the conversation. For example, if the user is feeling anxious, it can analyze that emotion.

[0426] A means to adjust the tone of generated responses based on perceived sentiment

[0427] The emotion engine adjusts the wording of the generated answer based on the user's emotions. For example, if the user is feeling anxious, the answer will be reassuringly phrased, such as "Don't worry, we've already obtained a patent for our new product."

[0428] A means of returning the generated answer to the user

[0429] The final adjusted answer is sent from the server to the device and displayed on the chat interface, allowing the user to view it and obtain the necessary information.

[0430] Specific example explanation

[0431] For example, if a user enters a question in a virtual store such as "Please tell me the patent status of this new product," the following processing will occur.

[0432] 1. User: Type a question into the chat interface: "What is the patent status of this new product?"

[0433] 2. Terminal: Receives the query and sends it to the server over the Internet.

[0434] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a question about a patent.

[0435] 4. Server: Selects a patent-specialized generative AI model and passes the query to it.

[0436] 5. Emotion engine: Analyzes the user's emotions when entering a question and determines whether the user is feeling anxious.

[0437] 6. Generative AI model: Generates answers such as "The new product has already been patented."

[0438] 7. Server: Based on the analysis of the emotion engine, adjust the response to a tone like "Don't worry, we've already patented our new product."

[0439] 8. Server: Formats the generated answer and sends it to the user's device.

[0440] 9. Terminal: The received response is displayed in the chat interface for the user to review.

[0441] Prompt Sentence Examples

[0442] Specific examples of prompts are as follows:

[0443] "Please tell me the patent status of this new product."

[0444] In this way, by combining a generative AI model with an emotion recognition engine, it is possible to provide optimal information according to the user's situation. This invention can support the efficient operation of small and medium-sized enterprises and improve user satisfaction.

[0445] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0446] Step 1:

[0447] User: Type a question into the chat interface, for example, "What is the patent status of this new product?"

[0448] Input: User question text

[0449] Output: The question text is sent to the terminal

[0450] Step 2:

[0451] Terminal: Receives the question text entered by the user.

[0452] Input: Question text

[0453] Output: The question text is converted into the appropriate format and sent to the server.

[0454] Step 3:

[0455] Server: Analyzes the received question using natural language processing techniques, for example, using text analysis to identify the subject of the question.

[0456] Input: Question text

[0457] Output: An appropriate domain model is selected based on the subject and intent of the question.

[0458] Step 4:

[0459] Server: Based on the results of the question analysis, selects an appropriate domain-specific model (e.g., a generative AI model specializing in patents).

[0460] Input: Parsed result of question

[0461] Output: Selected discipline model

[0462] Step 5:

[0463] Server: Uses the selected model to generate an answer to the question. The generated answer contains specialized information about the user's question.

[0464] Input: Question text and selected model

[0465] Output: Generated answer text

[0466] Step 6:

[0467] Server: Recognizes the user's emotions using an emotion engine. Extracts emotion data from the user's input text and identifies emotions such as anxiety, excitement, and stress.

[0468] Input: Question text

[0469] Output: User emotion data

[0470] Step 7:

[0471] Server: Adjust the tone of the generated answer based on the perceived emotion. For example, if the user is stressed, change the answer to a gentler tone.

[0472] Input: Generated answer text and sentiment data

[0473] Output: Tone-adjusted answer text

[0474] Step 8:

[0475] Server: Formats the finalized response and sends it to the user's device.

[0476] Input: Tone-adjusted answer text

[0477] Output: Formatted answer text to send to terminal

[0478] Step 9:

[0479] Terminal: Displays the received response in the chat interface, allowing the user to review the response.

[0480] Input: Received response text

[0481] Output: Reply to be displayed in the chat interface

[0482] 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.

[0483] 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.

[0484] 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.

[0485] [Second embodiment]

[0486] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0487] 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.

[0488] 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).

[0489] 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.

[0490] 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.

[0491] 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).

[0492] 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.

[0493] 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.

[0494] 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.

[0495] 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.

[0496] 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.

[0497] 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."

[0498] Understood. Below is the "Form for Carrying Out the Invention" from the patent specification.

[0499] ---

[0500] The present invention relates to a system that provides generative AI models specialized in four specialized fields, namely finance, legal affairs, patents, and IT, for small and medium-sized enterprises. Specific embodiments for implementing the present invention are described below.

[0501] System Overview

[0502] This system involves a series of processes: acquiring a question entered by a user, analyzing the question, selecting an appropriate domain model, generating an answer using the selected model, and returning it to the user. The system mainly involves three entities: a server, a terminal, and a user.

[0503] Program processing overview

[0504] 1. User inputs a question

[0505] User: Enters a question using the chat interface. For example, the user enters, "What are the legal procedures required to start a new business?"

[0506] 2. Submit your question

[0507] Terminal: The terminal formats the entered question appropriately and sends it over the Internet to a server.

[0508] 3. Question Analysis and Model Selection

[0509] Server: Receives the question and analyzes its intent using natural language processing technology. Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT) for the question.

[0510] 4. Answer Generation

[0511] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[0512] 5. Replying to Answers

[0513] Server: Formats the generated answer and sends it to the user's device.

[0514] Terminal: Displays the received response in the chat interface for the user to review.

[0515] Specific example explanation

[0516] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[0517] 1. User: Type a question into the chat interface: "What are the legal procedures required to start a new business?"

[0518] 2. Terminal: Receives the query and sends it to the server over the Internet.

[0519] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[0520] 4. Server: Selects a legal-focused generative AI model and passes the question to it.

[0521] 5. Generative AI model: Based on the question, it generates an answer such as, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[0522] 6. Server: Formats the generated answer and sends it to the user's device.

[0523] 7. Terminal: The received response is displayed in the chat interface, allowing the user to check the response and quickly understand the necessary steps.

[0524] In this way, this system provides an environment in which users can quickly acquire specialized knowledge at low cost, and supports the efficient operation of small and medium-sized enterprises.

[0525] The processing flow will be explained below.

[0526] Understood. Below I will explain the program process in concrete steps.

[0527] ---

[0528] Step 1:

[0529] User: Type a question into the chat interface. For example, "What are the legal steps required to start a new business?"

[0530] Step 2:

[0531] Terminal: Receives the questions entered by the user, formats them as text data, and sends the formatted data to a server over the Internet.

[0532] Step 3:

[0533] Server: Receives queries from the terminal via the receiving port. Decodes and extracts the received text data.

[0534] Step 4:

[0535] Server: Passes the decoded question data to a natural language processing (NLP) module, which analyzes the question and extracts key keywords and context.

[0536] Step 5:

[0537] NLP module: Analyzes the intent of the question, removes unnecessary information, and determines the appropriate area of ​​expertise. In this example, it determines that the question is related to "legal affairs."

[0538] Step 6:

[0539] Server: Based on the analysis results of the NLP module, select a legal-specialized generative AI model and pass the question data to the selected model.

[0540] Step 7:

[0541] Generative AI model: Generates appropriate answers based on question data. For example, it generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[0542] Step 8:

[0543] Server: Receives the generated answer, formats it in a format that is easy for the user to read, and sends the formatted answer data to the terminal.

[0544] Step 9:

[0545] Terminal: The received response data is decoded and displayed in the chat interface. The user can check the displayed response and obtain the necessary information.

[0546] ---

[0547] Through these steps, the system provides appropriate, expert answers to questions entered by users. The specific operations performed at each processing step allow users to quickly acquire expert knowledge and support the efficient operation of small and medium-sized businesses.

[0548] Example 1

[0549] 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."

[0550] Small and medium-sized enterprises (SMEs) face difficulties in obtaining prompt and appropriate answers to financial, legal, patent, and IT-related issues that require specialized knowledge. Obtaining expert advice often requires time and money, posing a significant challenge, especially for SMEs with limited resources. Furthermore, there is a lack of systems that can properly analyze questions and provide information in the most appropriate specialized field. Therefore, there is a need for a system that can provide specialized knowledge efficiently and economically.

[0551] 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.

[0552] In this invention, the server includes means for acquiring a question input by a user, means for formatting the question and sending it to the server, means for analyzing the question and selecting an appropriate domain model in the server, means for sending a prompt to the selected domain model and receiving a generated answer, and means for formatting the generated answer and sending it to the user's terminal, thereby enabling users to quickly acquire specialized knowledge at low cost.

[0553] Ok, below are some definitions of important words:

[0554] A "user" is an entity that utilizes the system to input and confirm questions.

[0555] A "question" is information that a user enters into the system using the chat interface.

[0556] A "terminal" is a hardware device that a user uses to enter questions and receive answers from a server.

[0557] A "server" is a computer system that processes questions submitted by users, selects appropriate domain expertise models, and returns generated answers to the users.

[0558] A "specialized domain model" is a generative AI model that specializes in a specific field, such as finance, law, patents, or IT.

[0559] A "generative AI model" is an artificial intelligence model that generates answers in natural language based on input prompts.

[0560] A "prompt" is an input sentence used to generate an answer for a generative AI model.

[0561] "Formatting" refers to shaping questions entered by users and answers generated by generative AI models into a form suitable for communication or display.

[0562] "Natural language processing technology" is a technology that enables computers to understand, interpret, and generate human language.

[0563] A "chat interface" is an interactive user interface that allows a user to enter questions and receive answers from a generative AI model.

[0564] An "answer" is information generated by a generative AI model based on a prompt.

[0565] This invention relates to a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, legal affairs, patents, and IT. This system efficiently acquires, analyzes, and answers user questions, and involves three entities: the user, the terminal, and the server.

[0566] System Overview

[0567] The system executes a series of processes: a user inputs a question using a chat interface, the question is received by a server, an answer is generated based on an appropriate generative AI model, and the answer is returned to the user. Details for specifically implementing the system are described below.

[0568] Hardware and software used

[0569] Hardware:

[0570] Device: A device used by a user, such as a PC, smartphone, or tablet.

[0571] Server: A server (e.g., a cloud service server) for processing questions and managing generative AI models.

[0572] software:

[0573] Chat interface: An interface for users to enter questions and view generated answers.

[0574] Natural language processing technology: Software to analyze the intent of the question (e.g., Google Cloud NLP API, IBM Watson).

[0575] Generative AI models: Artificial intelligence models for generating answers based on domain expertise (e.g., OpenAI GPT-3).

[0576] Example of a system

[0577] User Action:

[0578] Users access the chat interface from a PC or smartphone browser and enter questions such as, "Please tell me what legal procedures are required to start a new business."

[0579] Terminal behavior:

[0580] The terminal receives the user's question, formats it appropriately, and sends it to the server over the Internet using an HTTP request (for example, the POST method).

[0581] Server behavior:

[0582] The server analyzes the received question using natural language processing technology (e.g., Google Cloud NLP API or IBM Watson) and determines that the question is legal. It then selects a generative AI model (e.g., OpenAI GPT-3) specialized in the legal domain, generates a prompt (e.g., "What legal procedures are required to start a new business?"), and sends it to the AI ​​model.

[0583] Generative AI model answers:

[0584] Based on the prompt, the generative AI model generates an answer such as, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office," and sends this answer back to the server.

[0585] Sending the response from the server to the device:

[0586] The server then formats the answers received from the generative AI model and sends them to the user's device, which then displays the answers in the chat interface for the user to review.

[0587] Examples of prompt statements

[0588] Prompt: "What are the legal steps required to start a new business?"

[0589] The present invention allows small and medium-sized enterprises to acquire specialized knowledge quickly and at low cost, enabling efficient business operations. This system can handle everything from processing inquiries to providing specialized information.

[0590] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0591] Understood. Below is a detailed explanation of the system program processing flow.

[0592] Step 1: User enters question

[0593] User: Accesses the chat interface in a browser on a PC or smartphone and types in a question.

[0594] Input: A user types a question in text format, such as "What are the legal steps required to start a new business?"

[0595] Output: The question text is sent to the chat interface.

[0596] What it does: The user types a question into the chat interface using a keyboard or touchscreen.

[0597] Step 2: Format and submit your question

[0598] Terminal: Converts the user's question into an appropriate format (e.g., JSON format) and sends it to the server.

[0599] Input: The question text entered by the user.

[0600] Data processing: Format the question text into JSON format.

[0601] Output: The formatted question data is sent to the server as an HTTP request (e.g., a POST request).

[0602] What happens: The device's browser runs JavaScript code, formats the question data appropriately, and sends it to the server.

[0603] Step 3: Receiving and parsing the question

[0604] Server: Analyzes the received question using natural language processing technology and understands its intent.

[0605] Input: Formatted question data sent from the terminal.

[0606] Data Computing: Analyze the intent of the question using natural language processing technology (e.g., Google Cloud NLP API, IBM Watson).

[0607] Output: The analysis result will be the domain of expertise the question is related to. For example, it will be determined that the question is about "legal affairs."

[0608] Specific operation: The server calls the natural language processing API, analyzes the question data, and identifies the area of ​​expertise.

[0609] Step 4: Model selection and prompt generation

[0610] Server: Selects the appropriate domain model based on the analysis results and generates prompts.

[0611] Input: Analysis results (e.g., a question about "legal").

[0612] Data processing: Select a specialized domain model (e.g., a generative AI model specialized in legal matters) and generate a prompt. Create a prompt such as, "Please tell me the legal procedures required to start a new business."

[0613] Output: The selected generative AI model and the prompt.

[0614] Specific operation: The server automatically generates a prompt sentence and prepares to send it to the selected generative AI model.

[0615] Step 5: Submitting an Answer Generation Request

[0616] Server: Sends prompts to the generative AI model and requests an answer.

[0617] Input: The generated prompt statement.

[0618] Data computation: Send prompts to a generative AI model (e.g., OpenAI GPT-3) to generate an answer.

[0619] Output: The generated answer.

[0620] How it works: The server uses HTTP requests to send prompts to the generative AI model and receive response data.

[0621] Step 6: Receiving and formatting responses

[0622] Server: Formats the answers received from the generative AI model.

[0623] Input: The answer data received from the generative AI model.

[0624] Data processing: Format the received response data into an easy-to-read format (e.g., HTML format).

[0625] Output: Formatted response data.

[0626] What happens: The server organizes the response data and converts it into a format suitable for display.

[0627] Step 7: Submit and view your responses

[0628] Server: Sends the formatted response data to the user's device.

[0629] Input: Formatted response data.

[0630] Output: The answer data sent to the user's device.

[0631] Specific operation: The server sends the answer data as an HTTP response.

[0632] Terminal: Displays the response received from the server in the chat interface.

[0633] Input: The response data received from the server.

[0634] Output: The response displayed in the chat interface.

[0635] What happens: The device's browser runs the JavaScript code and displays the response in the chat interface, where the user can view the response on their screen.

[0636] The above is the specific processing flow of the system program.

[0637] (Application example 1)

[0638] 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."

[0639] In today's small and medium-sized enterprises, it is difficult to secure personnel with specialized knowledge in legal affairs, finance, patents, and IT security. As a result, strengthening corporate security and legal responses are often delayed. Furthermore, there are limitations to independently obtaining information in these specialized fields, making it difficult to take prompt and accurate measures. The present invention aims to solve these problems by providing specialized knowledge in real time and supporting the efficient and safe operation of small and medium-sized enterprises.

[0640] 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.

[0641] In this invention, the server includes means for acquiring a question input by a user, means for processing the question and selecting an appropriate specialty domain model, means for generating an answer using the selected specialty domain model, means for returning the generated answer to the user, means for inputting a question via a smart device, and means for providing advice on security measures in real time. This enables a user to easily input specialized questions via a smart device and quickly receive necessary security measures and legal advice in real time.

[0642] The "means for acquiring a question entered by a user" is an interface for receiving a question entered by a user through a smart device and transmitting the content of the question to the system.

[0643] The "means for processing the query and selecting an appropriate domain model" is a processing device for analyzing the received query and selecting the most appropriate domain model based on the content of the query.

[0644] The "means for generating an answer using a selected domain model" is an algorithm or software that uses a selected domain model to generate an answer to a user's question.

[0645] The "means for returning the generated answer to the user" refers to a communication means and interface for transmitting the generated answer to the user's smart device and displaying it.

[0646] "Means for inputting questions via a smart device" refers to a function that allows users to input questions by voice or text using a device such as a smartphone or smart glasses.

[0647] "Means for providing security advice in real time" refers to a system-wide function that provides appropriate measures and advice immediately in response to security questions.

[0648] This invention relates to a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, legal affairs, patents, and IT security. The entire system mainly involves three entities: a server, a terminal, and a user, and operates according to the following specific processing steps.

[0649] System Overview

[0650] Program processing overview

[0651] 1. User enters question:

[0652] User: Using a smart device (such as a smartphone or smart glasses), the user types a question, for example, "How can I improve the security of my company's network?"

[0653] 2. Submit your question:

[0654] Terminal: Receives the query and sends it to the server via the Internet.

[0655] 3. Question analysis and model selection:

[0656] Server: Receives the question and analyzes its intent using natural language processing technology (e.g., BERT, GPT-4). Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT security) for the question.

[0657] 4. Generate answers:

[0658] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about security, a generative AI model specializing in IT security will generate an answer such as, "To strengthen network security, it is important to first review your firewall settings and configure access control lists appropriately."

[0659] 5. Answer Response:

[0660] Server: Formats the generated answer and sends it to the user's device.

[0661] Terminal: The received answers are displayed on the interface of the smart device for the user to review.

[0662] Hardware and Software Configuration

[0663] Natural language processing engines: BERT, GPT-4

[0664] Communication API: HTTPS

[0665] Devices: Smartphones (iOS / Android), smart glasses

[0666] Specific processing explanation

[0667] 1. Data acceptance and transmission:

[0668] The device receives the question, formats it appropriately, and sends it to the server via HTTPS.

[0669] 2. Question Analysis:

[0670] The server analyzes the received question using natural language processing engines such as BERT and GPT-4, which identifies the intent of the question and selects the most appropriate domain expertise model.

[0671] 3. Answer generation:

[0672] The domain expertise model generates the best answer based on the question, for example, a security question will be answered with firewall and access control list settings.

[0673] 4. Submitting and Viewing Answers:

[0674] The server formats the generated response in JSON format and sends it to the device, which receives it and displays it on the device's interface. Furthermore, by using a voice output function, the system can be used in environments with limited visual capabilities.

[0675] Examples of concrete examples and prompts

[0676] Specific examples

[0677] A user asks a question on their smartphone: "How can I strengthen our company's network security?" The device receives the question and sends it to the server. The server analyzes the question and selects an AI model specialized in IT security. The AI ​​model generates an answer: "To strengthen network security, it is important to first review your firewall settings and configure the access control list appropriately." The generated answer is sent to the device, where the user confirms it.

[0678] Prompt Sentence Examples

[0679] "What settings or techniques would you recommend to enhance our network security?"

[0680] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0681] Step 1:

[0682] The user uses a smartphone or smart glasses to ask a question by voice or by entering text, such as "How can I improve the security of my company's network?" The entered data is converted into text on the fly and passed to the device's application.

[0683] Step 2:

[0684] The device then formats the received question appropriately and sends it to the server using the HTTPS protocol, where the question is transferred to the server in JSON format.

[0685] Step 3:

[0686] The server analyzes the received question data. First, it uses a natural language processing engine (BERT or GPT-4) to analyze the intent of the question. Specifically, it tokenizes the question text, understands the context, and identifies which domain the question relates to. For example, it extracts keywords such as "network security," "cloud," and "access control."

[0687] Step 4:

[0688] Based on the analysis results, the server selects the most appropriate specialized domain model. The server selects an appropriate model from multiple generative AI models for finance, legal affairs, patents, IT security, etc. In this case, it selects a generative AI model specialized in IT security.

[0689] Step 5:

[0690] The selected generative AI model generates an answer based on the question. The model references past data and knowledge bases to generate the best answer to the question. For example, it might generate an answer like, "To strengthen network security, it is important to first review your firewall settings and configure access control lists appropriately."

[0691] Step 6:

[0692] The server then formats the generated answer appropriately and sends it to the device in JSON format, which may include the answer as well as related references and links.

[0693] Step 7:

[0694] The device displays the received answers on the user interface. The display method can be selected as text display or voice output. The user can check the answer content on the screen and obtain the necessary information. For example, by using voice output, the content can be understood even when there are visual constraints.

[0695] Step 8:

[0696] The user can then take specific action based on the displayed answers. If necessary, they can enter further questions to obtain new information. For example, they can ask again, "Please tell me how to specifically configure the access control list."

[0697] Through this series of processing steps, small and medium-sized business users can receive expert advice in real time, enabling them to take security measures quickly and accurately.

[0698] 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.

[0699] Understood. Below is a description of the "Mode for Carrying Out the Invention" based on the claims.

[0700] ---

[0701] The present invention combines a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, law, patents, and IT, with an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described below.

[0702] System Overview

[0703] This system involves a series of processes: it acquires a question entered by a user, analyzes the question, selects an appropriate domain model, generates an answer using the selected model, and returns it to the user. Furthermore, by combining it with an emotion engine that recognizes the user's emotions, it provides a more personalized response.

[0704] The system mainly involves three entities: a server, a terminal, and a user.

[0705] Program processing overview

[0706] 1. User inputs a question

[0707] User: Enters a question using the chat interface, for example, "What are the legal steps required to start a new business?"

[0708] 2. Submit your question

[0709] Terminal: The terminal formats the entered question appropriately and sends it over the Internet to a server.

[0710] 3. Question Analysis and Model Selection

[0711] Server: Receives the question and analyzes its intent using natural language processing technology. Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT) for the question.

[0712] 4. User Emotion Recognition

[0713] Emotion engine: Analyzes the user's emotions from the text they input and other interactions, for example, to determine whether they are stressed or excited.

[0714] 5. Answer Generation

[0715] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[0716] 6. Adjust the tone of your response

[0717] Server: Adjust the tone and wording of the generated answer based on the user's emotions as recognized by the emotion engine. For example, if the user is feeling stressed, the answer will be more friendly and soothing.

[0718] 7. Replying to Answers

[0719] Server: Formats the generated answer and sends it to the user's device.

[0720] Terminal: Displays the received response in the chat interface for the user to review.

[0721] Specific example explanation

[0722] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[0723] 1. User: Type a question into the chat interface: "What are the legal procedures required to start a new business?"

[0724] 2. Terminal: Receives the query and sends it to the server over the Internet.

[0725] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[0726] 4. Server: Selects a legal-focused generative AI model and passes the question to it.

[0727] 5. Emotion engine: Analyzes the user's emotions when entering a question and determines whether the user is feeling stressed.

[0728] 6. Generative AI model: Generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[0729] 7. Server: Based on the analysis results of the emotion engine, adjust the response to a friendly tone, such as, "I know you're having a lot of trouble, but don't worry. To start a new business, you first need to create and submit a certificate of registered matters. Next, you'll need to file a notification with the tax office."

[0730] 8. Server: Formats the generated answer and sends it to the user's device.

[0731] 9. Terminal: The received response is displayed in the chat interface, allowing the user to check the response and obtain the necessary information.

[0732] In this way, the system recognizes users' emotions and provides responses tailored to each individual user's situation, enabling more personalized support, further supporting the efficient operation of small and medium-sized businesses.

[0733] The processing flow will be explained below.

[0734] Understood. Below I will explain the process in concrete steps.

[0735] ---

[0736] Step 1:

[0737] User: Type a question into the chat interface, for example, "What are the legal steps I need to take to start a new business?"

[0738] Step 2:

[0739] Terminal: Receives the question, formats it as text data, and sends the formatted data to a server over the Internet.

[0740] Step 3:

[0741] Server: Receives questions from the terminal via the receiving port. Decodes the received text data and extracts the question content.

[0742] Step 4:

[0743] Server: Passes the decoded question data to a natural language processing (NLP) module, which analyzes the question and extracts key keywords and context.

[0744] Step 5:

[0745] NLP module: Analyzes the intent of the question, removes unnecessary information, and determines the appropriate area of ​​expertise. In this example, it determines that the question is related to "legal affairs."

[0746] Step 6:

[0747] Server: Based on the analysis results of the NLP module, select a legal-specialized generative AI model and pass the question data to the selected model.

[0748] Step 7:

[0749] Emotion engine: Analyzes the user's emotions from the text they input and other interactions, for example, to determine whether they are stressed or excited.

[0750] Step 8:

[0751] Generative AI model: Generates appropriate answers based on question data. For example, it generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[0752] Step 9:

[0753] Server: Adjust the tone and wording of the generated answer based on the user's emotions as recognized by the emotion engine. For example, if the user is feeling stressed, the answer will be more friendly and soothing.

[0754] Step 10:

[0755] Server: Formats the generated answer and sends it to the user's device.

[0756] Step 11:

[0757] Terminal: The received response data is decoded and displayed in the chat interface. The user can check the displayed response and obtain the necessary information.

[0758] ---

[0759] Through these steps, the system can provide appropriate, expert answers to questions entered by users, and can also provide personalized responses that take the user's feelings into consideration. The specific operations performed at each processing step allow users to quickly acquire expert knowledge, thereby supporting the efficient operation of small and medium-sized businesses.

[0760] Example 2

[0761] 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."

[0762] Conventional technologies select domain expertise models to provide appropriate answers to questions entered by users, but they lack personalized responses based on the user's emotions and do not fully consider the stressful situations faced by small and medium-sized business owners, particularly those faced with complex problems. This reduces user satisfaction when obtaining information and hinders efficient operations.

[0763] 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.

[0764] In this invention, the server includes means for acquiring a question input by a user, means for processing the question and selecting an appropriate domain model, means for generating an answer using the selected domain model, means for adjusting the tone of the generated answer, and means for returning the generated answer to the user, thereby enabling a personalized answer that takes into account the user's emotions.

[0765] The "means for acquiring a question entered by a user" refers to a device or software for receiving a question entered by a user using a chat interface or the like and capturing it as data.

[0766] The "means for processing questions and selecting an appropriate specialized field model" refers to a device or software that analyzes the received question using natural language processing technology and selects a model corresponding to a specialized field such as finance, law, patents, or IT based on the content of the question.

[0767] A "means for generating an answer using a selected domain-specific model" is a device or software that executes a domain-specific model to generate an appropriate answer using a pre-trained dataset based on the analysis of the question.

[0768] The "means for adjusting the tone of the generated response" refers to a device or software that modifies the expression and style of the generated response content to match the user's emotional state, and presents it in a form that is familiar and easy for the user to understand.

[0769] The "means for returning the generated response to the user" is a device or software for transmitting the tone-adjusted response to the user's terminal and ultimately displaying it on the user's chat interface.

[0770] "Natural language processing technology" is a technology that enables computers to understand, interpret, and generate human language, and specifically, it is a technology that includes algorithms and models that analyze and generate text.

[0771] A "chat interface" is a user interface for users to input text, and is a platform for entering questions and displaying answers.

[0772] An "emotion engine" is a software module that analyzes the emotions a user feels from the text entered by the user and other interaction data.

[0773] A "data packet" is a small unit of data used when transmitting data over a network such as the Internet.

[0774] The present invention is a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, law, patents, and IT. It also combines an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described below.

[0775] Overall system configuration

[0776] The system involves a series of processes: it acquires a question entered by a user, analyzes the question, selects an appropriate domain model, generates an answer using the selected model, and returns it to the user. Furthermore, it combines an emotion engine that recognizes the user's emotions to provide a more personalized response.

[0777] The system mainly uses the following hardware and software:

[0778] Server: Responsible for question analysis, domain model selection, answer generation, and tone adjustment.

[0779] Terminal: Enter questions and display answers.

[0780] Emotion engine: A software module that analyzes the user's emotions.

[0781] Natural language processing technology: Techniques for analyzing the intent of questions (e.g., BERT, GPT-3).

[0782] Processing flow, data processing and data calculation

[0783] 1. User inputs a question

[0784] The user uses the chat interface to input a question, for example, "Please tell me the legal procedures required to start a new business." This question is captured as data on the device.

[0785] 2. Submit your question

[0786] The terminal converts the question into an appropriate format (e.g., JSON format) and sends it to a server over the Internet.

[0787] 3. Question Analysis and Domain Model Selection

[0788] The server analyzes the received question data, using natural language processing technology (such as BERT or GPT-3) to analyze the content and intent of the question, and based on the results, selects the domain model (finance, legal, patent, or IT) that best suits the question.

[0789] 4. Recognition of user emotions using an emotion engine

[0790] The emotion engine installed on the server recognizes the user's emotions from the text they input and other interaction data, analyzing their emotional state such as whether they are stressed, nervous, excited, etc.

[0791] 5. Answer Generation

[0792] The selected domain-specific model generates an appropriate answer to the question. For example, if the question is about legal matters, a legal-specific generative AI model will generate the answer. The generation process uses a pre-trained dataset to provide a specific answer based on the user's question.

[0793] 6. Adjust the tone of your response

[0794] Instead of generating responses as is, the emotion engine adjusts the tone and expression based on the user's emotional state. For example, if the user is feeling stressed, the response will be modified to be more friendly and soothing.

[0795] 7. Replying to Answers

[0796] The server converts the tone-adjusted response into an appropriate format and sends it to the user's terminal, which displays the received response in a chat interface for the user to review.

[0797] Specific example explanation

[0798] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[0799] User: Type a question into the chat interface: "What are the legal steps required to start a new business?"

[0800] Terminal: Receives the query and sends it to the server over the Internet.

[0801] Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[0802] Server: Selects a legal-focused generative AI model and passes the question to it.

[0803] Emotion engine: Analyzes the user's emotions and determines whether the user is feeling stressed.

[0804] Generative AI model: Generates the answer, "To start a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[0805] Server: Based on the analysis results of the emotion engine, adjust the response to a more friendly tone: "We understand that you are facing many challenges, but please rest assured. To start a new business, you must first create and submit a certificate of registered matters. Next, you will need to file a notification with the tax office."

[0806] Server: Formats the generated answer and sends it to the user's device.

[0807] Terminal: The received response is displayed in the chat interface, allowing the user to check the response and obtain the necessary information.

[0808] In this way, the system recognizes users' emotions and provides responses tailored to each individual user's situation, enabling more personalized support, which can help small and medium-sized businesses operate more efficiently.

[0809] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0810] Step 1:

[0811] User question input

[0812] User: Uses the chat interface to type a question, for example, "What are the legal steps required to start a new business?"

[0813] Input: Text data entered by the user.

[0814] Output: The retrieved question text.

[0815] Specific operation: When the user enters text in the chat box and presses the "Send" button, the text data is sent to the terminal.

[0816] Step 2:

[0817] Submit a Question

[0818] Terminal: Converts the question into an appropriate format, such as JSON, and sends it to a server over the Internet.

[0819] Input: The retrieved question text.

[0820] Output: Formatted question data.

[0821] Specific operation: Converts text data into JSON format and sends it to the server using an HTTP request.

[0822] Step 3:

[0823] Question analysis and domain model selection

[0824] Server: Analyzes the received question data. It uses natural language processing technology (e.g., BERT, GPT-3, etc.) to analyze the content and intent of the question, and based on the results, selects the most appropriate domain model (finance, legal, patent, or IT).

[0825] Input: Formatted question data.

[0826] Output: Information about the question intent and the selected domain model.

[0827] What it does: The natural language processing model analyzes the question text and extracts its intent. Based on the results, it selects the appropriate domain model.

[0828] Step 4:

[0829] Recognizing user emotions with an emotion engine

[0830] Server: The emotion engine analyzes the user's input text and determines the user's emotional state (stress, tension, excitement, etc.).

[0831] Input: The text entered by the user.

[0832] Output: User emotion data.

[0833] What it does: Sentiment analysis algorithms analyze text data and assign emotional tags (e.g., stress, relief, etc.).

[0834] Step 5:

[0835] Generate answers

[0836] Server: The selected domain model generates an appropriate answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "To start a new business, you must first create and submit a certificate of registered matters. Next, you will need to file a notification with the tax office."

[0837] Input: Question intent and selected domain model information.

[0838] Output: The generated answer.

[0839] What it does: The domain-specific model generates answer text based on a pre-trained dataset.

[0840] Step 6:

[0841] Adjusting the tone of your response

[0842] Server: Adjusts the tone and expression of the generated answer based on the user's emotional state as recognized by the emotion engine. If the user is feeling stressed, the answer will be modified to be more friendly and soothing.

[0843] Input: Generated answers and user sentiment data.

[0844] Output: Tone-adjusted answer.

[0845] What it does: Text processing algorithms analyze the generated answers and modify them to suit the user's sentiment and style.

[0846] Step 7:

[0847] Response to the answer

[0848] Server: Sends the formatted answer to the user's device.

[0849] Terminal: Displays the received response in the chat interface for the user to review.

[0850] Input: Tone-adjusted answer.

[0851] Output: The response displayed in the chat interface.

[0852] Specific operation: The server sends the answer data to the device as an HTTP response, and the device displays the received data in the chat box.

[0853] (Application example 2)

[0854] 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."

[0855] With conventional systems, it is difficult for small and medium-sized enterprises (SMEs) to quickly obtain appropriate information when seeking specialized knowledge in finance, law, patents, or IT. Furthermore, because the system responds uniformly without considering the user's feelings, it is difficult to provide appropriate support tailored to the user's situation. As a result, user satisfaction declines and effective communication becomes difficult.

[0856] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring a question entered by a user, means for processing the question and selecting an appropriate expertise domain model, means for generating an answer using the selected expertise domain model, means for recognizing the user's emotion, means for adjusting the tone of the generated answer based on the recognized emotion, and means for returning the generated answer to the user. This makes it possible to quickly provide the expert knowledge desired by the user and to provide a personalized response according to the user's emotion.

[0857] The "means for acquiring a user-input question" is an interface or device for acquiring text information entered by a user.

[0858] The "means for processing the question and selecting the appropriate domain-specific model" refers to an algorithm or device that analyzes the received question and selects the most appropriate domain-specific generative AI model based on the question's content.

[0859] A "means for generating an answer using a selected domain model" is a system or device that uses a selected generative AI model to create an appropriate answer to a user's question.

[0860] A "means for recognizing user emotions" is an algorithm or device that extracts and analyzes a user's emotional state from text entered by the user and other interactions.

[0861] "Means for adjusting the tone of the generated response based on the recognized emotion" refers to an algorithm or device that appropriately changes the expression or nuance of the generated response depending on the user's emotional state.

[0862] The "means for returning the generated answer to the user" is an interface or device for displaying or communicating the generated answer to the user.

[0863] This invention is a system that combines a system that provides generative AI models specialized in the fields of finance, law, patents, and IT for small and medium-sized enterprises with an emotion engine that recognizes user emotions.

[0864] System components and hardware / software

[0865] Hardware:

[0866] Server: A server with high-performance computing power

[0867] User devices: smartphones, tablets, PCs

[0868] software:

[0869] Natural language processing technology: Generative AI models such as GPT-4

[0870] Emotion recognition technology: EmotionAPI

[0871] System Operation Overview

[0872] A way to capture the question the user enters

[0873] Users can input questions through a chat interface on their smartphone or PC, such as, "Have you obtained a patent for your new product?"

[0874] A means of processing questions and selecting appropriate subject-matter models

[0875] The server uses natural language processing technology to analyze questions sent from the device to the server, and based on this analysis, selects an appropriate domain model, such as legal, financial, patent, or IT.

[0876] A means of generating answers using selected domain models

[0877] The selected generative AI model generates a detailed answer to the question, for example, for a patent-related question, it might generate an answer such as, "The patent for the new product has already been obtained and granted by the United States Patent and Trademark Office (USPTO)."

[0878] A means of recognizing user emotions

[0879] The emotion engine recognizes the user's emotions from the questions entered by the user and the context of the conversation. For example, if the user is feeling anxious, it can analyze that emotion.

[0880] A means to adjust the tone of generated responses based on perceived sentiment

[0881] The emotion engine adjusts the wording of the generated answer based on the user's emotions. For example, if the user is feeling anxious, the answer will be reassuringly phrased, such as "Don't worry, we've already obtained a patent for our new product."

[0882] A means of returning the generated answer to the user

[0883] The final adjusted answer is sent from the server to the device and displayed on the chat interface, allowing the user to view it and obtain the necessary information.

[0884] Specific example explanation

[0885] For example, if a user enters a question in a virtual store such as "Please tell me the patent status of this new product," the following processing will occur.

[0886] 1. User: Type a question into the chat interface: "What is the patent status of this new product?"

[0887] 2. Terminal: Receives the query and sends it to the server over the Internet.

[0888] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a question about a patent.

[0889] 4. Server: Selects a patent-specialized generative AI model and passes the query to it.

[0890] 5. Emotion engine: Analyzes the user's emotions when entering a question and determines whether the user is feeling anxious.

[0891] 6. Generative AI model: Generates answers such as "The new product has already been patented."

[0892] 7. Server: Based on the analysis of the emotion engine, adjust the response to a tone like "Don't worry, we've already patented our new product."

[0893] 8. Server: Formats the generated answer and sends it to the user's device.

[0894] 9. Terminal: The received response is displayed in the chat interface for the user to review.

[0895] Prompt Sentence Examples

[0896] Specific examples of prompts are as follows:

[0897] "Please tell me the patent status of this new product."

[0898] In this way, by combining a generative AI model with an emotion recognition engine, it is possible to provide optimal information according to the user's situation. This invention can support the efficient operation of small and medium-sized enterprises and improve user satisfaction.

[0899] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0900] Step 1:

[0901] User: Type a question into the chat interface, for example, "What is the patent status of this new product?"

[0902] Input: User question text

[0903] Output: The question text is sent to the terminal

[0904] Step 2:

[0905] Terminal: Receives the question text entered by the user.

[0906] Input: Question text

[0907] Output: The question text is converted into the appropriate format and sent to the server.

[0908] Step 3:

[0909] Server: Analyzes the received question using natural language processing techniques, for example, using text analysis to identify the subject of the question.

[0910] Input: Question text

[0911] Output: An appropriate domain model is selected based on the subject and intent of the question.

[0912] Step 4:

[0913] Server: Based on the results of the question analysis, selects an appropriate domain-specific model (e.g., a generative AI model specializing in patents).

[0914] Input: Parsed result of question

[0915] Output: Selected discipline model

[0916] Step 5:

[0917] Server: Uses the selected model to generate an answer to the question. The generated answer contains specialized information about the user's question.

[0918] Input: Question text and selected model

[0919] Output: Generated answer text

[0920] Step 6:

[0921] Server: Recognizes the user's emotions using an emotion engine. Extracts emotion data from the user's input text and identifies emotions such as anxiety, excitement, and stress.

[0922] Input: Question text

[0923] Output: User emotion data

[0924] Step 7:

[0925] Server: Adjust the tone of the generated answer based on the perceived emotion. For example, if the user is stressed, change the answer to a gentler tone.

[0926] Input: Generated answer text and sentiment data

[0927] Output: Tone-adjusted answer text

[0928] Step 8:

[0929] Server: Formats the finalized response and sends it to the user's device.

[0930] Input: Tone-adjusted answer text

[0931] Output: Formatted answer text to send to terminal

[0932] Step 9:

[0933] Terminal: Displays the received response in the chat interface, allowing the user to review the response.

[0934] Input: Received response text

[0935] Output: Reply to be displayed in the chat interface

[0936] 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.

[0937] 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.

[0938] 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.

[0939] [Third embodiment]

[0940] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0941] 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.

[0942] 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).

[0943] 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.

[0944] 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.

[0945] 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).

[0946] 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.

[0947] 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.

[0948] 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.

[0949] 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.

[0950] 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.

[0951] 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."

[0952] Understood. Below is the "Form for Carrying Out the Invention" from the patent specification.

[0953] ---

[0954] The present invention relates to a system that provides generative AI models specialized in four specialized fields, namely finance, legal affairs, patents, and IT, for small and medium-sized enterprises. Specific embodiments for implementing the present invention are described below.

[0955] System Overview

[0956] This system involves a series of processes: acquiring a question entered by a user, analyzing the question, selecting an appropriate domain model, generating an answer using the selected model, and returning it to the user. The system mainly involves three entities: a server, a terminal, and a user.

[0957] Program processing overview

[0958] 1. User inputs a question

[0959] User: Enters a question using the chat interface. For example, the user enters, "What are the legal procedures required to start a new business?"

[0960] 2. Submit your question

[0961] Terminal: The terminal formats the entered question appropriately and sends it over the Internet to a server.

[0962] 3. Question Analysis and Model Selection

[0963] Server: Receives the question and analyzes its intent using natural language processing technology. Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT) for the question.

[0964] 4. Answer Generation

[0965] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[0966] 5. Replying to Answers

[0967] Server: Formats the generated answer and sends it to the user's device.

[0968] Terminal: Displays the received response in the chat interface for the user to review.

[0969] Specific example explanation

[0970] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[0971] 1. User: Type a question into the chat interface: "What are the legal procedures required to start a new business?"

[0972] 2. Terminal: Receives the query and sends it to the server over the Internet.

[0973] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[0974] 4. Server: Selects a legal-focused generative AI model and passes the question to it.

[0975] 5. Generative AI model: Based on the question, it generates an answer such as, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[0976] 6. Server: Formats the generated answer and sends it to the user's device.

[0977] 7. Terminal: The received response is displayed in the chat interface, allowing the user to check the response and quickly understand the necessary steps.

[0978] In this way, this system provides an environment in which users can quickly acquire specialized knowledge at low cost, and supports the efficient operation of small and medium-sized enterprises.

[0979] The processing flow will be explained below.

[0980] Understood. Below I will explain the program process in concrete steps.

[0981] ---

[0982] Step 1:

[0983] User: Type a question into the chat interface. For example, "What are the legal steps required to start a new business?"

[0984] Step 2:

[0985] Terminal: Receives the questions entered by the user, formats them as text data, and sends the formatted data to a server over the Internet.

[0986] Step 3:

[0987] Server: Receives queries from the terminal via the receiving port. Decodes and extracts the received text data.

[0988] Step 4:

[0989] Server: Passes the decoded question data to a natural language processing (NLP) module, which analyzes the question and extracts key keywords and context.

[0990] Step 5:

[0991] NLP module: Analyzes the intent of the question, removes unnecessary information, and determines the appropriate area of ​​expertise. In this example, it determines that the question is related to "legal affairs."

[0992] Step 6:

[0993] Server: Based on the analysis results of the NLP module, select a legal-specialized generative AI model and pass the question data to the selected model.

[0994] Step 7:

[0995] Generative AI model: Generates appropriate answers based on question data. For example, it generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[0996] Step 8:

[0997] Server: Receives the generated answer, formats it in a format that is easy for the user to read, and sends the formatted answer data to the terminal.

[0998] Step 9:

[0999] Terminal: The received response data is decoded and displayed in the chat interface. The user can check the displayed response and obtain the necessary information.

[1000] ---

[1001] Through these steps, the system provides appropriate, expert answers to questions entered by users. The specific operations performed at each processing step allow users to quickly acquire expert knowledge and support the efficient operation of small and medium-sized businesses.

[1002] Example 1

[1003] 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."

[1004] Small and medium-sized enterprises (SMEs) face difficulties in obtaining prompt and appropriate answers to financial, legal, patent, and IT-related issues that require specialized knowledge. Obtaining expert advice often requires time and money, posing a significant challenge, especially for SMEs with limited resources. Furthermore, there is a lack of systems that can properly analyze questions and provide information in the most appropriate specialized field. Therefore, there is a need for a system that can provide specialized knowledge efficiently and economically.

[1005] 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.

[1006] In this invention, the server includes means for acquiring a question input by a user, means for formatting the question and sending it to the server, means for analyzing the question and selecting an appropriate domain model in the server, means for sending a prompt to the selected domain model and receiving a generated answer, and means for formatting the generated answer and sending it to the user's terminal, thereby enabling users to quickly acquire specialized knowledge at low cost.

[1007] Ok, below are some definitions of important words:

[1008] A "user" is an entity that utilizes the system to input and confirm questions.

[1009] A "question" is information that a user enters into the system using the chat interface.

[1010] A "terminal" is a hardware device that a user uses to enter questions and receive answers from a server.

[1011] A "server" is a computer system that processes questions submitted by users, selects appropriate domain expertise models, and returns generated answers to the users.

[1012] A "specialized domain model" is a generative AI model that specializes in a specific field, such as finance, law, patents, or IT.

[1013] A "generative AI model" is an artificial intelligence model that generates answers in natural language based on input prompts.

[1014] A "prompt" is an input sentence used to generate an answer for a generative AI model.

[1015] "Formatting" refers to shaping questions entered by users and answers generated by generative AI models into a form suitable for communication or display.

[1016] "Natural language processing technology" is a technology that enables computers to understand, interpret, and generate human language.

[1017] A "chat interface" is an interactive user interface that allows a user to enter questions and receive answers from a generative AI model.

[1018] An "answer" is information generated by a generative AI model based on a prompt.

[1019] This invention relates to a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, legal affairs, patents, and IT. This system efficiently acquires, analyzes, and answers user questions, and involves three entities: the user, the terminal, and the server.

[1020] System Overview

[1021] The system executes a series of processes: a user inputs a question using a chat interface, the question is received by a server, an answer is generated based on an appropriate generative AI model, and the answer is returned to the user. Details for specifically implementing the system are described below.

[1022] Hardware and software used

[1023] Hardware:

[1024] Device: A device used by a user, such as a PC, smartphone, or tablet.

[1025] Server: A server (e.g., a cloud service server) for processing questions and managing generative AI models.

[1026] software:

[1027] Chat interface: An interface for users to enter questions and view generated answers.

[1028] Natural language processing technology: Software to analyze the intent of the question (e.g., Google Cloud NLP API, IBM Watson).

[1029] Generative AI models: Artificial intelligence models for generating answers based on domain expertise (e.g., OpenAI GPT-3).

[1030] Example of a system

[1031] User Action:

[1032] Users access the chat interface from a PC or smartphone browser and enter questions such as, "Please tell me what legal procedures are required to start a new business."

[1033] Terminal behavior:

[1034] The terminal receives the user's question, formats it appropriately, and sends it to the server over the Internet using an HTTP request (for example, the POST method).

[1035] Server behavior:

[1036] The server analyzes the received question using natural language processing technology (e.g., Google Cloud NLP API or IBM Watson) and determines that the question is legal. It then selects a generative AI model (e.g., OpenAI GPT-3) specialized in the legal domain, generates a prompt (e.g., "What legal procedures are required to start a new business?"), and sends it to the AI ​​model.

[1037] Generative AI model answers:

[1038] Based on the prompt, the generative AI model generates an answer such as, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office," and sends this answer back to the server.

[1039] Sending the response from the server to the device:

[1040] The server then formats the answers received from the generative AI model and sends them to the user's device, which then displays the answers in the chat interface for the user to review.

[1041] Examples of prompt statements

[1042] Prompt: "What are the legal steps required to start a new business?"

[1043] The present invention allows small and medium-sized enterprises to acquire specialized knowledge quickly and at low cost, enabling efficient business operations. This system can handle everything from processing inquiries to providing specialized information.

[1044] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1045] Understood. Below is a detailed explanation of the system program processing flow.

[1046] Step 1: User enters question

[1047] User: Accesses the chat interface in a browser on a PC or smartphone and types in a question.

[1048] Input: A user types a question in text format, such as "What are the legal steps required to start a new business?"

[1049] Output: The question text is sent to the chat interface.

[1050] What it does: The user types a question into the chat interface using a keyboard or touchscreen.

[1051] Step 2: Format and submit your question

[1052] Terminal: Converts the user's question into an appropriate format (e.g., JSON format) and sends it to the server.

[1053] Input: The question text entered by the user.

[1054] Data processing: Format the question text into JSON format.

[1055] Output: The formatted question data is sent to the server as an HTTP request (e.g., a POST request).

[1056] What happens: The device's browser runs JavaScript code, formats the question data appropriately, and sends it to the server.

[1057] Step 3: Receiving and parsing the question

[1058] Server: Analyzes the received question using natural language processing technology and understands its intent.

[1059] Input: Formatted question data sent from the terminal.

[1060] Data Computing: Analyze the intent of the question using natural language processing technology (e.g., Google Cloud NLP API, IBM Watson).

[1061] Output: The analysis result will be the domain of expertise the question is related to. For example, it will be determined that the question is about "legal affairs."

[1062] Specific operation: The server calls the natural language processing API, analyzes the question data, and identifies the area of ​​expertise.

[1063] Step 4: Model selection and prompt generation

[1064] Server: Selects the appropriate domain model based on the analysis results and generates prompts.

[1065] Input: Analysis results (e.g., a question about "legal").

[1066] Data processing: Select a specialized domain model (e.g., a generative AI model specialized in legal matters) and generate a prompt. Create a prompt such as, "Please tell me the legal procedures required to start a new business."

[1067] Output: The selected generative AI model and the prompt.

[1068] Specific operation: The server automatically generates a prompt sentence and prepares to send it to the selected generative AI model.

[1069] Step 5: Submitting an Answer Generation Request

[1070] Server: Sends prompts to the generative AI model and requests an answer.

[1071] Input: The generated prompt statement.

[1072] Data computation: Send prompts to a generative AI model (e.g., OpenAI GPT-3) to generate an answer.

[1073] Output: The generated answer.

[1074] How it works: The server uses HTTP requests to send prompts to the generative AI model and receive response data.

[1075] Step 6: Receiving and formatting responses

[1076] Server: Formats the answers received from the generative AI model.

[1077] Input: The answer data received from the generative AI model.

[1078] Data processing: Format the received response data into an easy-to-read format (e.g., HTML format).

[1079] Output: Formatted response data.

[1080] What happens: The server organizes the response data and converts it into a format suitable for display.

[1081] Step 7: Submit and view your responses

[1082] Server: Sends the formatted response data to the user's device.

[1083] Input: Formatted response data.

[1084] Output: The answer data sent to the user's device.

[1085] Specific operation: The server sends the answer data as an HTTP response.

[1086] Terminal: Displays the response received from the server in the chat interface.

[1087] Input: The response data received from the server.

[1088] Output: The response displayed in the chat interface.

[1089] What happens: The device's browser runs the JavaScript code and displays the response in the chat interface, where the user can view the response on their screen.

[1090] The above is the specific processing flow of the system program.

[1091] (Application example 1)

[1092] 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."

[1093] In today's small and medium-sized enterprises, it is difficult to secure personnel with specialized knowledge in legal affairs, finance, patents, and IT security. As a result, strengthening corporate security and legal responses are often delayed. Furthermore, there are limitations to independently obtaining information in these specialized fields, making it difficult to take prompt and accurate measures. The present invention aims to solve these problems by providing specialized knowledge in real time and supporting the efficient and safe operation of small and medium-sized enterprises.

[1094] 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.

[1095] In this invention, the server includes means for acquiring a question input by a user, means for processing the question and selecting an appropriate specialty domain model, means for generating an answer using the selected specialty domain model, means for returning the generated answer to the user, means for inputting a question via a smart device, and means for providing advice on security measures in real time. This enables a user to easily input specialized questions via a smart device and quickly receive necessary security measures and legal advice in real time.

[1096] The "means for acquiring a question entered by a user" is an interface for receiving a question entered by a user through a smart device and transmitting the content of the question to the system.

[1097] The "means for processing the query and selecting an appropriate domain model" is a processing device for analyzing the received query and selecting the most appropriate domain model based on the content of the query.

[1098] The "means for generating an answer using a selected domain model" is an algorithm or software that uses a selected domain model to generate an answer to a user's question.

[1099] The "means for returning the generated answer to the user" refers to a communication means and interface for transmitting the generated answer to the user's smart device and displaying it.

[1100] "Means for inputting questions via a smart device" refers to a function that allows users to input questions by voice or text using a device such as a smartphone or smart glasses.

[1101] "Means for providing security advice in real time" refers to a system-wide function that provides appropriate measures and advice immediately in response to security questions.

[1102] This invention relates to a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, legal affairs, patents, and IT security. The entire system mainly involves three entities: a server, a terminal, and a user, and operates according to the following specific processing steps.

[1103] System Overview

[1104] Program processing overview

[1105] 1. User enters question:

[1106] User: Using a smart device (such as a smartphone or smart glasses), the user types a question, for example, "How can I improve the security of my company's network?"

[1107] 2. Submit your question:

[1108] Terminal: Receives the query and sends it to the server via the Internet.

[1109] 3. Question analysis and model selection:

[1110] Server: Receives the question and analyzes its intent using natural language processing technology (e.g., BERT, GPT-4). Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT security) for the question.

[1111] 4. Generate answers:

[1112] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about security, a generative AI model specializing in IT security will generate an answer such as, "To strengthen network security, it is important to first review your firewall settings and configure access control lists appropriately."

[1113] 5. Answer Response:

[1114] Server: Formats the generated answer and sends it to the user's device.

[1115] Terminal: The received answers are displayed on the interface of the smart device for the user to review.

[1116] Hardware and Software Configuration

[1117] Natural language processing engines: BERT, GPT-4

[1118] Communication API: HTTPS

[1119] Devices: Smartphones (iOS / Android), smart glasses

[1120] Specific processing explanation

[1121] 1. Data acceptance and transmission:

[1122] The device receives the question, formats it appropriately, and sends it to the server via HTTPS.

[1123] 2. Question Analysis:

[1124] The server analyzes the received question using natural language processing engines such as BERT and GPT-4, which identifies the intent of the question and selects the most appropriate domain expertise model.

[1125] 3. Answer generation:

[1126] The domain expertise model generates the best answer based on the question, for example, a security question will be answered with firewall and access control list settings.

[1127] 4. Submitting and Viewing Answers:

[1128] The server formats the generated response in JSON format and sends it to the device, which receives it and displays it on the device's interface. Furthermore, by using a voice output function, the system can be used in environments with limited visual capabilities.

[1129] Examples of concrete examples and prompts

[1130] Specific examples

[1131] A user asks a question on their smartphone: "How can I strengthen our company's network security?" The device receives the question and sends it to the server. The server analyzes the question and selects an AI model specialized in IT security. The AI ​​model generates an answer: "To strengthen network security, it is important to first review your firewall settings and configure the access control list appropriately." The generated answer is sent to the device, where the user confirms it.

[1132] Prompt Sentence Examples

[1133] "What settings or techniques would you recommend to enhance our network security?"

[1134] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1135] Step 1:

[1136] The user uses a smartphone or smart glasses to ask a question by voice or by entering text, such as "How can I improve the security of my company's network?" The entered data is converted into text on the fly and passed to the device's application.

[1137] Step 2:

[1138] The device then formats the received question appropriately and sends it to the server using the HTTPS protocol, where the question is transferred to the server in JSON format.

[1139] Step 3:

[1140] The server analyzes the received question data. First, it uses a natural language processing engine (BERT or GPT-4) to analyze the intent of the question. Specifically, it tokenizes the question text, understands the context, and identifies which domain the question relates to. For example, it extracts keywords such as "network security," "cloud," and "access control."

[1141] Step 4:

[1142] Based on the analysis results, the server selects the most appropriate specialized domain model. The server selects an appropriate model from multiple generative AI models for finance, legal affairs, patents, IT security, etc. In this case, it selects a generative AI model specialized in IT security.

[1143] Step 5:

[1144] The selected generative AI model generates an answer based on the question. The model references past data and knowledge bases to generate the best answer to the question. For example, it might generate an answer like, "To strengthen network security, it is important to first review your firewall settings and configure access control lists appropriately."

[1145] Step 6:

[1146] The server then formats the generated answer appropriately and sends it to the device in JSON format, which may include the answer as well as related references and links.

[1147] Step 7:

[1148] The device displays the received answers on the user interface. The display method can be selected as text display or voice output. The user can check the answer content on the screen and obtain the necessary information. For example, by using voice output, the content can be understood even when there are visual constraints.

[1149] Step 8:

[1150] The user can then take specific action based on the displayed answers. If necessary, they can enter further questions to obtain new information. For example, they can ask again, "Please tell me how to specifically configure the access control list."

[1151] Through this series of processing steps, small and medium-sized business users can receive expert advice in real time, enabling them to take security measures quickly and accurately.

[1152] 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.

[1153] Understood. Below is a description of the "Mode for Carrying Out the Invention" based on the claims.

[1154] ---

[1155] The present invention combines a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, law, patents, and IT, with an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described below.

[1156] System Overview

[1157] This system involves a series of processes: it acquires a question entered by a user, analyzes the question, selects an appropriate domain model, generates an answer using the selected model, and returns it to the user. Furthermore, by combining it with an emotion engine that recognizes the user's emotions, it provides a more personalized response.

[1158] The system mainly involves three entities: a server, a terminal, and a user.

[1159] Program processing overview

[1160] 1. User inputs a question

[1161] User: Enters a question using the chat interface, for example, "What are the legal steps required to start a new business?"

[1162] 2. Submit your question

[1163] Terminal: The terminal formats the entered question appropriately and sends it over the Internet to a server.

[1164] 3. Question Analysis and Model Selection

[1165] Server: Receives the question and analyzes its intent using natural language processing technology. Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT) for the question.

[1166] 4. User Emotion Recognition

[1167] Emotion engine: Analyzes the user's emotions from the text they input and other interactions, for example, to determine whether they are stressed or excited.

[1168] 5. Answer Generation

[1169] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[1170] 6. Adjust the tone of your response

[1171] Server: Adjust the tone and wording of the generated answer based on the user's emotions as recognized by the emotion engine. For example, if the user is feeling stressed, the answer will be more friendly and soothing.

[1172] 7. Replying to Answers

[1173] Server: Formats the generated answer and sends it to the user's device.

[1174] Terminal: Displays the received response in the chat interface for the user to review.

[1175] Specific example explanation

[1176] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[1177] 1. User: Type a question into the chat interface: "What are the legal procedures required to start a new business?"

[1178] 2. Terminal: Receives the query and sends it to the server over the Internet.

[1179] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[1180] 4. Server: Selects a legal-focused generative AI model and passes the question to it.

[1181] 5. Emotion engine: Analyzes the user's emotions when entering a question and determines whether the user is feeling stressed.

[1182] 6. Generative AI model: Generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[1183] 7. Server: Based on the analysis results of the emotion engine, adjust the response to a friendly tone, such as, "I know you're having a lot of trouble, but don't worry. To start a new business, you first need to create and submit a certificate of registered matters. Next, you'll need to file a notification with the tax office."

[1184] 8. Server: Formats the generated answer and sends it to the user's device.

[1185] 9. Terminal: The received response is displayed in the chat interface, allowing the user to check the response and obtain the necessary information.

[1186] In this way, the system recognizes users' emotions and provides responses tailored to each individual user's situation, enabling more personalized support, further supporting the efficient operation of small and medium-sized businesses.

[1187] The processing flow will be explained below.

[1188] Understood. Below I will explain the process in concrete steps.

[1189] ---

[1190] Step 1:

[1191] User: Type a question into the chat interface, for example, "What are the legal steps I need to take to start a new business?"

[1192] Step 2:

[1193] Terminal: Receives the question, formats it as text data, and sends the formatted data to a server over the Internet.

[1194] Step 3:

[1195] Server: Receives questions from the terminal via the receiving port. Decodes the received text data and extracts the question content.

[1196] Step 4:

[1197] Server: Passes the decoded question data to a natural language processing (NLP) module, which analyzes the question and extracts key keywords and context.

[1198] Step 5:

[1199] NLP module: Analyzes the intent of the question, removes unnecessary information, and determines the appropriate area of ​​expertise. In this example, it determines that the question is related to "legal affairs."

[1200] Step 6:

[1201] Server: Based on the analysis results of the NLP module, select a legal-specialized generative AI model and pass the question data to the selected model.

[1202] Step 7:

[1203] Emotion engine: Analyzes the user's emotions from the text they input and other interactions, for example, to determine whether they are stressed or excited.

[1204] Step 8:

[1205] Generative AI model: Generates appropriate answers based on question data. For example, it generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[1206] Step 9:

[1207] Server: Adjust the tone and wording of the generated answer based on the user's emotions as recognized by the emotion engine. For example, if the user is feeling stressed, the answer will be more friendly and soothing.

[1208] Step 10:

[1209] Server: Formats the generated answer and sends it to the user's device.

[1210] Step 11:

[1211] Terminal: The received response data is decoded and displayed in the chat interface. The user can check the displayed response and obtain the necessary information.

[1212] ---

[1213] Through these steps, the system can provide appropriate, expert answers to questions entered by users, and can also provide personalized responses that take the user's feelings into consideration. The specific operations performed at each processing step allow users to quickly acquire expert knowledge, thereby supporting the efficient operation of small and medium-sized businesses.

[1214] Example 2

[1215] 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."

[1216] Conventional technologies select domain expertise models to provide appropriate answers to questions entered by users, but they lack personalized responses based on the user's emotions and do not fully consider the stressful situations faced by small and medium-sized business owners, particularly those faced with complex problems. This reduces user satisfaction when obtaining information and hinders efficient operations.

[1217] 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.

[1218] In this invention, the server includes means for acquiring a question input by a user, means for processing the question and selecting an appropriate domain model, means for generating an answer using the selected domain model, means for adjusting the tone of the generated answer, and means for returning the generated answer to the user, thereby enabling a personalized answer that takes into account the user's emotions.

[1219] The "means for acquiring a question entered by a user" refers to a device or software for receiving a question entered by a user using a chat interface or the like and capturing it as data.

[1220] The "means for processing questions and selecting an appropriate specialized field model" refers to a device or software that analyzes the received question using natural language processing technology and selects a model corresponding to a specialized field such as finance, law, patents, or IT based on the content of the question.

[1221] A "means for generating an answer using a selected domain-specific model" is a device or software that executes a domain-specific model to generate an appropriate answer using a pre-trained dataset based on the analysis of the question.

[1222] The "means for adjusting the tone of the generated response" refers to a device or software that modifies the expression and style of the generated response content to match the user's emotional state, and presents it in a form that is familiar and easy for the user to understand.

[1223] The "means for returning the generated response to the user" is a device or software for transmitting the tone-adjusted response to the user's terminal and ultimately displaying it on the user's chat interface.

[1224] "Natural language processing technology" is a technology that enables computers to understand, interpret, and generate human language, and specifically, it is a technology that includes algorithms and models that analyze and generate text.

[1225] A "chat interface" is a user interface for users to input text, and is a platform for entering questions and displaying answers.

[1226] An "emotion engine" is a software module that analyzes the emotions a user feels from the text entered by the user and other interaction data.

[1227] A "data packet" is a small unit of data used when transmitting data over a network such as the Internet.

[1228] The present invention is a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, law, patents, and IT. It also combines an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described below.

[1229] Overall system configuration

[1230] The system involves a series of processes: it acquires a question entered by a user, analyzes the question, selects an appropriate domain model, generates an answer using the selected model, and returns it to the user. Furthermore, it combines an emotion engine that recognizes the user's emotions to provide a more personalized response.

[1231] The system mainly uses the following hardware and software:

[1232] Server: Responsible for question analysis, domain model selection, answer generation, and tone adjustment.

[1233] Terminal: Enter questions and display answers.

[1234] Emotion engine: A software module that analyzes the user's emotions.

[1235] Natural language processing technology: Techniques for analyzing the intent of questions (e.g., BERT, GPT-3).

[1236] Processing flow, data processing and data calculation

[1237] 1. User inputs a question

[1238] The user uses the chat interface to input a question, for example, "Please tell me the legal procedures required to start a new business." This question is captured as data on the device.

[1239] 2. Submit your question

[1240] The terminal converts the question into an appropriate format (e.g., JSON format) and sends it to a server over the Internet.

[1241] 3. Question Analysis and Domain Model Selection

[1242] The server analyzes the received question data, using natural language processing technology (such as BERT or GPT-3) to analyze the content and intent of the question, and based on the results, selects the domain model (finance, legal, patent, or IT) that best suits the question.

[1243] 4. Recognition of user emotions using an emotion engine

[1244] The emotion engine installed on the server recognizes the user's emotions from the text they input and other interaction data, analyzing their emotional state such as whether they are stressed, nervous, excited, etc.

[1245] 5. Answer Generation

[1246] The selected domain-specific model generates an appropriate answer to the question. For example, if the question is about legal matters, a legal-specific generative AI model will generate the answer. The generation process uses a pre-trained dataset to provide a specific answer based on the user's question.

[1247] 6. Adjust the tone of your response

[1248] Instead of generating responses as is, the emotion engine adjusts the tone and expression based on the user's emotional state. For example, if the user is feeling stressed, the response will be modified to be more friendly and soothing.

[1249] 7. Replying to Answers

[1250] The server converts the tone-adjusted response into an appropriate format and sends it to the user's terminal, which displays the received response in a chat interface for the user to review.

[1251] Specific example explanation

[1252] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[1253] User: Type a question into the chat interface: "What are the legal steps required to start a new business?"

[1254] Terminal: Receives the query and sends it to the server over the Internet.

[1255] Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[1256] Server: Selects a legal-focused generative AI model and passes the question to it.

[1257] Emotion engine: Analyzes the user's emotions and determines whether the user is feeling stressed.

[1258] Generative AI model: Generates the answer, "To start a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[1259] Server: Based on the analysis results of the emotion engine, adjust the response to a more friendly tone: "We understand that you are facing many challenges, but please rest assured. To start a new business, you must first create and submit a certificate of registered matters. Next, you will need to file a notification with the tax office."

[1260] Server: Formats the generated answer and sends it to the user's device.

[1261] Terminal: The received response is displayed in the chat interface, allowing the user to check the response and obtain the necessary information.

[1262] In this way, the system recognizes users' emotions and provides responses tailored to each individual user's situation, enabling more personalized support, which can help small and medium-sized businesses operate more efficiently.

[1263] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1264] Step 1:

[1265] User question input

[1266] User: Uses the chat interface to type a question, for example, "What are the legal steps required to start a new business?"

[1267] Input: Text data entered by the user.

[1268] Output: The retrieved question text.

[1269] Specific operation: When the user enters text in the chat box and presses the "Send" button, the text data is sent to the terminal.

[1270] Step 2:

[1271] Submit a Question

[1272] Terminal: Converts the question into an appropriate format, such as JSON, and sends it to a server over the Internet.

[1273] Input: The retrieved question text.

[1274] Output: Formatted question data.

[1275] Specific operation: Converts text data into JSON format and sends it to the server using an HTTP request.

[1276] Step 3:

[1277] Question analysis and domain model selection

[1278] Server: Analyzes the received question data. It uses natural language processing technology (e.g., BERT, GPT-3, etc.) to analyze the content and intent of the question, and based on the results, selects the most appropriate domain model (finance, legal, patent, or IT).

[1279] Input: Formatted question data.

[1280] Output: Information about the question intent and the selected domain model.

[1281] What it does: The natural language processing model analyzes the question text and extracts its intent. Based on the results, it selects the appropriate domain model.

[1282] Step 4:

[1283] Recognizing user emotions with an emotion engine

[1284] Server: The emotion engine analyzes the user's input text and determines the user's emotional state (stress, tension, excitement, etc.).

[1285] Input: The text entered by the user.

[1286] Output: User emotion data.

[1287] What it does: Sentiment analysis algorithms analyze text data and assign emotional tags (e.g., stress, relief, etc.).

[1288] Step 5:

[1289] Generate answers

[1290] Server: The selected domain model generates an appropriate answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "To start a new business, you must first create and submit a certificate of registered matters. Next, you will need to file a notification with the tax office."

[1291] Input: Question intent and selected domain model information.

[1292] Output: The generated answer.

[1293] What it does: The domain-specific model generates answer text based on a pre-trained dataset.

[1294] Step 6:

[1295] Adjusting the tone of your response

[1296] Server: Adjusts the tone and expression of the generated answer based on the user's emotional state as recognized by the emotion engine. If the user is feeling stressed, the answer will be modified to be more friendly and soothing.

[1297] Input: Generated answers and user sentiment data.

[1298] Output: Tone-adjusted answer.

[1299] What it does: Text processing algorithms analyze the generated answers and modify them to suit the user's sentiment and style.

[1300] Step 7:

[1301] Response to the answer

[1302] Server: Sends the formatted answer to the user's device.

[1303] Terminal: Displays the received response in the chat interface for the user to review.

[1304] Input: Tone-adjusted answer.

[1305] Output: The response displayed in the chat interface.

[1306] Specific operation: The server sends the answer data to the device as an HTTP response, and the device displays the received data in the chat box.

[1307] (Application example 2)

[1308] 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."

[1309] With conventional systems, it is difficult for small and medium-sized enterprises (SMEs) to quickly obtain appropriate information when seeking specialized knowledge in finance, law, patents, or IT. Furthermore, because the system responds uniformly without considering the user's feelings, it is difficult to provide appropriate support tailored to the user's situation. As a result, user satisfaction declines and effective communication becomes difficult.

[1310] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring a question entered by a user, means for processing the question and selecting an appropriate expertise domain model, means for generating an answer using the selected expertise domain model, means for recognizing the user's emotion, means for adjusting the tone of the generated answer based on the recognized emotion, and means for returning the generated answer to the user. This makes it possible to quickly provide the expert knowledge desired by the user and to provide a personalized response according to the user's emotion.

[1311] The "means for acquiring a user-input question" is an interface or device for acquiring text information entered by a user.

[1312] The "means for processing the question and selecting the appropriate domain-specific model" refers to an algorithm or device that analyzes the received question and selects the most appropriate domain-specific generative AI model based on the question's content.

[1313] A "means for generating an answer using a selected domain model" is a system or device that uses a selected generative AI model to create an appropriate answer to a user's question.

[1314] A "means for recognizing user emotions" is an algorithm or device that extracts and analyzes a user's emotional state from text entered by the user and other interactions.

[1315] "Means for adjusting the tone of the generated response based on the recognized emotion" refers to an algorithm or device that appropriately changes the expression or nuance of the generated response depending on the user's emotional state.

[1316] The "means for returning the generated answer to the user" is an interface or device for displaying or communicating the generated answer to the user.

[1317] This invention is a system that combines a system that provides generative AI models specialized in the fields of finance, law, patents, and IT for small and medium-sized enterprises with an emotion engine that recognizes user emotions.

[1318] System components and hardware / software

[1319] Hardware:

[1320] Server: A server with high-performance computing power

[1321] User devices: smartphones, tablets, PCs

[1322] software:

[1323] Natural language processing technology: Generative AI models such as GPT-4

[1324] Emotion recognition technology: EmotionAPI

[1325] System Operation Overview

[1326] A way to capture the question the user enters

[1327] Users can input questions through a chat interface on their smartphone or PC, such as, "Have you obtained a patent for your new product?"

[1328] A means of processing questions and selecting appropriate subject-matter models

[1329] The server uses natural language processing technology to analyze questions sent from the device to the server, and based on this analysis, selects an appropriate domain model, such as legal, financial, patent, or IT.

[1330] A means of generating answers using selected domain models

[1331] The selected generative AI model generates a detailed answer to the question, for example, for a patent-related question, it might generate an answer such as, "The patent for the new product has already been obtained and granted by the United States Patent and Trademark Office (USPTO)."

[1332] A means of recognizing user emotions

[1333] The emotion engine recognizes the user's emotions from the questions entered by the user and the context of the conversation. For example, if the user is feeling anxious, it can analyze that emotion.

[1334] A means to adjust the tone of generated responses based on perceived sentiment

[1335] The emotion engine adjusts the wording of the generated answer based on the user's emotions. For example, if the user is feeling anxious, the answer will be reassuringly phrased, such as "Don't worry, we've already obtained a patent for our new product."

[1336] A means of returning the generated answer to the user

[1337] The final adjusted answer is sent from the server to the device and displayed on the chat interface, allowing the user to view it and obtain the necessary information.

[1338] Specific example explanation

[1339] For example, if a user enters a question in a virtual store such as "Please tell me the patent status of this new product," the following processing will occur.

[1340] 1. User: Type a question into the chat interface: "What is the patent status of this new product?"

[1341] 2. Terminal: Receives the query and sends it to the server over the Internet.

[1342] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a question about a patent.

[1343] 4. Server: Selects a patent-specialized generative AI model and passes the query to it.

[1344] 5. Emotion engine: Analyzes the user's emotions when entering a question and determines whether the user is feeling anxious.

[1345] 6. Generative AI model: Generates answers such as "The new product has already been patented."

[1346] 7. Server: Based on the analysis of the emotion engine, adjust the response to a tone like "Don't worry, we've already patented our new product."

[1347] 8. Server: Formats the generated answer and sends it to the user's device.

[1348] 9. Terminal: The received response is displayed in the chat interface for the user to review.

[1349] Prompt Sentence Examples

[1350] Specific examples of prompts are as follows:

[1351] "Please tell me the patent status of this new product."

[1352] In this way, by combining a generative AI model with an emotion recognition engine, it is possible to provide optimal information according to the user's situation. This invention can support the efficient operation of small and medium-sized enterprises and improve user satisfaction.

[1353] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1354] Step 1:

[1355] User: Type a question into the chat interface, for example, "What is the patent status of this new product?"

[1356] Input: User question text

[1357] Output: The question text is sent to the terminal

[1358] Step 2:

[1359] Terminal: Receives the question text entered by the user.

[1360] Input: Question text

[1361] Output: The question text is converted into the appropriate format and sent to the server.

[1362] Step 3:

[1363] Server: Analyzes the received question using natural language processing techniques, for example, using text analysis to identify the subject of the question.

[1364] Input: Question text

[1365] Output: An appropriate domain model is selected based on the subject and intent of the question.

[1366] Step 4:

[1367] Server: Based on the results of the question analysis, selects an appropriate domain-specific model (e.g., a generative AI model specializing in patents).

[1368] Input: Parsed result of question

[1369] Output: Selected discipline model

[1370] Step 5:

[1371] Server: Uses the selected model to generate an answer to the question. The generated answer contains specialized information about the user's question.

[1372] Input: Question text and selected model

[1373] Output: Generated answer text

[1374] Step 6:

[1375] Server: Recognizes the user's emotions using an emotion engine. Extracts emotion data from the user's input text and identifies emotions such as anxiety, excitement, and stress.

[1376] Input: Question text

[1377] Output: User emotion data

[1378] Step 7:

[1379] Server: Adjust the tone of the generated answer based on the perceived emotion. For example, if the user is stressed, change the answer to a gentler tone.

[1380] Input: Generated answer text and sentiment data

[1381] Output: Tone-adjusted answer text

[1382] Step 8:

[1383] Server: Formats the finalized response and sends it to the user's device.

[1384] Input: Tone-adjusted answer text

[1385] Output: Formatted answer text to send to terminal

[1386] Step 9:

[1387] Terminal: Displays the received response in the chat interface, allowing the user to review the response.

[1388] Input: Received response text

[1389] Output: Reply to be displayed in the chat interface

[1390] 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.

[1391] 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.

[1392] 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.

[1393] [Fourth embodiment]

[1394] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1395] 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.

[1396] 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).

[1397] 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.

[1398] 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.

[1399] 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).

[1400] 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.

[1401] 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.

[1402] 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.

[1403] 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.

[1404] 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.

[1405] 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.

[1406] 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."

[1407] Understood. Below is the "Form for Carrying Out the Invention" from the patent specification.

[1408] ---

[1409] The present invention relates to a system that provides generative AI models specialized in four specialized fields, namely finance, legal affairs, patents, and IT, for small and medium-sized enterprises. Specific embodiments for implementing the present invention are described below.

[1410] System Overview

[1411] This system involves a series of processes: acquiring a question entered by a user, analyzing the question, selecting an appropriate domain model, generating an answer using the selected model, and returning it to the user. The system mainly involves three entities: a server, a terminal, and a user.

[1412] Program processing overview

[1413] 1. User inputs a question

[1414] User: Enters a question using the chat interface. For example, the user enters, "What are the legal procedures required to start a new business?"

[1415] 2. Submit your question

[1416] Terminal: The terminal formats the entered question appropriately and sends it over the Internet to a server.

[1417] 3. Question Analysis and Model Selection

[1418] Server: Receives the question and analyzes its intent using natural language processing technology. Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT) for the question.

[1419] 4. Answer Generation

[1420] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[1421] 5. Replying to Answers

[1422] Server: Formats the generated answer and sends it to the user's device.

[1423] Terminal: Displays the received response in the chat interface for the user to review.

[1424] Specific example explanation

[1425] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[1426] 1. User: Type a question into the chat interface: "What are the legal procedures required to start a new business?"

[1427] 2. Terminal: Receives the query and sends it to the server over the Internet.

[1428] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[1429] 4. Server: Selects a legal-focused generative AI model and passes the question to it.

[1430] 5. Generative AI model: Based on the question, it generates an answer such as, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[1431] 6. Server: Formats the generated answer and sends it to the user's device.

[1432] 7. Terminal: The received response is displayed in the chat interface, allowing the user to check the response and quickly understand the necessary steps.

[1433] In this way, this system provides an environment in which users can quickly acquire specialized knowledge at low cost, and supports the efficient operation of small and medium-sized enterprises.

[1434] The processing flow will be explained below.

[1435] Understood. Below I will explain the program process in concrete steps.

[1436] ---

[1437] Step 1:

[1438] User: Type a question into the chat interface. For example, "What are the legal steps required to start a new business?"

[1439] Step 2:

[1440] Terminal: Receives the questions entered by the user, formats them as text data, and sends the formatted data to a server over the Internet.

[1441] Step 3:

[1442] Server: Receives queries from the terminal via the receiving port. Decodes and extracts the received text data.

[1443] Step 4:

[1444] Server: Passes the decoded question data to a natural language processing (NLP) module, which analyzes the question and extracts key keywords and context.

[1445] Step 5:

[1446] NLP module: Analyzes the intent of the question, removes unnecessary information, and determines the appropriate area of ​​expertise. In this example, it determines that the question is related to "legal affairs."

[1447] Step 6:

[1448] Server: Based on the analysis results of the NLP module, select a legal-specialized generative AI model and pass the question data to the selected model.

[1449] Step 7:

[1450] Generative AI model: Generates appropriate answers based on question data. For example, it generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[1451] Step 8:

[1452] Server: Receives the generated answer, formats it in a format that is easy for the user to read, and sends the formatted answer data to the terminal.

[1453] Step 9:

[1454] Terminal: The received response data is decoded and displayed in the chat interface. The user can check the displayed response and obtain the necessary information.

[1455] ---

[1456] Through these steps, the system provides appropriate, expert answers to questions entered by users. The specific operations performed at each processing step allow users to quickly acquire expert knowledge and support the efficient operation of small and medium-sized businesses.

[1457] Example 1

[1458] 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."

[1459] Small and medium-sized enterprises (SMEs) face difficulties in obtaining prompt and appropriate answers to financial, legal, patent, and IT-related issues that require specialized knowledge. Obtaining expert advice often requires time and money, posing a significant challenge, especially for SMEs with limited resources. Furthermore, there is a lack of systems that can properly analyze questions and provide information in the most appropriate specialized field. Therefore, there is a need for a system that can provide specialized knowledge efficiently and economically.

[1460] 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.

[1461] In this invention, the server includes means for acquiring a question input by a user, means for formatting the question and sending it to the server, means for analyzing the question and selecting an appropriate domain model in the server, means for sending a prompt to the selected domain model and receiving a generated answer, and means for formatting the generated answer and sending it to the user's terminal, thereby enabling users to quickly acquire specialized knowledge at low cost.

[1462] Ok, below are some definitions of important words:

[1463] A "user" is an entity that utilizes the system to input and confirm questions.

[1464] A "question" is information that a user enters into the system using the chat interface.

[1465] A "terminal" is a hardware device that a user uses to enter questions and receive answers from a server.

[1466] A "server" is a computer system that processes questions submitted by users, selects appropriate domain expertise models, and returns generated answers to the users.

[1467] A "specialized domain model" is a generative AI model that specializes in a specific field, such as finance, law, patents, or IT.

[1468] A "generative AI model" is an artificial intelligence model that generates answers in natural language based on input prompts.

[1469] A "prompt" is an input sentence used to generate an answer for a generative AI model.

[1470] "Formatting" refers to shaping questions entered by users and answers generated by generative AI models into a form suitable for communication or display.

[1471] "Natural language processing technology" is a technology that enables computers to understand, interpret, and generate human language.

[1472] A "chat interface" is an interactive user interface that allows a user to enter questions and receive answers from a generative AI model.

[1473] An "answer" is information generated by a generative AI model based on a prompt.

[1474] This invention relates to a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, legal affairs, patents, and IT. This system efficiently acquires, analyzes, and answers user questions, and involves three entities: the user, the terminal, and the server.

[1475] System Overview

[1476] The system executes a series of processes: a user inputs a question using a chat interface, the question is received by a server, an answer is generated based on an appropriate generative AI model, and the answer is returned to the user. Details for specifically implementing the system are described below.

[1477] Hardware and software used

[1478] Hardware:

[1479] Device: A device used by a user, such as a PC, smartphone, or tablet.

[1480] Server: A server (e.g., a cloud service server) for processing questions and managing generative AI models.

[1481] software:

[1482] Chat interface: An interface for users to enter questions and view generated answers.

[1483] Natural language processing technology: Software to analyze the intent of the question (e.g., Google Cloud NLP API, IBM Watson).

[1484] Generative AI models: Artificial intelligence models for generating answers based on domain expertise (e.g., OpenAI GPT-3).

[1485] Example of a system

[1486] User Action:

[1487] Users access the chat interface from a PC or smartphone browser and enter questions such as, "Please tell me what legal procedures are required to start a new business."

[1488] Terminal behavior:

[1489] The terminal receives the user's question, formats it appropriately, and sends it to the server over the Internet using an HTTP request (for example, the POST method).

[1490] Server behavior:

[1491] The server analyzes the received question using natural language processing technology (e.g., Google Cloud NLP API or IBM Watson) and determines that the question is legal. It then selects a generative AI model (e.g., OpenAI GPT-3) specialized in the legal domain, generates a prompt (e.g., "What legal procedures are required to start a new business?"), and sends it to the AI ​​model.

[1492] Generative AI model answers:

[1493] Based on the prompt, the generative AI model generates an answer such as, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office," and sends this answer back to the server.

[1494] Sending the response from the server to the device:

[1495] The server then formats the answers received from the generative AI model and sends them to the user's device, which then displays the answers in the chat interface for the user to review.

[1496] Examples of prompt statements

[1497] Prompt: "What are the legal steps required to start a new business?"

[1498] The present invention allows small and medium-sized enterprises to acquire specialized knowledge quickly and at low cost, enabling efficient business operations. This system can handle everything from processing inquiries to providing specialized information.

[1499] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1500] Understood. Below is a detailed explanation of the system program processing flow.

[1501] Step 1: User enters question

[1502] User: Accesses the chat interface in a browser on a PC or smartphone and types in a question.

[1503] Input: A user types a question in text format, such as "What are the legal steps required to start a new business?"

[1504] Output: The question text is sent to the chat interface.

[1505] What it does: The user types a question into the chat interface using a keyboard or touchscreen.

[1506] Step 2: Format and submit your question

[1507] Terminal: Converts the user's question into an appropriate format (e.g., JSON format) and sends it to the server.

[1508] Input: The question text entered by the user.

[1509] Data processing: Format the question text into JSON format.

[1510] Output: The formatted question data is sent to the server as an HTTP request (e.g., a POST request).

[1511] What happens: The device's browser runs JavaScript code, formats the question data appropriately, and sends it to the server.

[1512] Step 3: Receiving and parsing the question

[1513] Server: Analyzes the received question using natural language processing technology and understands its intent.

[1514] Input: Formatted question data sent from the terminal.

[1515] Data Computing: Analyze the intent of the question using natural language processing technology (e.g., Google Cloud NLP API, IBM Watson).

[1516] Output: The analysis result will be the domain of expertise the question is related to. For example, it will be determined that the question is about "legal affairs."

[1517] Specific operation: The server calls the natural language processing API, analyzes the question data, and identifies the area of ​​expertise.

[1518] Step 4: Model selection and prompt generation

[1519] Server: Selects the appropriate domain model based on the analysis results and generates prompts.

[1520] Input: Analysis results (e.g., a question about "legal").

[1521] Data processing: Select a specialized domain model (e.g., a generative AI model specialized in legal matters) and generate a prompt. Create a prompt such as, "Please tell me the legal procedures required to start a new business."

[1522] Output: The selected generative AI model and the prompt.

[1523] Specific operation: The server automatically generates a prompt sentence and prepares to send it to the selected generative AI model.

[1524] Step 5: Submitting an Answer Generation Request

[1525] Server: Sends prompts to the generative AI model and requests an answer.

[1526] Input: The generated prompt statement.

[1527] Data computation: Send prompts to a generative AI model (e.g., OpenAI GPT-3) to generate an answer.

[1528] Output: The generated answer.

[1529] How it works: The server uses HTTP requests to send prompts to the generative AI model and receive response data.

[1530] Step 6: Receiving and formatting responses

[1531] Server: Formats the answers received from the generative AI model.

[1532] Input: The answer data received from the generative AI model.

[1533] Data processing: Format the received response data into an easy-to-read format (e.g., HTML format).

[1534] Output: Formatted response data.

[1535] What happens: The server organizes the response data and converts it into a format suitable for display.

[1536] Step 7: Submit and view your responses

[1537] Server: Sends the formatted response data to the user's device.

[1538] Input: Formatted response data.

[1539] Output: The answer data sent to the user's device.

[1540] Specific operation: The server sends the answer data as an HTTP response.

[1541] Terminal: Displays the response received from the server in the chat interface.

[1542] Input: The response data received from the server.

[1543] Output: The response displayed in the chat interface.

[1544] What happens: The device's browser runs the JavaScript code and displays the response in the chat interface, where the user can view the response on their screen.

[1545] The above is the specific processing flow of the system program.

[1546] (Application example 1)

[1547] 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."

[1548] In today's small and medium-sized enterprises, it is difficult to secure personnel with specialized knowledge in legal affairs, finance, patents, and IT security. As a result, strengthening corporate security and legal responses are often delayed. Furthermore, there are limitations to independently obtaining information in these specialized fields, making it difficult to take prompt and accurate measures. The present invention aims to solve these problems by providing specialized knowledge in real time and supporting the efficient and safe operation of small and medium-sized enterprises.

[1549] 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.

[1550] In this invention, the server includes means for acquiring a question input by a user, means for processing the question and selecting an appropriate specialty domain model, means for generating an answer using the selected specialty domain model, means for returning the generated answer to the user, means for inputting a question via a smart device, and means for providing advice on security measures in real time. This enables a user to easily input specialized questions via a smart device and quickly receive necessary security measures and legal advice in real time.

[1551] The "means for acquiring a question entered by a user" is an interface for receiving a question entered by a user through a smart device and transmitting the content of the question to the system.

[1552] The "means for processing the query and selecting an appropriate domain model" is a processing device for analyzing the received query and selecting the most appropriate domain model based on the content of the query.

[1553] The "means for generating an answer using a selected domain model" is an algorithm or software that uses a selected domain model to generate an answer to a user's question.

[1554] The "means for returning the generated answer to the user" refers to a communication means and interface for transmitting the generated answer to the user's smart device and displaying it.

[1555] "Means for inputting questions via a smart device" refers to a function that allows users to input questions by voice or text using a device such as a smartphone or smart glasses.

[1556] "Means for providing security advice in real time" refers to a system-wide function that provides appropriate measures and advice immediately in response to security questions.

[1557] This invention relates to a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, legal affairs, patents, and IT security. The entire system mainly involves three entities: a server, a terminal, and a user, and operates according to the following specific processing steps.

[1558] System Overview

[1559] Program processing overview

[1560] 1. User enters question:

[1561] User: Using a smart device (such as a smartphone or smart glasses), the user types a question, for example, "How can I improve the security of my company's network?"

[1562] 2. Submit your question:

[1563] Terminal: Receives the query and sends it to the server via the Internet.

[1564] 3. Question analysis and model selection:

[1565] Server: Receives the question and analyzes its intent using natural language processing technology (e.g., BERT, GPT-4). Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT security) for the question.

[1566] 4. Generate answers:

[1567] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about security, a generative AI model specializing in IT security will generate an answer such as, "To strengthen network security, it is important to first review your firewall settings and configure access control lists appropriately."

[1568] 5. Answer Response:

[1569] Server: Formats the generated answer and sends it to the user's device.

[1570] Terminal: The received answers are displayed on the interface of the smart device for the user to review.

[1571] Hardware and Software Configuration

[1572] Natural language processing engines: BERT, GPT-4

[1573] Communication API: HTTPS

[1574] Devices: Smartphones (iOS / Android), smart glasses

[1575] Specific processing explanation

[1576] 1. Data acceptance and transmission:

[1577] The device receives the question, formats it appropriately, and sends it to the server via HTTPS.

[1578] 2. Question Analysis:

[1579] The server analyzes the received question using natural language processing engines such as BERT and GPT-4, which identifies the intent of the question and selects the most appropriate domain expertise model.

[1580] 3. Answer generation:

[1581] The domain expertise model generates the best answer based on the question, for example, a security question will be answered with firewall and access control list settings.

[1582] 4. Submitting and Viewing Answers:

[1583] The server formats the generated response in JSON format and sends it to the device, which receives it and displays it on the device's interface. Furthermore, by using a voice output function, the system can be used in environments with limited visual capabilities.

[1584] Examples of concrete examples and prompts

[1585] Specific examples

[1586] A user asks a question on their smartphone: "How can I strengthen our company's network security?" The device receives the question and sends it to the server. The server analyzes the question and selects an AI model specialized in IT security. The AI ​​model generates an answer: "To strengthen network security, it is important to first review your firewall settings and configure the access control list appropriately." The generated answer is sent to the device, where the user confirms it.

[1587] Prompt Sentence Examples

[1588] "What settings or techniques would you recommend to enhance our network security?"

[1589] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1590] Step 1:

[1591] The user uses a smartphone or smart glasses to ask a question by voice or by entering text, such as "How can I improve the security of my company's network?" The entered data is converted into text on the fly and passed to the device's application.

[1592] Step 2:

[1593] The device then formats the received question appropriately and sends it to the server using the HTTPS protocol, where the question is transferred to the server in JSON format.

[1594] Step 3:

[1595] The server analyzes the received question data. First, it uses a natural language processing engine (BERT or GPT-4) to analyze the intent of the question. Specifically, it tokenizes the question text, understands the context, and identifies which domain the question relates to. For example, it extracts keywords such as "network security," "cloud," and "access control."

[1596] Step 4:

[1597] Based on the analysis results, the server selects the most appropriate specialized domain model. The server selects an appropriate model from multiple generative AI models for finance, legal affairs, patents, IT security, etc. In this case, it selects a generative AI model specialized in IT security.

[1598] Step 5:

[1599] The selected generative AI model generates an answer based on the question. The model references past data and knowledge bases to generate the best answer to the question. For example, it might generate an answer like, "To strengthen network security, it is important to first review your firewall settings and configure access control lists appropriately."

[1600] Step 6:

[1601] The server then formats the generated answer appropriately and sends it to the device in JSON format, which may include the answer as well as related references and links.

[1602] Step 7:

[1603] The device displays the received answers on the user interface. The display method can be selected as text display or voice output. The user can check the answer content on the screen and obtain the necessary information. For example, by using voice output, the content can be understood even when there are visual constraints.

[1604] Step 8:

[1605] The user can then take specific action based on the displayed answers. If necessary, they can enter further questions to obtain new information. For example, they can ask again, "Please tell me how to specifically configure the access control list."

[1606] Through this series of processing steps, small and medium-sized business users can receive expert advice in real time, enabling them to take security measures quickly and accurately.

[1607] 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.

[1608] Understood. Below is a description of the "Mode for Carrying Out the Invention" based on the claims.

[1609] ---

[1610] The present invention combines a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, law, patents, and IT, with an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described below.

[1611] System Overview

[1612] This system involves a series of processes: it acquires a question entered by a user, analyzes the question, selects an appropriate domain model, generates an answer using the selected model, and returns it to the user. Furthermore, by combining it with an emotion engine that recognizes the user's emotions, it provides a more personalized response.

[1613] The system mainly involves three entities: a server, a terminal, and a user.

[1614] Program processing overview

[1615] 1. User inputs a question

[1616] User: Enters a question using the chat interface, for example, "What are the legal steps required to start a new business?"

[1617] 2. Submit your question

[1618] Terminal: The terminal formats the entered question appropriately and sends it over the Internet to a server.

[1619] 3. Question Analysis and Model Selection

[1620] Server: Receives the question and analyzes its intent using natural language processing technology. Based on the results, it selects the most appropriate domain model (finance, legal, patent, or IT) for the question.

[1621] 4. User Emotion Recognition

[1622] Emotion engine: Analyzes the user's emotions from the text they input and other interactions, for example, to determine whether they are stressed or excited.

[1623] 5. Answer Generation

[1624] Generative AI model: A selected domain-specific model generates an answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[1625] 6. Adjust the tone of your response

[1626] Server: Adjust the tone and wording of the generated answer based on the user's emotions as recognized by the emotion engine. For example, if the user is feeling stressed, the answer will be more friendly and soothing.

[1627] 7. Replying to Answers

[1628] Server: Formats the generated answer and sends it to the user's device.

[1629] Terminal: Displays the received response in the chat interface for the user to review.

[1630] Specific example explanation

[1631] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[1632] 1. User: Type a question into the chat interface: "What are the legal procedures required to start a new business?"

[1633] 2. Terminal: Receives the query and sends it to the server over the Internet.

[1634] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[1635] 4. Server: Selects a legal-focused generative AI model and passes the question to it.

[1636] 5. Emotion engine: Analyzes the user's emotions when entering a question and determines whether the user is feeling stressed.

[1637] 6. Generative AI model: Generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[1638] 7. Server: Based on the analysis results of the emotion engine, adjust the response to a friendly tone, such as, "I know you're having a lot of trouble, but don't worry. To start a new business, you first need to create and submit a certificate of registered matters. Next, you'll need to file a notification with the tax office."

[1639] 8. Server: Formats the generated answer and sends it to the user's device.

[1640] 9. Terminal: The received response is displayed in the chat interface, allowing the user to check the response and obtain the necessary information.

[1641] In this way, the system recognizes users' emotions and provides responses tailored to each individual user's situation, enabling more personalized support, further supporting the efficient operation of small and medium-sized businesses.

[1642] The processing flow will be explained below.

[1643] Understood. Below I will explain the process in concrete steps.

[1644] ---

[1645] Step 1:

[1646] User: Type a question into the chat interface, for example, "What are the legal steps I need to take to start a new business?"

[1647] Step 2:

[1648] Terminal: Receives the question, formats it as text data, and sends the formatted data to a server over the Internet.

[1649] Step 3:

[1650] Server: Receives questions from the terminal via the receiving port. Decodes the received text data and extracts the question content.

[1651] Step 4:

[1652] Server: Passes the decoded question data to a natural language processing (NLP) module, which analyzes the question and extracts key keywords and context.

[1653] Step 5:

[1654] NLP module: Analyzes the intent of the question, removes unnecessary information, and determines the appropriate area of ​​expertise. In this example, it determines that the question is related to "legal affairs."

[1655] Step 6:

[1656] Server: Based on the analysis results of the NLP module, select a legal-specialized generative AI model and pass the question data to the selected model.

[1657] Step 7:

[1658] Emotion engine: Analyzes the user's emotions from the text they input and other interactions, for example, to determine whether they are stressed or excited.

[1659] Step 8:

[1660] Generative AI model: Generates appropriate answers based on question data. For example, it generates the answer, "When starting a new business, you must first create and submit a certificate of registered matters. Next, you will be required to notify the tax office."

[1661] Step 9:

[1662] Server: Adjust the tone and wording of the generated answer based on the user's emotions as recognized by the emotion engine. For example, if the user is feeling stressed, the answer will be more friendly and soothing.

[1663] Step 10:

[1664] Server: Formats the generated answer and sends it to the user's device.

[1665] Step 11:

[1666] Terminal: The received response data is decoded and displayed in the chat interface. The user can check the displayed response and obtain the necessary information.

[1667] ---

[1668] Through these steps, the system can provide appropriate, expert answers to questions entered by users, and can also provide personalized responses that take the user's feelings into consideration. The specific operations performed at each processing step allow users to quickly acquire expert knowledge, thereby supporting the efficient operation of small and medium-sized businesses.

[1669] Example 2

[1670] 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."

[1671] Conventional technologies select domain expertise models to provide appropriate answers to questions entered by users, but they lack personalized responses based on the user's emotions and do not fully consider the stressful situations faced by small and medium-sized business owners, particularly those faced with complex problems. This reduces user satisfaction when obtaining information and hinders efficient operations.

[1672] 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.

[1673] In this invention, the server includes means for acquiring a question input by a user, means for processing the question and selecting an appropriate domain model, means for generating an answer using the selected domain model, means for adjusting the tone of the generated answer, and means for returning the generated answer to the user, thereby enabling a personalized answer that takes into account the user's emotions.

[1674] The "means for acquiring a question entered by a user" refers to a device or software for receiving a question entered by a user using a chat interface or the like and capturing it as data.

[1675] The "means for processing questions and selecting an appropriate specialized field model" refers to a device or software that analyzes the received question using natural language processing technology and selects a model corresponding to a specialized field such as finance, law, patents, or IT based on the content of the question.

[1676] A "means for generating an answer using a selected domain-specific model" is a device or software that executes a domain-specific model to generate an appropriate answer using a pre-trained dataset based on the analysis of the question.

[1677] The "means for adjusting the tone of the generated response" refers to a device or software that modifies the expression and style of the generated response content to match the user's emotional state, and presents it in a form that is familiar and easy for the user to understand.

[1678] The "means for returning the generated response to the user" is a device or software for transmitting the tone-adjusted response to the user's terminal and ultimately displaying it on the user's chat interface.

[1679] "Natural language processing technology" is a technology that enables computers to understand, interpret, and generate human language, and specifically, it is a technology that includes algorithms and models that analyze and generate text.

[1680] A "chat interface" is a user interface for users to input text, and is a platform for entering questions and displaying answers.

[1681] An "emotion engine" is a software module that analyzes the emotions a user feels from the text entered by the user and other interaction data.

[1682] A "data packet" is a small unit of data used when transmitting data over a network such as the Internet.

[1683] The present invention is a system that provides generative AI models specialized in four specialized fields for small and medium-sized enterprises: finance, law, patents, and IT. It also combines an emotion engine that recognizes user emotions. Specific embodiments for implementing the present invention are described below.

[1684] Overall system configuration

[1685] The system involves a series of processes: it acquires a question entered by a user, analyzes the question, selects an appropriate domain model, generates an answer using the selected model, and returns it to the user. Furthermore, it combines an emotion engine that recognizes the user's emotions to provide a more personalized response.

[1686] The system mainly uses the following hardware and software:

[1687] Server: Responsible for question analysis, domain model selection, answer generation, and tone adjustment.

[1688] Terminal: Enter questions and display answers.

[1689] Emotion engine: A software module that analyzes the user's emotions.

[1690] Natural language processing technology: Techniques for analyzing the intent of questions (e.g., BERT, GPT-3).

[1691] Processing flow, data processing and data calculation

[1692] 1. User inputs a question

[1693] The user uses the chat interface to input a question, for example, "Please tell me the legal procedures required to start a new business." This question is captured as data on the device.

[1694] 2. Submit your question

[1695] The terminal converts the question into an appropriate format (e.g., JSON format) and sends it to a server over the Internet.

[1696] 3. Question Analysis and Domain Model Selection

[1697] The server analyzes the received question data, using natural language processing technology (such as BERT or GPT-3) to analyze the content and intent of the question, and based on the results, selects the domain model (finance, legal, patent, or IT) that best suits the question.

[1698] 4. Recognition of user emotions using an emotion engine

[1699] The emotion engine installed on the server recognizes the user's emotions from the text they input and other interaction data, analyzing their emotional state such as whether they are stressed, nervous, excited, etc.

[1700] 5. Answer Generation

[1701] The selected domain-specific model generates an appropriate answer to the question. For example, if the question is about legal matters, a legal-specific generative AI model will generate the answer. The generation process uses a pre-trained dataset to provide a specific answer based on the user's question.

[1702] 6. Adjust the tone of your response

[1703] Instead of generating responses as is, the emotion engine adjusts the tone and expression based on the user's emotional state. For example, if the user is feeling stressed, the response will be modified to be more friendly and soothing.

[1704] 7. Replying to Answers

[1705] The server converts the tone-adjusted response into an appropriate format and sends it to the user's terminal, which displays the received response in a chat interface for the user to review.

[1706] Specific example explanation

[1707] For example, if the owner of a small or medium-sized business about to start a new business wants to use this system to learn about legal procedures, the system will operate in the following manner.

[1708] User: Type a question into the chat interface: "What are the legal steps required to start a new business?"

[1709] Terminal: Receives the query and sends it to the server over the Internet.

[1710] Server: Analyzes the received question using natural language processing technology and determines that it is a legal question.

[1711] Server: Selects a legal-focused generative AI model and passes the question to it.

[1712] Emotion engine: Analyzes the user's emotions and determines whether the user is feeling stressed.

[1713] Generative AI model: Generates the answer, "To start a new business, you must first create and submit a certificate of registered matters. Next, you will be required to file a notification with the tax office."

[1714] Server: Based on the analysis results of the emotion engine, adjust the response to a more friendly tone: "We understand that you are facing many challenges, but please rest assured. To start a new business, you must first create and submit a certificate of registered matters. Next, you will need to file a notification with the tax office."

[1715] Server: Formats the generated answer and sends it to the user's device.

[1716] Terminal: The received response is displayed in the chat interface, allowing the user to check the response and obtain the necessary information.

[1717] In this way, the system recognizes users' emotions and provides responses tailored to each individual user's situation, enabling more personalized support, which can help small and medium-sized businesses operate more efficiently.

[1718] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1719] Step 1:

[1720] User question input

[1721] User: Uses the chat interface to type a question, for example, "What are the legal steps required to start a new business?"

[1722] Input: Text data entered by the user.

[1723] Output: The retrieved question text.

[1724] Specific operation: When the user enters text in the chat box and presses the "Send" button, the text data is sent to the terminal.

[1725] Step 2:

[1726] Submit a Question

[1727] Terminal: Converts the question into an appropriate format, such as JSON, and sends it to a server over the Internet.

[1728] Input: The retrieved question text.

[1729] Output: Formatted question data.

[1730] Specific operation: Converts text data into JSON format and sends it to the server using an HTTP request.

[1731] Step 3:

[1732] Question analysis and domain model selection

[1733] Server: Analyzes the received question data. It uses natural language processing technology (e.g., BERT, GPT-3, etc.) to analyze the content and intent of the question, and based on the results, selects the most appropriate domain model (finance, legal, patent, or IT).

[1734] Input: Formatted question data.

[1735] Output: Information about the question intent and the selected domain model.

[1736] What it does: The natural language processing model analyzes the question text and extracts its intent. Based on the results, it selects the appropriate domain model.

[1737] Step 4:

[1738] Recognizing user emotions with an emotion engine

[1739] Server: The emotion engine analyzes the user's input text and determines the user's emotional state (stress, tension, excitement, etc.).

[1740] Input: The text entered by the user.

[1741] Output: User emotion data.

[1742] What it does: Sentiment analysis algorithms analyze text data and assign emotional tags (e.g., stress, relief, etc.).

[1743] Step 5:

[1744] Generate answers

[1745] Server: The selected domain model generates an appropriate answer to the question. For example, if the question is about legal matters, a legal-specialized generative AI model will generate the answer, "To start a new business, you must first create and submit a certificate of registered matters. Next, you will need to file a notification with the tax office."

[1746] Input: Question intent and selected domain model information.

[1747] Output: The generated answer.

[1748] What it does: The domain-specific model generates answer text based on a pre-trained dataset.

[1749] Step 6:

[1750] Adjusting the tone of your response

[1751] Server: Adjusts the tone and expression of the generated answer based on the user's emotional state as recognized by the emotion engine. If the user is feeling stressed, the answer will be modified to be more friendly and soothing.

[1752] Input: Generated answers and user sentiment data.

[1753] Output: Tone-adjusted answer.

[1754] What it does: Text processing algorithms analyze the generated answers and modify them to suit the user's sentiment and style.

[1755] Step 7:

[1756] Response to the answer

[1757] Server: Sends the formatted answer to the user's device.

[1758] Terminal: Displays the received response in the chat interface for the user to review.

[1759] Input: Tone-adjusted answer.

[1760] Output: The response displayed in the chat interface.

[1761] Specific operation: The server sends the answer data to the device as an HTTP response, and the device displays the received data in the chat box.

[1762] (Application example 2)

[1763] 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."

[1764] With conventional systems, it is difficult for small and medium-sized enterprises (SMEs) to quickly obtain appropriate information when seeking specialized knowledge in finance, law, patents, or IT. Furthermore, because the system responds uniformly without considering the user's feelings, it is difficult to provide appropriate support tailored to the user's situation. As a result, user satisfaction declines and effective communication becomes difficult.

[1765] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring a question entered by a user, means for processing the question and selecting an appropriate expertise domain model, means for generating an answer using the selected expertise domain model, means for recognizing the user's emotion, means for adjusting the tone of the generated answer based on the recognized emotion, and means for returning the generated answer to the user. This makes it possible to quickly provide the expert knowledge desired by the user and to provide a personalized response according to the user's emotion.

[1766] The "means for acquiring a user-input question" is an interface or device for acquiring text information entered by a user.

[1767] The "means for processing the question and selecting the appropriate domain-specific model" refers to an algorithm or device that analyzes the received question and selects the most appropriate domain-specific generative AI model based on the question's content.

[1768] A "means for generating an answer using a selected domain model" is a system or device that uses a selected generative AI model to create an appropriate answer to a user's question.

[1769] A "means for recognizing user emotions" is an algorithm or device that extracts and analyzes a user's emotional state from text entered by the user and other interactions.

[1770] "Means for adjusting the tone of the generated response based on the recognized emotion" refers to an algorithm or device that appropriately changes the expression or nuance of the generated response depending on the user's emotional state.

[1771] The "means for returning the generated answer to the user" is an interface or device for displaying or communicating the generated answer to the user.

[1772] This invention is a system that combines a system that provides generative AI models specialized in the fields of finance, law, patents, and IT for small and medium-sized enterprises with an emotion engine that recognizes user emotions.

[1773] System components and hardware / software

[1774] Hardware:

[1775] Server: A server with high-performance computing power

[1776] User devices: smartphones, tablets, PCs

[1777] software:

[1778] Natural language processing technology: Generative AI models such as GPT-4

[1779] Emotion recognition technology: EmotionAPI

[1780] System Operation Overview

[1781] A way to capture the question the user enters

[1782] Users can input questions through a chat interface on their smartphone or PC, such as, "Have you obtained a patent for your new product?"

[1783] A means of processing questions and selecting appropriate subject-matter models

[1784] The server uses natural language processing technology to analyze questions sent from the device to the server, and based on this analysis, selects an appropriate domain model, such as legal, financial, patent, or IT.

[1785] A means of generating answers using selected domain models

[1786] The selected generative AI model generates a detailed answer to the question, for example, for a patent-related question, it might generate an answer such as, "The patent for the new product has already been obtained and granted by the United States Patent and Trademark Office (USPTO)."

[1787] A means of recognizing user emotions

[1788] The emotion engine recognizes the user's emotions from the questions entered by the user and the context of the conversation. For example, if the user is feeling anxious, it can analyze that emotion.

[1789] A means to adjust the tone of generated responses based on perceived sentiment

[1790] The emotion engine adjusts the wording of the generated answer based on the user's emotions. For example, if the user is feeling anxious, the answer will be reassuringly phrased, such as "Don't worry, we've already obtained a patent for our new product."

[1791] A means of returning the generated answer to the user

[1792] The final adjusted answer is sent from the server to the device and displayed on the chat interface, allowing the user to view it and obtain the necessary information.

[1793] Specific example explanation

[1794] For example, if a user enters a question in a virtual store such as "Please tell me the patent status of this new product," the following processing will occur.

[1795] 1. User: Type a question into the chat interface: "What is the patent status of this new product?"

[1796] 2. Terminal: Receives the query and sends it to the server over the Internet.

[1797] 3. Server: Analyzes the received question using natural language processing technology and determines that it is a question about a patent.

[1798] 4. Server: Selects a patent-specialized generative AI model and passes the query to it.

[1799] 5. Emotion engine: Analyzes the user's emotions when entering a question and determines whether the user is feeling anxious.

[1800] 6. Generative AI model: Generates answers such as "The new product has already been patented."

[1801] 7. Server: Based on the analysis of the emotion engine, adjust the response to a tone like "Don't worry, we've already patented our new product."

[1802] 8. Server: Formats the generated answer and sends it to the user's device.

[1803] 9. Terminal: The received response is displayed in the chat interface for the user to review.

[1804] Prompt Sentence Examples

[1805] Specific examples of prompts are as follows:

[1806] "Please tell me the patent status of this new product."

[1807] In this way, by combining a generative AI model with an emotion recognition engine, it is possible to provide optimal information according to the user's situation. This invention can support the efficient operation of small and medium-sized enterprises and improve user satisfaction.

[1808] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1809] Step 1:

[1810] User: Type a question into the chat interface, for example, "What is the patent status of this new product?"

[1811] Input: User question text

[1812] Output: The question text is sent to the terminal

[1813] Step 2:

[1814] Terminal: Receives the question text entered by the user.

[1815] Input: Question text

[1816] Output: The question text is converted into the appropriate format and sent to the server.

[1817] Step 3:

[1818] Server: Analyzes the received question using natural language processing techniques, for example, using text analysis to identify the subject of the question.

[1819] Input: Question text

[1820] Output: An appropriate domain model is selected based on the subject and intent of the question.

[1821] Step 4:

[1822] Server: Based on the results of the question analysis, selects an appropriate domain-specific model (e.g., a generative AI model specializing in patents).

[1823] Input: Parsed result of question

[1824] Output: Selected discipline model

[1825] Step 5:

[1826] Server: Uses the selected model to generate an answer to the question. The generated answer contains specialized information about the user's question.

[1827] Input: Question text and selected model

[1828] Output: Generated answer text

[1829] Step 6:

[1830] Server: Recognizes the user's emotions using an emotion engine. Extracts emotion data from the user's input text and identifies emotions such as anxiety, excitement, and stress.

[1831] Input: Question text

[1832] Output: User emotion data

[1833] Step 7:

[1834] Server: Adjust the tone of the generated answer based on the perceived emotion. For example, if the user is stressed, change the answer to a gentler tone.

[1835] Input: Generated answer text and sentiment data

[1836] Output: Tone-adjusted answer text

[1837] Step 8:

[1838] Server: Formats the finalized response and sends it to the user's device.

[1839] Input: Tone-adjusted answer text

[1840] Output: Formatted answer text to send to terminal

[1841] Step 9:

[1842] Terminal: Displays the received response in the chat interface, allowing the user to review the response.

[1843] Input: Received response text

[1844] Output: Reply to be displayed in the chat interface

[1845] 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.

[1846] 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.

[1847] 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.

[1848] 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.

[1849] 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.

[1850] 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.

[1851] 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).

[1852] 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.

[1853] 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."

[1854] 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.

[1855] 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).

[1856] 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.

[1857] 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.

[1858] 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.

[1859] 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.

[1860] 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.

[1861] 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.

[1862] 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.

[1863] 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.

[1864] 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.

[1865] 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.

[1866] The following is further disclosed regarding the above embodiment.

[1867] Understood. Below are the draft claims.

[1868] (Claim 1)

[1869] means for obtaining a question entered by a user;

[1870] a means of processing the question and selecting the appropriate disciplinary model;

[1871] means for generating an answer using the selected domain model;

[1872] means for returning the generated answer to the user;

[1873] A system including:

[1874] (Claim 2)

[1875] 10. The system of claim 1, further comprising means for analyzing intent of the question using natural language processing techniques.

[1876] (Claim 3)

[1877] 10. The system of claim 1, further comprising means for inputting and displaying user questions in a chat interface.

[1878] "Example 1"

[1879] Understood. Below, I will add the features of the system and write the claims.

[1880] (Claim 1)

[1881] means for obtaining a question entered by a user;

[1882] a means of processing the question and selecting the appropriate disciplinary model;

[1883] means for generating an answer using the selected domain model;

[1884] means for returning the generated answer to the user;

[1885] means for formatting and transmitting the retrieved questions to a server;

[1886] means for parsing the query and selecting a domain expertise model at the server;

[1887] means for sending prompts to the selected model and receiving generated answers;

[1888] means for generating answers using questions and prompts, and a generative AI model;

[1889] means for formatting and transmitting the generated answer to the user's terminal;

[1890] A system including:

[1891] (Claim 2)

[1892] 10. The system of claim 1, further comprising means for analyzing intent of the question using natural language processing techniques.

[1893] (Claim 3)

[1894] 10. The system of claim 1, further comprising means for inputting and displaying user questions in a chat interface.

[1895] "Application Example 1"

[1896] (Claim 1)

[1897] means for obtaining a question entered by a user;

[1898] a means of processing the question and selecting the appropriate disciplinary model;

[1899] means for generating an answer using the selected domain model;

[1900] means for returning the generated answer to the user;

[1901] A means for inputting questions via a smart device;

[1902] A means of providing real-time security advice;

[1903] A system including:

[1904] (Claim 2)

[1905] 10. The system of claim 1, further comprising means for analyzing intent of the question using natural language processing techniques.

[1906] (Claim 3)

[1907] 10. The system of claim 1, further comprising means for inputting and displaying user questions in a chat interface.

[1908] "Example 2: Combining Emotion Engines"

[1909] (Claim 1)

[1910] means for obtaining a question entered by a user;

[1911] a means of processing the question and selecting the appropriate disciplinary model;

[1912] means for generating an answer using the selected domain model;

[1913] a means to adjust the tone of the responses generated;

[1914] means for returning the generated answer to the user;

[1915] A system including:

[1916] (Claim 2)

[1917] 10. The system of claim 1, further comprising means for analyzing intent of the question using natural language processing techniques.

[1918] (Claim 3)

[1919] 10. The system of claim 1, further comprising means for inputting and displaying user questions in a chat interface.

[1920] "Application example 2 when combining emotion engines"

[1921] Understood. Below are the original patent claims rewritten to include the characteristic parts of the application example.

[1922] (Claim 1)

[1923] means for obtaining a question entered by a user;

[1924] a means of processing the question and selecting the appropriate disciplinary model;

[1925] means for generating an answer using the selected domain model;

[1926] means for recognizing a user's emotion;

[1927] a means for adjusting the tone of the generated response based on the perceived sentiment;

[1928] means for returning the generated answer to the user;

[1929] A system including:

[1930] (Claim 2)

[1931] 10. The system of claim 1, further comprising means for analyzing intent of the question using natural language processing techniques.

[1932] (Claim 3)

[1933] 10. The system of claim 1, further comprising means for inputting and displaying user questions in a chat interface. [Explanation of symbols]

[1934] 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. means for obtaining a question entered by a user; a means of processing the question and selecting the appropriate disciplinary model; means for generating an answer using the selected domain model; means for returning the generated answer to the user; A system including:

2. The system of claim 1 , further comprising means for analyzing the intent of the question using natural language processing techniques.

3. The system of claim 1 further comprising means for inputting and displaying user questions in a chat interface.

Images