system

Abstract

We provide the system. [Solution] A speech recognition means that receives voice input from the user and converts that voice into text, An analysis method that analyzes text data using a natural language processing engine and classifies the content of inquiries into specific categories, A model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on the analysis results, A response generation means that generates a response to a query using a selected generative artificial intelligence model, A response output means that provides the generated response to the user in text or voice, A system that includes this.

Description

Technical Field

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

Background Art

[0002] Patent Document 1 discloses a method for controlling a persona chatbot, which is performed by at least one processor, the method including steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a character of the chatbot, 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

Summary of the Invention

Problems to be Solved by the Invention

[0004] In modern society, the need to respond quickly and accurately to various inquiries and consultations is increasing. However, with existing methods, it is difficult to provide a consistent and high-quality response to user inquiries. In particular, users who use public telephones or smartphones are required to provide appropriate answers in real time. Furthermore, there are few systems that generate responses considering the user's inquiry history and individual context information. Therefore, there is a need for a system that can respond to user inquiries quickly and with high accuracy.

Means for Solving the Problems

[0005] The present invention provides a system that includes: speech recognition means for receiving voice input from a user and converting the voice into text; analysis means for analyzing the text data using a natural language processing engine and classifying the inquiry into specific categories; model selection means for selecting a generative artificial intelligence model to generate an appropriate response based on the analysis results; response generation means for generating a response to the inquiry using the selected generative artificial intelligence model; and response output means for providing the generated response to the user in text or voice. This system also includes history analysis means for storing the user's inquiry history and using that history data for analysis to generate a response that takes contextual information into account, and communication means for sending and receiving data between the terminal and the server via a communication line, thereby realizing immediate and appropriate responses that meet the diverse needs of the user. Furthermore, the present invention not only improves user satisfaction but also contributes to improving operational efficiency.

[0006] "Speech recognition means" refers to technology and devices for receiving speech input from a user and converting that speech into text data.

[0007] "Analysis means" refers to a technology and apparatus that analyzes text data obtained by speech recognition means using a natural language processing engine and classifies the content of inquiries into specific categories.

[0008] "Model selection means" refers to a technology and apparatus for selecting an appropriate generative artificial intelligence model based on the results of the analysis means.

[0009] "Response generation means" refers to a technology and apparatus that generates a response to a user's inquiry using a selected generative artificial intelligence model.

[0010] "Response output means" refers to a technology and apparatus that provides the generated response to the user in text or audio format.

[0011] "History analysis means" refers to a technology and apparatus that stores a user's inquiry history and uses that history data for analysis to generate a response that takes contextual information into account.

[0012] "Communication means" refers to the technology and equipment used by terminals and servers to send and receive data via communication lines. [Brief explanation of the drawing]

[0013] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] This is a sequence diagram showing the processing flow of the data processing system in Example 2, which incorporates an emotion engine. [Figure 14]It is a sequence diagram showing the processing flow of a data processing system in Application Example 2 when a sentiment engine is combined.

Embodiments for Carrying Out the Invention

[0014] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0015] First, the terms used in the following description will be explained.

[0016] In the following embodiments, a numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0017] In the following embodiments, a numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0018] In the following embodiments, a numbered storage is one or more non-volatile storage devices that store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, etc.

[0019] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0020] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0021] [First Embodiment]

[0022] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0023] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0024] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0025] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0026] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0027] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0028] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0029] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0030] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0031] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0032] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0033] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0034] This invention relates to a system for responding quickly and accurately to a variety of user inquiries, and combines speech recognition technology, natural language processing technology, and generative artificial intelligence.

[0035] System-wide configuration

[0036] This system consists of a user, a terminal, and a server. The user makes a query through the terminal, which converts it into text data using speech recognition. The server then receives the text, analyzes its content using analysis tools, selects the optimal model to generate a response, and returns it to the terminal for the user to receive.

[0037] Server operation

[0038] 1. Inquiry reception

[0039] The server receives text data sent from the terminal.

[0040] The received text is passed to the analysis device, and the analysis begins.

[0041] 2. Analysis and Categorization

[0042] The server uses parsing tools to analyze the context of the text data and classify it into specific categories.

[0043] Based on the classified categories, an appropriate generative artificial intelligence model (e.g., fortune-telling, elderly care, housing, etc.) is selected.

[0044] 3. Response generation

[0045] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection method.

[0046] The response generated using the response generation means is formatted into an appropriate format and converted into audio data as needed.

[0047] 4. Send a response

[0048] The server sends the generated response to the terminal.

[0049] Terminal operation

[0050] 1. Speech recognition

[0051] The device receives voice input from the user.

[0052] The system uses speech recognition to convert speech into text data and then sends that text data to a server.

[0053] 2. Response reception and output

[0054] The terminal receives the response data sent from the server.

[0055] The system uses a response output mechanism to provide the user with the received data as text or audio.

[0056] User actions

[0057] 1. Inquiry

[0058] Users ask questions using a device (such as a public telephone, smartphone, or tablet).

[0059] By asking questions using voice, information can be easily obtained.

[0060] 2. Confirmation of response

[0061] The user checks the response provided by the device.

[0062] You can ask further questions or inquire about additional relevant information as needed.

[0063] Specific usage examples

[0064] Fortune-telling inquiries

[0065] User: "Please tell me my fortune for this month."

[0066] Terminal: Converts audio into text data and sends it to the server.

[0067] Server: Using analysis tools, the text "Please tell me my fortune for this month" is classified into the fortune-telling category, and a generative artificial intelligence model specifically for fortune-telling is selected.

[0068] Server: The fortune-telling model generates a response such as "Your fortune this month is good" and sends it to the terminal.

[0069] Terminal: Conveys received responses to the user via voice.

[0070] Caregiving inquiries

[0071] User: "I'm looking for a senior living facility."

[0072] Terminal: Converts audio into text data and sends it to the server.

[0073] Server: Using analysis tools, the text "I am looking for a facility for the elderly" is classified into the care category, and a generative artificial intelligence model specifically for care is selected.

[0074] Server: The caregiving model generates a response such as "The recommended elderly care facility is XX," and sends it to the terminal.

[0075] Terminal: Conveys received responses to the user via text or voice.

[0076] This embodiment of the present invention enables rapid response to diverse user needs and improves user satisfaction. The system provides advanced analysis and appropriate responses to various inquiries, thus enabling it to meet a wide range of user needs.

[0077] The following describes the processing flow.

[0078] Step 1:

[0079] The user speaks a specific question into the device's microphone (e.g., "Please tell me my fortune for this month").

[0080] Step 2:

[0081] The device uses speech recognition to convert the user's voice into text data.

[0082] Send the converted text data to the server.

[0083] Step 3:

[0084] The server receives text data sent from the terminal.

[0085] The received text data is passed to an analysis tool, which uses a natural language processing engine to analyze the context of the text.

[0086] Step 4:

[0087] The server uses analysis tools to classify text data into specific categories (e.g., fortune-telling, elderly care, housing).

[0088] Based on the analysis results, an appropriate generative artificial intelligence model will be selected.

[0089] Step 5:

[0090] The server generates an appropriate response to the query using a selected generative artificial intelligence model.

[0091] The generated response is formatted into an appropriate format using a response generation means.

[0092] Step 6:

[0093] The server sends the generated response data to the terminal.

[0094] Step 7:

[0095] The terminal receives the response data sent from the server.

[0096] The system uses a response output mechanism to provide the user with the received data in text or audio format.

[0097] Step 8:

[0098] The user checks the response provided by the device.

[0099] Ask additional questions as needed, then return to step 1.

[0100] (Example 1)

[0101] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0102] Traditional systems struggled to provide quick and accurate responses to user inquiries. Furthermore, generating responses that fully considered the context and history of the inquiry was difficult, making improving user satisfaction a challenge.

[0103] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0104] In this invention, the server includes: speech recognition means for receiving voice input from a user and converting the voice into text; analysis means for analyzing the text data using a natural language engine and classifying the inquiry into specific categories; model selection means for selecting a generative artificial intelligence engine that generates an appropriate response based on the analysis results; response generation means for generating a response to the inquiry using the selected generative artificial intelligence engine; and response output means for providing the generated response data to the user as text or voice. This enables the rapid and accurate provision of responses to user inquiries and the generation of sophisticated responses that take into account context and history.

[0105] A "speech recognition means" is a technological device that receives speech input from a user and converts that speech into text data.

[0106] A "natural language engine" is a technology that analyzes text data, understands its context and meaning, and classifies the content of inquiries into specific categories.

[0107] "Analysis means" refers to a system component that uses a natural language engine to analyze the context of text data and classify the query content into a specific category.

[0108] A "generative artificial intelligence engine" is an artificial intelligence technology component that generates appropriate responses based on a specific category.

[0109] A "model selection means" is a system component that selects an appropriate generative artificial intelligence engine based on the analysis results.

[0110] "Response generation means" refers to a technology that generates responses to inquiries using a selected generative artificial intelligence engine.

[0111] A "response output means" is a system component that provides the generated response data to the user as text or audio.

[0112] "History analysis means" refers to a technology that stores the user's inquiry history and uses that history data for analysis to generate a response that takes contextual information into account.

[0113] "Communication means" refers to the technology by which a terminal and a server send and receive data via a communication path.

[0114] This invention relates to a system that responds quickly and accurately to a variety of user inquiries, and is comprised of a combination of advanced speech recognition technology, natural language processing technology, and a generative artificial intelligence engine.

[0115] System-wide configuration

[0116] This system primarily consists of three elements: user, terminal, and server. The user makes inquiries via voice through the terminal, which converts them into text data using speech recognition technology. The text data is sent to the server, which analyzes the received text and uses an optimal generative artificial intelligence engine to provide a response to the user.

[0117] Server operation

[0118] 1. Inquiry reception

[0119] The server receives text data sent from the terminal.

[0120] The received text is passed to the analysis device, and the analysis begins.

[0121] 2. Analysis and Categorization

[0122] The server uses analysis tools to analyze the context of the text data and classify it into specific categories.

[0123] Specifically, a natural language engine is used to classify the data into categories such as "fortune telling," "elderly care," and "housing."

[0124] 3. Model Selection

[0125] The server selects an appropriate generative artificial intelligence engine (e.g., GPT-3®, DALL-E, etc.) based on the analysis results.

[0126] For example, if a question like "Please tell me my fortune for this month" is classified under the fortune-telling category, a generative artificial intelligence engine specifically for fortune-telling will be selected.

[0127] 4. Response generation

[0128] The server inputs a prompt (e.g., "Please tell me my fortune for this month") into the selected generative artificial intelligence engine and generates a response.

[0129] Using a response generation tool, specific answers such as "Your luck this month is good" are created.

[0130] 5. Send a response

[0131] The server formats the generated response into the appropriate format and converts it into audio data if necessary.

[0132] Send the response to the terminal.

[0133] Terminal operation

[0134] 1. Speech recognition

[0135] The device receives voice input from the user.

[0136] Speech-to-Text API is used to convert speech into text data.

[0137] Specifically, the audio "Please tell me my fortune for this month" is converted into the text format "Please tell me my fortune for this month".

[0138] 2. Response reception and output

[0139] The terminal receives the response data sent from the server.

[0140] The system uses a response output mechanism to provide the user with the received data as text or audio.

[0141] For example, the system could tell the user via voice message, "Your fortune this month is good."

[0142] User actions

[0143] 1. Inquiry

[0144] Users ask questions using their devices (such as public telephones, smartphones, or tablets) via voice.

[0145] For example, the device receives voice input when the user says, "I'm looking for a senior care facility."

[0146] 2. Confirmation of response

[0147] The user checks the response provided by the device.

[0148] You can ask further questions or inquire about additional relevant information as needed.

[0149] Specific usage examples

[0150] Fortune-telling inquiries

[0151] User: "Please tell me my fortune for this month."

[0152] Terminal: Converts audio into text data and sends it to the server.

[0153] Server: Classifies text data into fortune-telling categories and selects a generative artificial intelligence engine specifically for fortune-telling.

[0154] Server: Uses a fortune-telling engine to generate a response such as "Your fortune this month is good" and sends it to the terminal.

[0155] Terminal: Communicates responses to the user via voice.

[0156] Caregiving inquiries

[0157] User: "I'm looking for a senior living facility."

[0158] Terminal: Converts audio into text data and sends it to the server.

[0159] Server: Classifies text data into caregiving categories and selects a generative artificial intelligence engine specifically for caregiving.

[0160] Server: Uses an engine designed for elderly care to generate responses such as "We recommend XX elderly care facility" and send them to the terminal.

[0161] Terminal: Communicates responses to the user via voice.

[0162] This system enables quick and accurate responses to a wide range of user inquiries, thereby improving user satisfaction. Furthermore, by taking past inquiry history into account, it can provide more personalized responses.

[0163] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0164] Program execution flow: Explained in steps.

[0165] Step 1: User voice input

[0166] Users make inquiries to their devices (such as smartphones and tablets) using voice commands.

[0167] For example, a user might say, "Please tell me my fortune for this month."

[0168] Input: User's voice data

[0169] Output: Audio data is input to the terminal.

[0170] Step 2: Voice recognition on the device

[0171] The device uses speech recognition technology (e.g., Google Speech-to-Text API) to convert the user's voice into text data.

[0172] The system processes the audio data and generates the text, "Please tell me my fortune for this month."

[0173] Input: User's voice data

[0174] Data processing: Converting audio data to text data.

[0175] Output: Text data "Please tell me my fortune for this month"

[0176] Step 3: Send text data

[0177] The terminal sends the converted text data to the server. HTTP or similar communication protocols are used.

[0178] The text data is sent to the server and prepared for analysis.

[0179] Input: Text data "Please tell me my fortune for this month"

[0180] Data processing: Packeting and preparing text data for transmission.

[0181] Output: Text data sent to the server

[0182] Step 4: Text message received

[0183] The server receives text data sent from the terminal.

[0184] The received text data is passed to the analysis tool.

[0185] Input: Text data sent from the device

[0186] Data processing: Preparing for analysis of received data.

[0187] Output: Text data to be analyzed

[0188] Step 5: Analyze the inquiry

[0189] The server uses a natural language engine to analyze the context of the text data and then parses the query content.

[0190] For example, the text "Please tell me my fortune for this month" would be classified under the fortune-telling category.

[0191] Input: Text data to be analyzed

[0192] Data processing: Categorization using natural language processing

[0193] Output: Category information (e.g., "Fortune Telling")

[0194] Step 6: Model Selection

[0195] The server selects an appropriate generative artificial intelligence engine based on the analysis results.

[0196] For example, if it is classified as a fortune-telling category, a generative artificial intelligence engine specifically for fortune-telling (e.g., a GPT-3 model for fortune-telling) will be selected.

[0197] Input: Category information (e.g., "Fortune Telling")

[0198] Data processing: Selection of an appropriate AI model

[0199] Output: Selected generative artificial intelligence engine

[0200] Step 7: Generating the response

[0201] The server inputs a prompt (e.g., "Please tell me my fortune for this month") into the selected generative artificial intelligence engine and generates an appropriate response.

[0202] For example, a response such as "Your luck this month is good" is generated.

[0203] Input: Selected generative artificial intelligence engine, prompt

[0204] Data processing: Response generation by generative AI

[0205] Output: Generated response data (e.g., "Your fortune this month is good")

[0206] Step 8: Sending a response

[0207] The server formats the generated response into the appropriate format and converts it into audio data as needed.

[0208] The response data is sent to the terminal.

[0209] Input: Generated response data

[0210] Data processing: Formatting of response data and conversion to audio data (if necessary).

[0211] Output: Response data ready to be sent to the terminal

[0212] Step 9: Receive response / correction

[0213] The terminal receives the response data sent from the server.

[0214] Input: Response data sent from the server

[0215] Data processing: Processing of received data

[0216] Output: Response data in a state that can be provided to the user.

[0217] Step 10: Provide a response to the user

[0218] The terminal uses a response output means to provide the user with the received data as text or audio.

[0219] For example, the system could tell the user via voice message, "Your fortune this month is good."

[0220] Input: Response data in a state that can be provided to the user.

[0221] Data calculation: Output of response data

[0222] Output: The response provided to the user (text or audio)

[0223] This processing flow enables quick and accurate responses to a wide range of user inquiries.

[0224] (Application Example 1)

[0225] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0226] In autonomous vehicles, it is crucial for users to quickly and accurately obtain various information while driving. However, existing systems often lack sufficient accuracy in speech recognition and natural language processing, sometimes providing only incorrect or incomplete information. Furthermore, their ability to appropriately acquire and respond to real-time updated information is insufficient, which can impair user convenience. The problem this invention aims to solve is to address these challenges and provide a system that allows users to always obtain the latest information appropriately.

[0227] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0228] In this invention, the server includes speech recognition means for converting speech to text, analysis means for analyzing text data and classifying the inquiry content into specific categories, and model selection means for selecting a generative artificial intelligence model based on the analysis results. This makes it possible to generate and provide a quick and accurate response when a user queries for information in real time by voice within an autonomous vehicle.

[0229] "Voice recognition means" refers to a device or software that converts voice data input by a user into text data.

[0230] "Analysis means" refers to a device or software that analyzes text data using a natural language processing engine and classifies the content of a query into a specific category.

[0231] "Model selection means" refers to a device or software that selects an appropriate generative artificial intelligence model based on the analysis results.

[0232] "Response generation means" refers to a device or software that generates a response to an inquiry using a selected generative artificial intelligence model.

[0233] "Response output means" refers to a device or software that provides the generated response to the user in text or voice.

[0234] "Communication means" refers to a device or software that connects to an online server via a communication system within a vehicle and transmits and receives data in real time.

[0235] "Voice output means" refers to a device or software that provides the user with a voice response generated using the acoustic equipment inside the vehicle.

[0236] "History analysis means" refers to a device or software that stores the user's inquiry history and uses that history data for analysis to generate a response that takes contextual information into account.

[0237] This invention provides a system that responds quickly and accurately to user inquiries made by voice within an autonomous vehicle. The following is a specific embodiment of this system.

[0238] Hardware and software configuration

[0239] Hardware configuration:

[0240] 1. Microphone: A device installed inside a vehicle to capture sound.

[0241] 2. In-vehicle communication system: This is a communication system that connects to the internet and sends and receives data with online servers.

[0242] 3. Speaker: A device for providing the generated voice response to the user.

[0243] Software configuration:

[0244] 1. Speech Recognition Method: The speech acquired from the microphone is converted into text data using the Python library speech_recognition.

[0245] 2. Analysis method: A natural language processing engine is used to analyze the context of the text data and classify the query content into specific categories.

[0246] 3. Model Selection Method: Based on the analysis results, an appropriate generative artificial intelligence model (generative AI model) is selected.

[0247] 4. Response generation means: A response to the query is generated using a selected generative AI model.

[0248] 5. Response output method: The generated response is converted into audio data using the Google Text-to-Speech (gTTS) library and provided to the user through the speaker.

[0249] Specific operation of the system

[0250] 1. Voice Recognition: The user asks a question into the microphone inside the car. For example, "Where is the nearest gas station?"

[0251] 2. Text conversion: The speech recognition system converts the user's speech into text data.

[0252] 3. Analysis: The analysis tool analyzes the text data using a natural language processing engine and classifies the sentence "Where is the nearest gas station?" into the traffic information category.

[0253] 4. Model Selection: The model selection method selects a generation AI model for traffic information.

[0254] 5. Response Generation: The response generation means uses a selected generative AI model to generate a response such as, "The nearest gas station is 2km ahead on your right."

[0255] 6. Response Output: The response output means converts the generated response into audio data and transmits it to the user through the speaker.

[0256] Specific example

[0257] For example, if a user asks "Where is the nearest gas station?" while inside an autonomous vehicle, the system will respond in the following order:

[0258] 1. The microphone captures audio data, and the speech recognition system converts it into text data such as "Where is the nearest gas station?".

[0259] 2. The analysis tool analyzes the text data and classifies it into traffic information categories.

[0260] 3. The model selection method selects an AI model for generating traffic information.

[0261] 4. The response generation means generates the response, "The nearest gas station is 2km ahead on your right."

[0262] 5. The response output means converts the response into audio data using the gTTS library and provides it to the user through the speaker.

[0263] Example of a prompt

[0264] Example of a user inquiry: "Where is the nearest gas station?"

[0265] Prompt text: "In response to the voice-recognized inquiry 'Where is the nearest gas station?', the system will provide the location of the best gas station."

[0266] Examples of input content for a generative AI model:

[0267] Inquiry: "Where is the nearest gas station?"

[0268] Reply: "The nearest gas station is 2km ahead on your right."

[0269] In this way, users can obtain quick and accurate information using voice commands within autonomous vehicles.

[0270] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0271] Step 1:

[0272] The user makes a voice inquiry into the microphone inside the vehicle. The input is the user's voice data, and the output is the voice data captured by the microphone.

[0273] Step 2:

[0274] The device's speech recognition system converts the voice data input by the user into text data. The input for this step is the captured voice data, and the output is the converted text data. Specifically, the Python library speech_recognition is used to analyze the voice data and convert it into string information.

[0275] Step 3:

[0276] The terminal sends the converted text data to the server via the in-vehicle communication system. The input for this step is the converted text data, and the output is the text data sent to the server. Communication means are used for sending and receiving data.

[0277] Step 4:

[0278] The server's analysis mechanism analyzes the received text data using a natural language processing engine. The input for this step is the received text data, and the output is the analyzed query content and its categorization. Natural language processing techniques are used for the analysis, understanding the context and classifying it into appropriate categories.

[0279] Step 5:

[0280] The server's model selection mechanism selects an appropriate generative AI model based on the analyzed results. The input to this step is the analysis results and their category classification, and the output is the selected generative AI model. The selection mechanism determines the optimal generative AI model for each query category.

[0281] Step 6:

[0282] The response generation means of the server generates an appropriate response using the selected generation AI model. The input to this step is the selected generation AI model and the analyzed inquiry content, and the output is the generated response data. This includes the process by which the generation AI model generates an optimal answer to the inquiry.

[0283] Step 7:

[0284] The server transmits the generated response data to the terminal. The input to this step is the generated response data, and the output is the response data transmitted to the terminal. Communication means are used for data transmission and reception.

[0285] Step 8:

[0286] The response output means of the terminal converts the received response data into voice data. The input to this step is the received response data, and the output is the converted voice data. Specifically, the Google Text-to-Speech (gTTS) library is used to convert text data into voice data.

[0287] Step 9:

[0288] The speaker of the terminal plays the converted voice data for the user. The input to this step is the converted voice data, and the output is the provision of voice information to the user. Through the speaker, the user can listen to the generated response.

[0289] Through this series of processes, the user can obtain quick and accurate information using voice in an autonomous vehicle.

[0290] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0291] This invention relates to a system that recognizes a user's emotions and generates a response based on those emotions, and combines speech recognition technology, natural language processing technology, generative artificial intelligence, and an emotion engine.

[0292] System-wide configuration

[0293] This system consists of a user, a terminal, and a server. The user makes inquiries through the terminal, which converts them into text data using speech recognition. The server then receives the text data and speech sentiment data, analyzes them using an analysis tool, selects an appropriate model to generate a response, and returns it to the terminal for the user to receive.

[0294] Server operation

[0295] 1. Inquiry reception

[0296] The server receives text data and sentiment data sent from the terminal.

[0297] The received data is passed to the analysis device, and the analysis begins.

[0298] 2. Analysis and categorization of text data

[0299] The server uses parsing tools to analyze the context of text data and classify it into specific categories (e.g., fortune-telling, elderly care, housing).

[0300] Based on the classified categories, an appropriate generative artificial intelligence model is selected.

[0301] 3. Analysis of emotional data

[0302] The server analyzes the emotion data obtained from the voice using an emotion engine and recognizes the user's emotional state.

[0303] The recognized emotion data is combined with the inquiry content to adjust the content and tone of the response.

[0304] 4. Response Generation

[0305] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection means.

[0306] Using the response generation means, the generated response is formatted appropriately and converted into voice data if necessary.

[0307] 5. Response Transmission

[0308] The server transmits the generated response data to the terminal.

[0309] Terminal Operations

[0310] 1. Speech Recognition

[0311] The terminal receives voice input from the user.

[0312] Using speech recognition means, the voice is converted into text data, and the text data is transmitted to the server.

[0313] At the same time, emotion data is extracted from the voice using an emotion engine and transmitted to the server.

[0314] 2. Response Reception and Output

[0315] The terminal receives the response data transmitted from the server.

[0316] Using the response output means, the received data is provided to the user as text or voice.

[0317] User actions

[0318] 1. Inquiry

[0319] Users ask questions using a device (such as a public telephone, smartphone, or tablet).

[0320] By asking questions using voice, information can be easily obtained.

[0321] 2. Confirmation of response

[0322] The user checks the response provided by the device.

[0323] You can ask further questions or inquire about additional relevant information as needed.

[0324] Specific usage examples

[0325] Fortune-telling inquiries

[0326] User: "Please tell me my fortune for this month (in an excited voice)."

[0327] Terminal: Converts audio into text data, extracts emotional data from the audio, and sends it to the server.

[0328] Server: Using analysis tools, the text "Please tell me my fortune for this month" is classified into the fortune-telling category, and a generative artificial intelligence model specifically for fortune-telling is selected.

[0329] Server: Uses an emotion engine to analyze emotional data and recognizes the state of being "excited."

[0330] Server: Based on the recognized emotion data, it generates a response such as, "Your luck this month is good, especially your career luck!" and sends it to the terminal.

[0331] Terminal: Conveys received responses to the user via voice.

[0332] Caregiving inquiries

[0333] User: "I'm looking for a senior living facility (in a tired voice)."

[0334] Terminal: Converts audio into text data, extracts emotional data from the audio, and sends it to the server.

[0335] Server: Using analysis tools, the text "I am looking for a facility for the elderly" is classified into the care category, and a generative artificial intelligence model specifically for care is selected.

[0336] Server: Uses an emotion engine to analyze emotional data and recognize the state of being "tired".

[0337] Server: Based on recognized emotion data, it generates a response such as, "We recommend XX senior care facility. You seem tired, rest is important," and sends it to the terminal.

[0338] Terminal: Conveys received responses to the user via voice.

[0339] In this embodiment of the present invention, it is possible to recognize the user's emotional state and provide an appropriate response accordingly, thereby greatly improving user satisfaction. The system can combine advanced analysis and emotion recognition for various inquiries, enabling it to provide users with more personalized services.

[0340] The following describes the processing flow.

[0341] Step 1:

[0342] The user speaks a specific question into the device's microphone (e.g., "Please tell me my fortune for this month").

[0343] Step 2:

[0344] The device uses speech recognition to convert the user's voice into text data.

[0345] Send the converted text data to the server.

[0346] Step 3:

[0347] The device uses an emotion engine to extract emotional data from the audio.

[0348] The extracted emotion data is sent to the server.

[0349] Step 4:

[0350] The server receives text data and sentiment data sent from the terminal.

[0351] Step 5:

[0352] The server passes the text data to the analysis tool, which then uses a natural language processing engine to analyze the context.

[0353] Step 6:

[0354] The server uses analysis tools to classify text data into specific categories (e.g., fortune-telling, elderly care, housing).

[0355] Step 7:

[0356] The server selects an appropriate generative artificial intelligence model based on the analysis results.

[0357] Step 8:

[0358] The server uses an emotion engine to analyze emotional data and recognize the user's emotional state.

[0359] Step 9:

[0360] The server combines the recognized emotion data with the query content to adjust the content and tone of the response.

[0361] Step 10:

[0362] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection method.

[0363] Step 11:

[0364] The server formats the response generated using the response generation means into an appropriate format and converts it into audio data as needed.

[0365] Step 12:

[0366] The server sends the generated response data to the terminal.

[0367] Step 13:

[0368] The terminal receives the response data sent from the server.

[0369] Step 14:

[0370] The terminal uses a response output mechanism to provide the user with the received data in text or voice.

[0371] Step 15:

[0372] The user checks the response provided by the device.

[0373] Ask additional questions as needed, then return to step 1.

[0374] (Example 2)

[0375] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0376] Conventional speech recognition systems have been immature in generating responses that take user emotions into account, making it difficult to enhance user satisfaction. Furthermore, they often lack accuracy in analyzing inquiries and the appropriateness of their responses, particularly in providing personalized responses that reflect emotions. As a result, the user experience deteriorates, limiting the effectiveness of the system.

[0377] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0378] In this invention, the server includes emotion analysis means for extracting emotion data from speech, analysis means for analyzing text data using a natural language processing engine and classifying the inquiry content into specific categories, and model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on the analysis results. This makes it possible to generate personalized responses that take the user's emotions into consideration.

[0379] "Voice recognition means" refers to a technology or device for converting a user's voice input into text data.

[0380] "Emotional analysis means" refers to a technology or device for extracting emotional data from voice input and determining the user's emotional state.

[0381] A "natural language processing engine" is a technology or software that analyzes text data to understand its context and meaning.

[0382] "Analysis means" refers to a technique or device for classifying text data into specific categories.

[0383] A "generative artificial intelligence model" is an artificial intelligence model designed to generate appropriate responses based on input data.

[0384] "Model selection means" refers to a technique or device for selecting an appropriate generative artificial intelligence model based on analysis results.

[0385] "Response generation means" refers to a technology or apparatus that generates a response to a user's inquiry using a selected generative artificial intelligence model.

[0386] "Response output means" refers to a technology or device that provides the generated response to the user as text or audio.

[0387] "History analysis means" refers to a technology or device that analyzes user inquiry history data and generates a response that takes contextual information into account.

[0388] "Communication means" refers to the technology or device used by a terminal and a server to send and receive data via a communication line.

[0389] "Distributed processing means" refers to a technology or device that distributes processing to efficiently perform data analysis and response generation within a server.

[0390] This invention relates to a system that recognizes a user's emotions and generates a response based on those emotions. Specifically, it is a system that combines speech recognition technology, natural language processing technology, generative artificial intelligence, and emotion analysis technology.

[0391] System-wide configuration

[0392] This system consists of a user, a terminal, and a server. The user makes inquiries through the terminal, which converts them into text data using speech recognition. The server then receives the text data and speech emotion data, analyzes them using an analysis tool, selects an appropriate generative artificial intelligence model to generate a response, and returns it to the terminal for the user to receive.

[0393] Hardware and software used

[0394] Speech recognition means: The terminal uses a microphone to receive voice input from the user and a speech recognition service such as the Google Speech-to-Text API.

[0395] Sentiment analysis methods: Sentiment analysis utilizes IBM Watson® Tone Analyzer and Microsoft® Azure® Emotion API.

[0396] Natural Language Processing Engine: Uses natural language processing libraries such as the BERT model, spaCy, and NLTK.

[0397] Generative AI Models: Advanced generative AI models such as GPT-4 (registered trademark) are used.

[0398] Communication method: The terminal and server use the HTTP protocol over the internet to send and receive data.

[0399] Distributed processing method: Cloud infrastructure (e.g., AWS®, Google Cloud) is used to efficiently perform data analysis and response generation within the server.

[0400] Specific usage examples

[0401] Fortune-telling inquiries

[0402] Example: The user asks a voice question, "Please tell me my fortune for this month (in an excited voice)."

[0403] Terminal: Converts the audio into text data and sends the text "Please tell me my fortune for this month" to the server. At the same time, it extracts the emotion data "excited" from the audio and also sends this to the server.

[0404] Server: Uses analysis tools to classify text data into the "fortune-telling" category and selects a generative artificial intelligence model specifically for fortune-telling. Uses emotion analysis technology to recognize when the user is in an "excited" state.

[0405] Server: Based on recognized sentiment data and text data, it generates the response, "Your luck this month is good, especially your career luck!"

[0406] Terminal: Converts the generated response into speech data using speech synthesis technology and provides it to the user.

[0407] Caregiving inquiries

[0408] Example: The user asks a question via voice, "I'm looking for a senior living facility (in a tired voice)."

[0409] Terminal: Converts the voice into text data and sends the text "I am looking for a senior care facility" to the server. At the same time, it extracts emotion data, "I am tired," and also sends this to the server.

[0410] Server: Uses analysis tools to classify text data into the "caregiving" category and selects a generative artificial intelligence model specifically for caregiving. Uses emotion analysis technology to recognize that the user is in a "tired" state.

[0411] Server: Based on recognized sentiment data and text data, it generates a response such as, "We recommend XX senior living facility. You seem tired; rest is important."

[0412] Terminal: Converts the generated response into speech data using speech synthesis technology and provides it to the user.

[0413] This invention makes it possible to recognize a user's emotional state and provide an appropriate response accordingly. The system can combine advanced analysis and emotion recognition to provide users with more personalized services in response to various inquiries.

[0414] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0415] Step 1:

[0416] Receiving and preprocessing voice input

[0417] The device receives voice input from the user. For example, it captures voice data such as "Please tell me my fortune for this month" through the smartphone's microphone.

[0418] Input: User's voice data (e.g., "Please tell me my fortune for this month.")

[0419] Specific operation: The device uses noise cancellation technology to improve the quality of audio data.

[0420] Output: Clear audio data with noise removed.

[0421] Step 2:

[0422] Speech recognition and text conversion

[0423] The device converts speech data into text data using speech recognition technology. Specifically, it uses the Google Speech-to-Text API for speech recognition.

[0424] Input: Clear audio data (e.g., "Please tell me my fortune for this month")

[0425] Specific operation: Send audio data to the API and retrieve text data as a response.

[0426] Output: Text data (Example: "Please tell me my fortune for this month")

[0427] Step 3:

[0428] Emotion analysis

[0429] The terminal uses sentiment analysis tools to extract sentiment data from voice data. In this case, IBM Watson Tone Analyzer is used.

[0430] Input: Clear audio data (e.g., "Please tell me my fortune for this month")

[0431] Specific operation: Voice data is sent to the emotion analysis engine, and emotion data is obtained as a result of the analysis.

[0432] Output: Sentiment data (e.g., "excited")

[0433] Step 4:

[0434] Sending data

[0435] The device sends text data and sentiment data to the server.

[0436] Input: Text data and sentiment data (e.g., "Please tell me my horoscope for this month," "I'm excited")

[0437] Specific operation: Send data to the server using an HTTP request.

[0438] Output: The server receives text data and sentiment data.

[0439] Step 5:

[0440] Text data analysis and categorization

[0441] The server analyzes the context of text data using analytical tools and classifies it into specific categories (e.g., fortune-telling, elderly care, housing). Here, the BERT model and spaCy are used.

[0442] Input: Text data (Example: "Please tell me my fortune for this month.")

[0443] Specific operation: Text data is fed into a natural language processing engine, and categories are determined based on the analysis results.

[0444] Output: Category information (e.g., "Fortune Telling")

[0445] Step 6:

[0446] Category-based model selection

[0447] The server selects an appropriate generative artificial intelligence model based on the analysis results. In this case, a generative artificial intelligence model such as GPT-4 is selected for the fortune-telling category.

[0448] Input: Category information (e.g., "Fortune Telling")

[0449] Specific operation: Use the model selection method within the server to select a relevant artificial intelligence model.

[0450] Output: Generative artificial intelligence model (e.g., GPT-4)

[0451] Step 7:

[0452] Response generation

[0453] The server uses a selected generative artificial intelligence model to generate responses based on text data and sentiment data. Examples of prompts include "Please tell me my fortune for this month" and the state of being "excited."

[0454] Input: Text data, sentiment data, generative artificial intelligence model (e.g., "Please tell me my fortune for this month," "I'm excited")

[0455] Specific operation: A prompt is fed into a generative artificial intelligence model, and the generated response is retrieved.

[0456] Output: Generated response data (Example: "Your luck this month is good, especially your career luck!")

[0457] Step 8:

[0458] Response formatting and speech conversion

[0459] The server formats the generated response into an appropriate format (e.g., JSON) and uses speech synthesis technology to convert it into audio data. In this case, the Google Text-to-Speech API is used.

[0460] Input: Generated response data (Example: "Your luck this month is good, especially your career luck!")

[0461] Specific operation: Sends a text response to the speech synthesis engine and generates speech data.

[0462] Output: Audio data or formatted text data

[0463] Step 9:

[0464] Sending a response

[0465] The server sends the generated response data to the terminal.

[0466] Input: Voice data or formatted text data (e.g., "Your luck this month is good, especially your career luck!")

[0467] Specific operation: Send data to the terminal using an HTTP response.

[0468] Output: The terminal receives the response data.

[0469] Step 10:

[0470] Providing a response

[0471] The terminal receives response data sent from the server and provides it to the user.

[0472] Input: Voice data or formatted text data (e.g., "Your luck this month is good, especially your career luck!")

[0473] Specific operation: If the data is audio, it will be played as sound from the speaker; if the data is text, it will be displayed on the screen.

[0474] Output: Provides a response to the user.

[0475] The above outlines the specific program processing flow of this system. By describing the inputs and outputs at each step in detail, the overall operation of the system is made easier to understand.

[0476] (Application Example 2)

[0477] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0478] In modern food delivery services, users experience a range of emotions when placing an order, and a response tailored to those emotions is required. However, traditional systems lacked the ability to analyze user emotions and could only provide uniform responses, resulting in a poor user experience. Furthermore, there was a lack of technology to generate appropriate responses that considered context and emotions in response to user inquiries. As a result, gaining user trust was difficult, and improving the quality of service was essential.

[0479] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes: speech recognition means that receive voice input from the user and convert the voice into text; analysis means that analyze the text data using a natural language processing engine and classify the inquiry into specific categories; model selection means that select a generative artificial intelligence model that generates an appropriate response based on the analysis results; emotion analysis means that extract emotion data from the user's voice and adjust the response content based on that emotion; response generation means that generate a response to the inquiry using the selected generative artificial intelligence model; and response output means that provide the generated response to the user in text or voice. This makes it possible to provide an appropriate response according to the user's emotions and improve the user experience.

[0480] "A speech recognition means that receives voice input from a user and converts that voice into text" refers to a technology or device that converts a user's voice into digital data and generates text data from the voice.

[0481] "Analysis means for analyzing text data using a natural language processing engine and classifying query content into specific categories" refers to a technology or apparatus that performs a process of analyzing generated text data, identifying its content, and classifying it into specific categories.

[0482] "Model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on analysis results" refers to a technology or device that selects the most appropriate generative artificial intelligence model based on the analyzed data.

[0483] "Response generation means for generating a response to a query using a selected generative artificial intelligence model" refers to a technology or apparatus that uses a selected generative artificial intelligence model to generate an appropriate response to a query.

[0484] "An emotion analysis means that extracts emotion data from a user's voice and adjusts the response content based on that emotion" refers to a technology or device that analyzes a user's voice information to recognize their emotional state and adjusts the response content based on that emotional state.

[0485] "Response output means that provides the generated response to the user in text or audio" refers to a technology or device that conveys the generated response to the user in text or audio format.

[0486] This invention is a system that combines speech recognition technology, natural language processing technology, generative artificial intelligence models, and an emotion engine to provide customizable responses tailored to the user's emotions when using a food delivery service. Specific embodiments are described below.

[0487] System-wide configuration

[0488] The system of the present invention consists of a user, a terminal, and a server. The user provides voice input through the terminal, which converts it into text data using speech recognition means. Subsequently, the server receives the text data and sentiment data, analyzes them using analysis means, selects an appropriate generative artificial intelligence model to generate a response, and returns it to the terminal for the user to receive.

[0489] Server operation

[0490] Inquiry reception:

[0491] The server receives text data and sentiment data sent from the terminal. It then passes the received data to the analysis system and begins the analysis.

[0492] Text data analysis and categorization:

[0493] The server uses analysis tools to analyze the context of text data and classify it into specific categories (e.g., food categories). Based on the classified categories, it selects an appropriate generative artificial intelligence model.

[0494] Analysis of emotional data:

[0495] The server uses an emotion engine to analyze emotional data obtained from the voice and recognize the user's emotional state. Based on the recognized emotion, it adjusts the content and tone of the response.

[0496] Response generation:

[0497] The server generates an appropriate response using a generative artificial intelligence model selected by the model selection means. The response generation means then formats the generated response into an appropriate format and converts it into audio data as needed.

[0498] Send response:

[0499] The server sends the generated response data to the terminal.

[0500] Terminal operation

[0501] Speech recognition:

[0502] The terminal receives voice input from the user, converts the voice into text data using speech recognition technology, and sends that text data to the server. Simultaneously, it extracts emotion data from the voice using an emotion engine and sends it to the server.

[0503] Response reception and output:

[0504] The terminal receives response data sent from the server and uses a response output means to provide the received data to the user as text or audio.

[0505] User actions

[0506] inquiry:

[0507] The user asks questions using their device via voice. Example: "I'd like to order a pizza (in a worried voice)."

[0508] Response confirmation:

[0509] The user confirms the response provided by the device. Example: "Thank you for ordering a pizza. Please let us know if you have any questions."

[0510] Core technology

[0511] This system is built using the following technologies:

[0512] Speech recognition technology: Uses the speech_recognition library.

[0513] Natural language processing technology: Uses the natural_language_processing engine.

[0514] Generative artificial intelligence models: Generative AI models

[0515] Emotion engine: Uses emotion_recognition technology

[0516] Specific example

[0517] For example, if a user types "I'd like to order a pizza (in a worried voice)," the server converts the speech into text data and recognizes the emotion of "worry." Based on the recognized emotion, it generates a response such as "Thank you for ordering a pizza. Please let us know if you have any questions," and provides this to the user in voice.

[0518] Example of a prompt

[0519] Examples of prompts for a generative AI model include the following:

[0520] "Generate a reassuring response when the user is worried. For example, if the user's voice input is 'I'd like to order a pizza (in a worried voice),' create a response like, 'Thank you for ordering a pizza. Please let us know if you have any questions.'"

[0521] In this way, it becomes possible to automatically generate responses that respond to the user's emotions and improve the user experience.

[0522] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0523] Step 1:

[0524] The user uses their device to perform voice input. For example, they might say, "I'd like to order a pizza (in a worried voice)." This voice is the input.

[0525] Step 2:

[0526] The device receives voice input from the user and converts the voice into text data using speech recognition technology. The process of converting voice data into text data is performed, and the output is the text data "I want to order a pizza."

[0527] Step 3:

[0528] Simultaneously, the device uses emotion analysis to extract emotion data from the user's voice. The emotion in the voice data is analyzed, and the emotion data "worry" is output.

[0529] Step 4:

[0530] The terminal sends the text data from step 2 and the sentiment data from step 3 to the server. The server then receives the user's inquiry and sentiment state.

[0531] Step 5:

[0532] The server passes the received text data to a natural language processing engine, which then categorizes the query into a specific category. In this case, the text data "I want to order a pizza" is classified under the "Food Category."

[0533] Step 6:

[0534] The server uses a model selection mechanism to select an appropriate generative artificial intelligence model based on the analysis results of the text data. In this case, a generative artificial intelligence model specifically for food delivery is selected.

[0535] Step 7:

[0536] The server uses emotion analysis tools to re-analyze the user's emotional data and adjusts its response accordingly. For example, based on the emotional data of "worry," it provides a prompt to the AI ​​model to generate a reassuring response.

[0537] Step 8:

[0538] The server uses a selected generative artificial intelligence model to generate a tailored response. Here, the text response "Thank you for ordering pizza. Please let us know if you have any questions." is generated.

[0539] Step 9:

[0540] The server formats the generated response text appropriately and converts it into audio data as needed. This stage includes the process of converting text data to audio data.

[0541] Step 10:

[0542] The server sends the generated response data to the terminal. This data includes responses in both text and audio formats.

[0543] Step 11:

[0544] The terminal receives response data from the server and provides it to the user using a response output mechanism. Specifically, it plays a voice message saying, "Thank you for ordering pizza. If you have any questions, we will be happy to assist you."

[0545] This series of steps is expected to improve the user experience of food delivery services by allowing users to receive appropriate responses tailored to their emotional state.

[0546] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0547] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0548] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0549] [Second Embodiment]

[0550] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0551] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0552] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0553] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0554] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0555] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0556] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0557] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0558] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0559] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0560] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0561] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0562] This invention relates to a system for responding quickly and accurately to a variety of user inquiries, and combines speech recognition technology, natural language processing technology, and generative artificial intelligence.

[0563] System-wide configuration

[0564] This system consists of a user, a terminal, and a server. The user makes a query through the terminal, which converts it into text data using speech recognition. The server then receives the text, analyzes its content using analysis tools, selects the optimal model to generate a response, and returns it to the terminal for the user to receive.

[0565] Server operation

[0566] 1. Inquiry reception

[0567] The server receives text data sent from the terminal.

[0568] The received text is passed to the analysis device, and the analysis begins.

[0569] 2. Analysis and Categorization

[0570] The server uses parsing tools to analyze the context of the text data and classify it into specific categories.

[0571] Based on the classified categories, an appropriate generative artificial intelligence model (e.g., fortune-telling, elderly care, housing, etc.) is selected.

[0572] 3. Response generation

[0573] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection method.

[0574] The response generated using the response generation means is formatted into an appropriate format and converted into audio data as needed.

[0575] 4. Send a response

[0576] The server sends the generated response to the terminal.

[0577] Terminal operation

[0578] 1. Speech recognition

[0579] The device receives voice input from the user.

[0580] The system uses speech recognition to convert speech into text data and then sends that text data to a server.

[0581] 2. Response reception and output

[0582] The terminal receives the response data sent from the server.

[0583] The system uses a response output mechanism to provide the user with the received data as text or audio.

[0584] User actions

[0585] 1. Inquiry

[0586] Users ask questions using a device (such as a public telephone, smartphone, or tablet).

[0587] By asking questions using voice, information can be easily obtained.

[0588] 2. Confirmation of response

[0589] The user checks the response provided by the device.

[0590] You can ask further questions or inquire about additional relevant information as needed.

[0591] Specific usage examples

[0592] Fortune-telling inquiries

[0593] User: "Please tell me my fortune for this month."

[0594] Terminal: Converts audio into text data and sends it to the server.

[0595] Server: Using analysis tools, the text "Please tell me my fortune for this month" is classified into the fortune-telling category, and a generative artificial intelligence model specifically for fortune-telling is selected.

[0596] Server: The fortune-telling model generates a response such as "Your fortune this month is good" and sends it to the terminal.

[0597] Terminal: Conveys received responses to the user via voice.

[0598] Caregiving inquiries

[0599] User: "I'm looking for a senior living facility."

[0600] Terminal: Converts audio into text data and sends it to the server.

[0601] Server: Using analysis tools, the text "I am looking for a facility for the elderly" is classified into the care category, and a generative artificial intelligence model specifically for care is selected.

[0602] Server: The caregiving model generates a response such as "The recommended elderly care facility is XX," and sends it to the terminal.

[0603] Terminal: Conveys received responses to the user via text or voice.

[0604] This embodiment of the present invention enables rapid response to diverse user needs and improves user satisfaction. The system provides advanced analysis and appropriate responses to various inquiries, thus enabling it to meet a wide range of user needs.

[0605] The following describes the processing flow.

[0606] Step 1:

[0607] The user speaks a specific question into the device's microphone (e.g., "Please tell me my fortune for this month").

[0608] Step 2:

[0609] The device uses speech recognition to convert the user's voice into text data.

[0610] Send the converted text data to the server.

[0611] Step 3:

[0612] The server receives text data sent from the terminal.

[0613] The received text data is passed to an analysis tool, which uses a natural language processing engine to analyze the context of the text.

[0614] Step 4:

[0615] The server uses analysis tools to classify text data into specific categories (e.g., fortune-telling, elderly care, housing).

[0616] Based on the analysis results, an appropriate generative artificial intelligence model will be selected.

[0617] Step 5:

[0618] The server generates an appropriate response to the query using a selected generative artificial intelligence model.

[0619] The generated response is formatted into an appropriate format using a response generation means.

[0620] Step 6:

[0621] The server sends the generated response data to the terminal.

[0622] Step 7:

[0623] The terminal receives the response data sent from the server.

[0624] The system uses a response output mechanism to provide the user with the received data in text or audio format.

[0625] Step 8:

[0626] The user checks the response provided by the device.

[0627] Ask additional questions as needed, then return to step 1.

[0628] (Example 1)

[0629] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0630] Traditional systems struggled to provide quick and accurate responses to user inquiries. Furthermore, generating responses that fully considered the context and history of the inquiry was difficult, making improving user satisfaction a challenge.

[0631] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0632] In this invention, the server includes: speech recognition means for receiving voice input from a user and converting the voice into text; analysis means for analyzing the text data using a natural language engine and classifying the inquiry into specific categories; model selection means for selecting a generative artificial intelligence engine that generates an appropriate response based on the analysis results; response generation means for generating a response to the inquiry using the selected generative artificial intelligence engine; and response output means for providing the generated response data to the user as text or voice. This enables the rapid and accurate provision of responses to user inquiries and the generation of sophisticated responses that take into account context and history.

[0633] A "speech recognition means" is a technological device that receives speech input from a user and converts that speech into text data.

[0634] A "natural language engine" is a technology that analyzes text data, understands its context and meaning, and classifies the content of inquiries into specific categories.

[0635] "Analysis means" refers to a system component that uses a natural language engine to analyze the context of text data and classify the query content into a specific category.

[0636] A "generative artificial intelligence engine" is an artificial intelligence technology component that generates appropriate responses based on a specific category.

[0637] A "model selection means" is a system component that selects an appropriate generative artificial intelligence engine based on the analysis results.

[0638] "Response generation means" refers to a technology that generates responses to inquiries using a selected generative artificial intelligence engine.

[0639] A "response output means" is a system component that provides the generated response data to the user as text or audio.

[0640] "History analysis means" refers to a technology that stores the user's inquiry history and uses that history data for analysis to generate a response that takes contextual information into account.

[0641] "Communication means" refers to the technology by which a terminal and a server send and receive data via a communication path.

[0642] This invention relates to a system that responds quickly and accurately to a variety of user inquiries, and is comprised of a combination of advanced speech recognition technology, natural language processing technology, and a generative artificial intelligence engine.

[0643] System-wide configuration

[0644] This system primarily consists of three elements: user, terminal, and server. The user makes inquiries via voice through the terminal, which converts them into text data using speech recognition technology. The text data is sent to the server, which analyzes the received text and uses an optimal generative artificial intelligence engine to provide a response to the user.

[0645] Server operation

[0646] 1. Inquiry reception

[0647] The server receives text data sent from the terminal.

[0648] The received text is passed to the analysis device, and the analysis begins.

[0649] 2. Analysis and Categorization

[0650] The server uses analysis tools to analyze the context of the text data and classify it into specific categories.

[0651] Specifically, a natural language engine is used to classify the data into categories such as "fortune telling," "elderly care," and "housing."

[0652] 3. Model Selection

[0653] The server selects an appropriate generative artificial intelligence engine (e.g., GPT-3, DALL-E, etc.) based on the analysis results.

[0654] For example, if a question like "Please tell me my fortune for this month" is classified under the fortune-telling category, a generative artificial intelligence engine specifically for fortune-telling will be selected.

[0655] 4. Response generation

[0656] The server inputs a prompt (e.g., "Please tell me my fortune for this month") into the selected generative artificial intelligence engine and generates a response.

[0657] Using a response generation tool, specific answers such as "Your luck this month is good" are created.

[0658] 5. Send a response

[0659] The server formats the generated response into the appropriate format and converts it into audio data if necessary.

[0660] Send the response to the terminal.

[0661] Terminal operation

[0662] 1. Speech recognition

[0663] The device receives voice input from the user.

[0664] Speech-to-Text API is used to convert speech into text data.

[0665] Specifically, the audio "Please tell me my fortune for this month" is converted into the text format "Please tell me my fortune for this month".

[0666] 2. Response reception and output

[0667] The terminal receives the response data sent from the server.

[0668] The system uses a response output mechanism to provide the user with the received data as text or audio.

[0669] For example, the system could tell the user via voice message, "Your fortune this month is good."

[0670] User actions

[0671] 1. Inquiry

[0672] Users ask questions using their devices (such as public telephones, smartphones, or tablets) via voice.

[0673] For example, the device receives voice input when the user says, "I'm looking for a senior care facility."

[0674] 2. Confirmation of response

[0675] The user checks the response provided by the device.

[0676] You can ask further questions or inquire about additional relevant information as needed.

[0677] Specific usage examples

[0678] Fortune-telling inquiries

[0679] User: "Please tell me my fortune for this month."

[0680] Terminal: Converts audio into text data and sends it to the server.

[0681] Server: Classifies text data into fortune-telling categories and selects a generative artificial intelligence engine specifically for fortune-telling.

[0682] Server: Uses a fortune-telling engine to generate a response such as "Your fortune this month is good" and sends it to the terminal.

[0683] Terminal: Communicates responses to the user via voice.

[0684] Caregiving inquiries

[0685] User: "I'm looking for a senior living facility."

[0686] Terminal: Converts audio into text data and sends it to the server.

[0687] Server: Classifies text data into caregiving categories and selects a generative artificial intelligence engine specifically for caregiving.

[0688] Server: Uses an engine designed for elderly care to generate responses such as "We recommend XX elderly care facility" and send them to the terminal.

[0689] Terminal: Communicates responses to the user via voice.

[0690] This system enables quick and accurate responses to a wide range of user inquiries, thereby improving user satisfaction. Furthermore, by taking past inquiry history into account, it can provide more personalized responses.

[0691] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0692] Program execution flow: Explained in steps.

[0693] Step 1: User voice input

[0694] Users make inquiries to their devices (such as smartphones and tablets) using voice commands.

[0695] For example, a user might say, "Please tell me my fortune for this month."

[0696] Input: User's voice data

[0697] Output: Audio data is input to the terminal.

[0698] Step 2: Voice recognition on the device

[0699] The device uses speech recognition technology (e.g., Google Speech-to-Text API) to convert the user's voice into text data.

[0700] The system processes the audio data and generates the text, "Please tell me my fortune for this month."

[0701] Input: User's voice data

[0702] Data processing: Converting audio data to text data.

[0703] Output: Text data "Please tell me my fortune for this month"

[0704] Step 3: Send text data

[0705] The terminal sends the converted text data to the server. HTTP or similar communication protocols are used.

[0706] The text data is sent to the server and prepared for analysis.

[0707] Input: Text data "Please tell me my fortune for this month"

[0708] Data processing: Packeting and preparing text data for transmission.

[0709] Output: Text data sent to the server

[0710] Step 4: Text message received

[0711] The server receives text data sent from the terminal.

[0712] The received text data is passed to the analysis tool.

[0713] Input: Text data sent from the device

[0714] Data processing: Preparing for analysis of received data.

[0715] Output: Text data to be analyzed

[0716] Step 5: Analyze the inquiry

[0717] The server uses a natural language engine to analyze the context of the text data and then parses the query content.

[0718] For example, the text "Please tell me my fortune for this month" would be classified under the fortune-telling category.

[0719] Input: Text data to be analyzed

[0720] Data processing: Categorization using natural language processing

[0721] Output: Category information (e.g., "Fortune Telling")

[0722] Step 6: Model Selection

[0723] The server selects an appropriate generative artificial intelligence engine based on the analysis results.

[0724] For example, if it is classified as a fortune-telling category, a generative artificial intelligence engine specifically for fortune-telling (e.g., a GPT-3 model for fortune-telling) will be selected.

[0725] Input: Category information (e.g., "Fortune Telling")

[0726] Data processing: Selection of an appropriate AI model

[0727] Output: Selected generative artificial intelligence engine

[0728] Step 7: Generating the response

[0729] The server inputs a prompt (e.g., "Please tell me my fortune for this month") into the selected generative artificial intelligence engine and generates an appropriate response.

[0730] For example, a response such as "Your luck this month is good" is generated.

[0731] Input: Selected generative artificial intelligence engine, prompt

[0732] Data processing: Response generation by generative AI

[0733] Output: Generated response data (e.g., "Your fortune this month is good")

[0734] Step 8: Sending a response

[0735] The server formats the generated response into the appropriate format and converts it into audio data as needed.

[0736] The response data is sent to the terminal.

[0737] Input: Generated response data

[0738] Data processing: Formatting of response data and conversion to audio data (if necessary).

[0739] Output: Response data ready to be sent to the terminal

[0740] Step 9: Receive response / correction

[0741] The terminal receives the response data sent from the server.

[0742] Input: Response data sent from the server

[0743] Data processing: Processing of received data

[0744] Output: Response data in a state that can be provided to the user.

[0745] Step 10: Provide a response to the user

[0746] The terminal uses a response output means to provide the user with the received data as text or audio.

[0747] For example, the system could tell the user via voice message, "Your fortune this month is good."

[0748] Input: Response data in a state that can be provided to the user.

[0749] Data calculation: Output of response data

[0750] Output: The response provided to the user (text or audio)

[0751] This processing flow enables quick and accurate responses to a wide range of user inquiries.

[0752] (Application Example 1)

[0753] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0754] In autonomous vehicles, it is crucial for users to quickly and accurately obtain various information while driving. However, existing systems often lack sufficient accuracy in speech recognition and natural language processing, sometimes providing only incorrect or incomplete information. Furthermore, their ability to appropriately acquire and respond to real-time updated information is insufficient, which can impair user convenience. The problem this invention aims to solve is to address these challenges and provide a system that allows users to always obtain the latest information appropriately.

[0755] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0756] In this invention, the server includes speech recognition means for converting speech to text, analysis means for analyzing text data and classifying the inquiry content into specific categories, and model selection means for selecting a generative artificial intelligence model based on the analysis results. This makes it possible to generate and provide a quick and accurate response when a user queries for information in real time by voice within an autonomous vehicle.

[0757] "Voice recognition means" refers to a device or software that converts voice data input by a user into text data.

[0758] "Analysis means" refers to a device or software that analyzes text data using a natural language processing engine and classifies the content of a query into a specific category.

[0759] "Model selection means" refers to a device or software that selects an appropriate generative artificial intelligence model based on the analysis results.

[0760] "Response generation means" refers to a device or software that generates a response to an inquiry using a selected generative artificial intelligence model.

[0761] "Response output means" refers to a device or software that provides the generated response to the user in text or voice.

[0762] "Communication means" refers to a device or software that connects to an online server via a communication system within a vehicle and transmits and receives data in real time.

[0763] "Voice output means" refers to a device or software that provides the user with a voice response generated using the acoustic equipment inside the vehicle.

[0764] "History analysis means" refers to a device or software that stores the user's inquiry history and uses that history data for analysis to generate a response that takes contextual information into account.

[0765] This invention provides a system that responds quickly and accurately to user inquiries made by voice within an autonomous vehicle. The following is a specific embodiment of this system.

[0766] Hardware and software configuration

[0767] Hardware configuration:

[0768] 1. Microphone: A device installed inside a vehicle to capture sound.

[0769] 2. In-vehicle communication system: This is a communication system that connects to the internet and sends and receives data with online servers.

[0770] 3. Speaker: A device for providing the generated voice response to the user.

[0771] Software configuration:

[0772] 1. Speech Recognition Method: The speech acquired from the microphone is converted into text data using the Python library speech_recognition.

[0773] 2. Analysis method: A natural language processing engine is used to analyze the context of the text data and classify the query content into specific categories.

[0774] 3. Model Selection Method: Based on the analysis results, an appropriate generative artificial intelligence model (generative AI model) is selected.

[0775] 4. Response generation means: A response to the query is generated using a selected generative AI model.

[0776] 5. Response output method: The generated response is converted into audio data using the Google Text-to-Speech (gTTS) library and provided to the user through the speaker.

[0777] Specific operation of the system

[0778] 1. Voice Recognition: The user asks a question into the microphone inside the car. For example, "Where is the nearest gas station?"

[0779] 2. Text conversion: The speech recognition system converts the user's speech into text data.

[0780] 3. Analysis: The analysis tool analyzes the text data using a natural language processing engine and classifies the sentence "Where is the nearest gas station?" into the traffic information category.

[0781] 4. Model Selection: The model selection method selects a generation AI model for traffic information.

[0782] 5. Response Generation: The response generation means uses a selected generative AI model to generate a response such as, "The nearest gas station is 2km ahead on your right."

[0783] 6. Response Output: The response output means converts the generated response into audio data and transmits it to the user through the speaker.

[0784] Specific example

[0785] For example, if a user asks "Where is the nearest gas station?" while inside an autonomous vehicle, the system will respond in the following order:

[0786] 1. The microphone captures audio data, and the speech recognition system converts it into text data such as "Where is the nearest gas station?".

[0787] 2. The analysis tool analyzes the text data and classifies it into traffic information categories.

[0788] 3. The model selection method selects an AI model for generating traffic information.

[0789] 4. The response generation means generates the response, "The nearest gas station is 2km ahead on your right."

[0790] 5. The response output means converts the response into audio data using the gTTS library and provides it to the user through the speaker.

[0791] Example of a prompt

[0792] Example of a user inquiry: "Where is the nearest gas station?"

[0793] Prompt text: "In response to the voice-recognized inquiry 'Where is the nearest gas station?', the system will provide the location of the best gas station."

[0794] Examples of input content for a generative AI model:

[0795] Inquiry: "Where is the nearest gas station?"

[0796] Reply: "The nearest gas station is 2km ahead on your right."

[0797] In this way, users can obtain quick and accurate information using voice commands within autonomous vehicles.

[0798] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0799] Step 1:

[0800] The user makes a voice inquiry into the microphone inside the vehicle. The input is the user's voice data, and the output is the voice data captured by the microphone.

[0801] Step 2:

[0802] The device's speech recognition system converts the voice data input by the user into text data. The input for this step is the captured voice data, and the output is the converted text data. Specifically, the Python library speech_recognition is used to analyze the voice data and convert it into string information.

[0803] Step 3:

[0804] The terminal sends the converted text data to the server via the in-vehicle communication system. The input for this step is the converted text data, and the output is the text data sent to the server. Communication means are used for sending and receiving data.

[0805] Step 4:

[0806] The server's analysis mechanism analyzes the received text data using a natural language processing engine. The input for this step is the received text data, and the output is the analyzed query content and its categorization. Natural language processing techniques are used for the analysis, understanding the context and classifying it into appropriate categories.

[0807] Step 5:

[0808] The server's model selection mechanism selects an appropriate generative AI model based on the analyzed results. The input to this step is the analysis results and their category classification, and the output is the selected generative AI model. The selection mechanism determines the optimal generative AI model for each query category.

[0809] Step 6:

[0810] The server's response generation mechanism generates an appropriate response using a selected generative AI model. The inputs to this step are the selected generative AI model and the analyzed query, while the output is the generated response data. This includes the process by which the generative AI model generates the optimal answer to the query.

[0811] Step 7:

[0812] The server sends the generated response data to the terminal. The input to this step is the generated response data, and the output is the response data sent to the terminal. Communication means are used for sending and receiving data.

[0813] Step 8:

[0814] The terminal's response output mechanism converts the received response data into speech data. The input for this step is the received response data, and the output is the converted speech data. Specifically, the Google Text-to-Speech (gTTS) library is used to convert the text data into speech data.

[0815] Step 9:

[0816] The device's speaker plays the converted audio data to the user. The input in this step is the converted audio data, and the output is the provision of audio information to the user. Through the speaker, the user can hear the generated response.

[0817] This series of processes will enable users to obtain quick and accurate information using voice commands within autonomous vehicles.

[0818] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0819] This invention relates to a system that recognizes a user's emotions and generates a response based on those emotions, and combines speech recognition technology, natural language processing technology, generative artificial intelligence, and an emotion engine.

[0820] System-wide configuration

[0821] This system consists of a user, a terminal, and a server. The user makes inquiries through the terminal, which converts them into text data using speech recognition. The server then receives the text data and speech sentiment data, analyzes them using an analysis tool, selects an appropriate model to generate a response, and returns it to the terminal for the user to receive.

[0822] Server operation

[0823] 1. Inquiry reception

[0824] The server receives text data and sentiment data sent from the terminal.

[0825] The received data is passed to the analysis device, and the analysis begins.

[0826] 2. Analysis and categorization of text data

[0827] The server uses parsing tools to analyze the context of text data and classify it into specific categories (e.g., fortune-telling, elderly care, housing).

[0828] Based on the classified categories, an appropriate generative artificial intelligence model is selected.

[0829] 3. Analysis of emotional data

[0830] The server uses an emotion engine to analyze emotional data obtained from the voice and recognize the user's emotional state.

[0831] The recognized emotion data is combined with the query content to adjust the content and tone of the response.

[0832] 4. Response generation

[0833] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection method.

[0834] Using a response generation means, the generated response is formatted into an appropriate format and converted into audio data as needed.

[0835] 5. Send a response

[0836] The server sends the generated response data to the terminal.

[0837] Terminal operation

[0838] 1. Speech recognition

[0839] The device receives voice input from the user.

[0840] The system uses speech recognition to convert speech into text data and then sends that text data to a server.

[0841] Simultaneously, an emotion engine is used to extract emotion data from the audio and send it to the server.

[0842] 2. Response reception and output

[0843] The terminal receives the response data sent from the server.

[0844] The system uses a response output mechanism to provide the user with the received data as text or audio.

[0845] User actions

[0846] 1. Inquiry

[0847] Users ask questions using a device (such as a public telephone, smartphone, or tablet).

[0848] By asking questions using voice, information can be easily obtained.

[0849] 2. Confirmation of response

[0850] The user checks the response provided by the device.

[0851] You can ask further questions or inquire about additional relevant information as needed.

[0852] Specific usage examples

[0853] Fortune-telling inquiries

[0854] User: "Please tell me my fortune for this month (in an excited voice)."

[0855] Terminal: Converts audio into text data, extracts emotional data from the audio, and sends it to the server.

[0856] Server: Using analysis tools, the text "Please tell me my fortune for this month" is classified into the fortune-telling category, and a generative artificial intelligence model specifically for fortune-telling is selected.

[0857] Server: Uses an emotion engine to analyze emotional data and recognizes the state of being "excited."

[0858] Server: Based on the recognized emotion data, it generates a response such as, "Your luck this month is good, especially your career luck!" and sends it to the terminal.

[0859] Terminal: Conveys received responses to the user via voice.

[0860] Caregiving inquiries

[0861] User: "I'm looking for a senior living facility (in a tired voice)."

[0862] Terminal: Converts audio into text data, extracts emotional data from the audio, and sends it to the server.

[0863] Server: Using analysis tools, the text "I am looking for a facility for the elderly" is classified into the care category, and a generative artificial intelligence model specifically for care is selected.

[0864] Server: Uses an emotion engine to analyze emotional data and recognize the state of being "tired".

[0865] Server: Based on recognized emotion data, it generates a response such as, "We recommend XX senior care facility. You seem tired, rest is important," and sends it to the terminal.

[0866] Terminal: Conveys received responses to the user via voice.

[0867] In this embodiment of the present invention, it is possible to recognize the user's emotional state and provide an appropriate response accordingly, thereby greatly improving user satisfaction. The system can combine advanced analysis and emotion recognition for various inquiries, enabling it to provide users with more personalized services.

[0868] The following describes the processing flow.

[0869] Step 1:

[0870] The user speaks a specific question into the device's microphone (e.g., "Please tell me my fortune for this month").

[0871] Step 2:

[0872] The device uses speech recognition to convert the user's voice into text data.

[0873] Send the converted text data to the server.

[0874] Step 3:

[0875] The device uses an emotion engine to extract emotional data from the audio.

[0876] The extracted emotion data is sent to the server.

[0877] Step 4:

[0878] The server receives text data and sentiment data sent from the terminal.

[0879] Step 5:

[0880] The server passes the text data to the analysis tool, which then uses a natural language processing engine to analyze the context.

[0881] Step 6:

[0882] The server uses analysis tools to classify text data into specific categories (e.g., fortune-telling, elderly care, housing).

[0883] Step 7:

[0884] The server selects an appropriate generative artificial intelligence model based on the analysis results.

[0885] Step 8:

[0886] The server uses an emotion engine to analyze emotional data and recognize the user's emotional state.

[0887] Step 9:

[0888] The server combines the recognized emotion data with the query content to adjust the content and tone of the response.

[0889] Step 10:

[0890] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection method.

[0891] Step 11:

[0892] The server formats the response generated using the response generation means into an appropriate format and converts it into audio data as needed.

[0893] Step 12:

[0894] The server sends the generated response data to the terminal.

[0895] Step 13:

[0896] The terminal receives the response data sent from the server.

[0897] Step 14:

[0898] The terminal uses a response output mechanism to provide the user with the received data in text or voice.

[0899] Step 15:

[0900] The user checks the response provided by the device.

[0901] Ask additional questions as needed, then return to step 1.

[0902] (Example 2)

[0903] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0904] Conventional speech recognition systems have been immature in generating responses that take user emotions into account, making it difficult to enhance user satisfaction. Furthermore, they often lack accuracy in analyzing inquiries and the appropriateness of their responses, particularly in providing personalized responses that reflect emotions. As a result, the user experience deteriorates, limiting the effectiveness of the system.

[0905] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0906] In this invention, the server includes emotion analysis means for extracting emotion data from speech, analysis means for analyzing text data using a natural language processing engine and classifying the inquiry content into specific categories, and model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on the analysis results. This makes it possible to generate personalized responses that take the user's emotions into consideration.

[0907] "Voice recognition means" refers to a technology or device for converting a user's voice input into text data.

[0908] "Emotional analysis means" refers to a technology or device for extracting emotional data from voice input and determining the user's emotional state.

[0909] A "natural language processing engine" is a technology or software that analyzes text data to understand its context and meaning.

[0910] "Analysis means" refers to a technique or device for classifying text data into specific categories.

[0911] A "generative artificial intelligence model" is an artificial intelligence model designed to generate appropriate responses based on input data.

[0912] "Model selection means" refers to a technique or device for selecting an appropriate generative artificial intelligence model based on analysis results.

[0913] "Response generation means" refers to a technology or apparatus that generates a response to a user's inquiry using a selected generative artificial intelligence model.

[0914] "Response output means" refers to a technology or device that provides the generated response to the user as text or audio.

[0915] "History analysis means" refers to a technology or device that analyzes user inquiry history data and generates a response that takes contextual information into account.

[0916] "Communication means" refers to the technology or device used by a terminal and a server to send and receive data via a communication line.

[0917] "Distributed processing means" refers to a technology or device that distributes processing to efficiently perform data analysis and response generation within a server.

[0918] This invention relates to a system that recognizes a user's emotions and generates a response based on those emotions. Specifically, it is a system that combines speech recognition technology, natural language processing technology, generative artificial intelligence, and emotion analysis technology.

[0919] System-wide configuration

[0920] This system consists of a user, a terminal, and a server. The user makes inquiries through the terminal, which converts them into text data using speech recognition. The server then receives the text data and speech emotion data, analyzes them using an analysis tool, selects an appropriate generative artificial intelligence model to generate a response, and returns it to the terminal for the user to receive.

[0921] Hardware and software used

[0922] Speech recognition means: The terminal uses a microphone to receive voice input from the user and a speech recognition service such as the Google Speech-to-Text API.

[0923] Sentiment analysis methods: IBM Watson Tone Analyzer and Microsoft Azure Emotion API are used for sentiment analysis.

[0924] Natural Language Processing Engine: Uses natural language processing libraries such as the BERT model, spaCy, and NLTK.

[0925] Generative AI Models: Advanced generative AI models such as GPT-4 are used.

[0926] Communication method: The terminal and server use the HTTP protocol over the internet to send and receive data.

[0927] Distributed processing methods: Cloud infrastructure (e.g., AWS, Google Cloud) is used to efficiently perform data analysis and response generation within the server.

[0928] Specific usage examples

[0929] Fortune-telling inquiries

[0930] Example: The user asks a voice question, "Please tell me my fortune for this month (in an excited voice)."

[0931] Terminal: Converts the audio into text data and sends the text "Please tell me my fortune for this month" to the server. At the same time, it extracts the emotion data "excited" from the audio and also sends this to the server.

[0932] Server: Uses analysis tools to classify text data into the "fortune-telling" category and selects a generative artificial intelligence model specifically for fortune-telling. Uses emotion analysis technology to recognize when the user is in an "excited" state.

[0933] Server: Based on recognized sentiment data and text data, it generates the response, "Your luck this month is good, especially your career luck!"

[0934] Terminal: Converts the generated response into speech data using speech synthesis technology and provides it to the user.

[0935] Caregiving inquiries

[0936] Example: The user asks a question via voice, "I'm looking for a senior living facility (in a tired voice)."

[0937] Terminal: Converts the voice into text data and sends the text "I am looking for a senior care facility" to the server. At the same time, it extracts emotion data, "I am tired," and also sends this to the server.

[0938] Server: Uses analysis tools to classify text data into the "caregiving" category and selects a generative artificial intelligence model specifically for caregiving. Uses emotion analysis technology to recognize that the user is in a "tired" state.

[0939] Server: Based on recognized sentiment data and text data, it generates a response such as, "We recommend XX senior living facility. You seem tired; rest is important."

[0940] Terminal: Converts the generated response into speech data using speech synthesis technology and provides it to the user.

[0941] This invention makes it possible to recognize a user's emotional state and provide an appropriate response accordingly. The system can combine advanced analysis and emotion recognition to provide users with more personalized services in response to various inquiries.

[0942] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0943] Step 1:

[0944] Receiving and preprocessing voice input

[0945] The device receives voice input from the user. For example, it captures voice data such as "Please tell me my fortune for this month" through the smartphone's microphone.

[0946] Input: User's voice data (e.g., "Please tell me my fortune for this month.")

[0947] Specific operation: The device uses noise cancellation technology to improve the quality of audio data.

[0948] Output: Clear audio data with noise removed.

[0949] Step 2:

[0950] Speech recognition and text conversion

[0951] The device converts speech data into text data using speech recognition technology. Specifically, it uses the Google Speech-to-Text API for speech recognition.

[0952] Input: Clear audio data (e.g., "Please tell me my fortune for this month")

[0953] Specific operation: Send audio data to the API and retrieve text data as a response.

[0954] Output: Text data (Example: "Please tell me my fortune for this month")

[0955] Step 3:

[0956] Emotion analysis

[0957] The terminal uses sentiment analysis tools to extract sentiment data from voice data. In this case, IBM Watson Tone Analyzer is used.

[0958] Input: Clear audio data (e.g., "Please tell me my fortune for this month")

[0959] Specific operation: Voice data is sent to the emotion analysis engine, and emotion data is obtained as a result of the analysis.

[0960] Output: Sentiment data (e.g., "excited")

[0961] Step 4:

[0962] Sending data

[0963] The device sends text data and sentiment data to the server.

[0964] Input: Text data and sentiment data (e.g., "Please tell me my horoscope for this month," "I'm excited")

[0965] Specific operation: Send data to the server using an HTTP request.

[0966] Output: The server receives text data and sentiment data.

[0967] Step 5:

[0968] Text data analysis and categorization

[0969] The server analyzes the context of text data using analytical tools and classifies it into specific categories (e.g., fortune-telling, elderly care, housing). Here, the BERT model and spaCy are used.

[0970] Input: Text data (Example: "Please tell me my fortune for this month.")

[0971] Specific operation: Text data is fed into a natural language processing engine, and categories are determined based on the analysis results.

[0972] Output: Category information (e.g., "Fortune Telling")

[0973] Step 6:

[0974] Category-based model selection

[0975] The server selects an appropriate generative artificial intelligence model based on the analysis results. In this case, a generative artificial intelligence model such as GPT-4 is selected for the fortune-telling category.

[0976] Input: Category information (e.g., "Fortune Telling")

[0977] Specific operation: Use the model selection method within the server to select a relevant artificial intelligence model.

[0978] Output: Generative artificial intelligence model (e.g., GPT-4)

[0979] Step 7:

[0980] Response generation

[0981] The server uses a selected generative artificial intelligence model to generate responses based on text data and sentiment data. Examples of prompts include "Please tell me my fortune for this month" and the state of being "excited."

[0982] Input: Text data, sentiment data, generative artificial intelligence model (e.g., "Please tell me my fortune for this month," "I'm excited")

[0983] Specific operation: A prompt is fed into a generative artificial intelligence model, and the generated response is retrieved.

[0984] Output: Generated response data (Example: "Your luck this month is good, especially your career luck!")

[0985] Step 8:

[0986] Response formatting and speech conversion

[0987] The server formats the generated response into an appropriate format (e.g., JSON) and uses speech synthesis technology to convert it into audio data. In this case, the Google Text-to-Speech API is used.

[0988] Input: Generated response data (Example: "Your luck this month is good, especially your career luck!")

[0989] Specific operation: Sends a text response to the speech synthesis engine and generates speech data.

[0990] Output: Audio data or formatted text data

[0991] Step 9:

[0992] Sending a response

[0993] The server sends the generated response data to the terminal.

[0994] Input: Voice data or formatted text data (e.g., "Your luck this month is good, especially your career luck!")

[0995] Specific operation: Send data to the terminal using an HTTP response.

[0996] Output: The terminal receives the response data.

[0997] Step 10:

[0998] Providing a response

[0999] The terminal receives response data sent from the server and provides it to the user.

[1000] Input: Voice data or formatted text data (e.g., "Your luck this month is good, especially your career luck!")

[1001] Specific operation: If the data is audio, it will be played as sound from the speaker; if the data is text, it will be displayed on the screen.

[1002] Output: Provides a response to the user.

[1003] The above outlines the specific program processing flow of this system. By describing the inputs and outputs at each step in detail, the overall operation of the system is made easier to understand.

[1004] (Application Example 2)

[1005] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[1006] In modern food delivery services, users experience a range of emotions when placing an order, and a response tailored to those emotions is required. However, traditional systems lacked the ability to analyze user emotions and could only provide uniform responses, resulting in a poor user experience. Furthermore, there was a lack of technology to generate appropriate responses that considered context and emotions in response to user inquiries. As a result, gaining user trust was difficult, and improving the quality of service was essential.

[1007] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes: speech recognition means that receive voice input from the user and convert the voice into text; analysis means that analyze the text data using a natural language processing engine and classify the inquiry into specific categories; model selection means that select a generative artificial intelligence model that generates an appropriate response based on the analysis results; emotion analysis means that extract emotion data from the user's voice and adjust the response content based on that emotion; response generation means that generate a response to the inquiry using the selected generative artificial intelligence model; and response output means that provide the generated response to the user in text or voice. This makes it possible to provide an appropriate response according to the user's emotions and improve the user experience.

[1008] "A speech recognition means that receives voice input from a user and converts that voice into text" refers to a technology or device that converts a user's voice into digital data and generates text data from the voice.

[1009] "Analysis means for analyzing text data using a natural language processing engine and classifying query content into specific categories" refers to a technology or apparatus that performs a process of analyzing generated text data, identifying its content, and classifying it into specific categories.

[1010] "Model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on analysis results" refers to a technology or device that selects the most appropriate generative artificial intelligence model based on the analyzed data.

[1011] "Response generation means for generating a response to a query using a selected generative artificial intelligence model" refers to a technology or apparatus that uses a selected generative artificial intelligence model to generate an appropriate response to a query.

[1012] "An emotion analysis means that extracts emotion data from a user's voice and adjusts the response content based on that emotion" refers to a technology or device that analyzes a user's voice information to recognize their emotional state and adjusts the response content based on that emotional state.

[1013] "Response output means that provides the generated response to the user in text or audio" refers to a technology or device that conveys the generated response to the user in text or audio format.

[1014] This invention is a system that combines speech recognition technology, natural language processing technology, generative artificial intelligence models, and an emotion engine to provide customizable responses tailored to the user's emotions when using a food delivery service. Specific embodiments are described below.

[1015] System-wide configuration

[1016] The system of the present invention consists of a user, a terminal, and a server. The user provides voice input through the terminal, which converts it into text data using speech recognition means. Subsequently, the server receives the text data and sentiment data, analyzes them using analysis means, selects an appropriate generative artificial intelligence model to generate a response, and returns it to the terminal for the user to receive.

[1017] Server operation

[1018] Inquiry reception:

[1019] The server receives text data and sentiment data sent from the terminal. It then passes the received data to the analysis system and begins the analysis.

[1020] Text data analysis and categorization:

[1021] The server uses analysis tools to analyze the context of text data and classify it into specific categories (e.g., food categories). Based on the classified categories, it selects an appropriate generative artificial intelligence model.

[1022] Analysis of emotional data:

[1023] The server uses an emotion engine to analyze emotional data obtained from the voice and recognize the user's emotional state. Based on the recognized emotion, it adjusts the content and tone of the response.

[1024] Response generation:

[1025] The server generates an appropriate response using a generative artificial intelligence model selected by the model selection means. The response generation means then formats the generated response into an appropriate format and converts it into audio data as needed.

[1026] Send response:

[1027] The server sends the generated response data to the terminal.

[1028] Terminal operation

[1029] Speech recognition:

[1030] The terminal receives voice input from the user, converts the voice into text data using speech recognition technology, and sends that text data to the server. Simultaneously, it extracts emotion data from the voice using an emotion engine and sends it to the server.

[1031] Response reception and output:

[1032] The terminal receives response data sent from the server and uses a response output means to provide the received data to the user as text or audio.

[1033] User actions

[1034] inquiry:

[1035] The user asks questions using their device via voice. Example: "I'd like to order a pizza (in a worried voice)."

[1036] Response confirmation:

[1037] The user confirms the response provided by the device. Example: "Thank you for ordering a pizza. Please let us know if you have any questions."

[1038] Core technology

[1039] This system is built using the following technologies:

[1040] Speech recognition technology: Uses the speech_recognition library.

[1041] Natural language processing technology: Uses the natural_language_processing engine.

[1042] Generative artificial intelligence models: Generative AI models

[1043] Emotion engine: Uses emotion_recognition technology

[1044] Specific example

[1045] For example, if a user types "I'd like to order a pizza (in a worried voice)," the server converts the speech into text data and recognizes the emotion of "worry." Based on the recognized emotion, it generates a response such as "Thank you for ordering a pizza. Please let us know if you have any questions," and provides this to the user in voice.

[1046] Example of a prompt

[1047] Examples of prompts for a generative AI model include the following:

[1048] "Generate a reassuring response when the user is worried. For example, if the user's voice input is 'I'd like to order a pizza (in a worried voice),' create a response like, 'Thank you for ordering a pizza. Please let us know if you have any questions.'"

[1049] In this way, it becomes possible to automatically generate responses that respond to the user's emotions and improve the user experience.

[1050] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1051] Step 1:

[1052] The user uses their device to perform voice input. For example, they might say, "I'd like to order a pizza (in a worried voice)." This voice is the input.

[1053] Step 2:

[1054] The device receives voice input from the user and converts the voice into text data using speech recognition technology. The process of converting voice data into text data is performed, and the output is the text data "I want to order a pizza."

[1055] Step 3:

[1056] Simultaneously, the device uses emotion analysis to extract emotion data from the user's voice. The emotion in the voice data is analyzed, and the emotion data "worry" is output.

[1057] Step 4:

[1058] The terminal sends the text data from step 2 and the sentiment data from step 3 to the server. The server then receives the user's inquiry and sentiment state.

[1059] Step 5:

[1060] The server passes the received text data to a natural language processing engine, which then categorizes the query into a specific category. In this case, the text data "I want to order a pizza" is classified under the "Food Category."

[1061] Step 6:

[1062] The server uses a model selection mechanism to select an appropriate generative artificial intelligence model based on the analysis results of the text data. In this case, a generative artificial intelligence model specifically for food delivery is selected.

[1063] Step 7:

[1064] The server uses emotion analysis tools to re-analyze the user's emotional data and adjusts its response accordingly. For example, based on the emotional data of "worry," it provides a prompt to the AI ​​model to generate a reassuring response.

[1065] Step 8:

[1066] The server uses a selected generative artificial intelligence model to generate a tailored response. Here, the text response "Thank you for ordering pizza. Please let us know if you have any questions." is generated.

[1067] Step 9:

[1068] The server formats the generated response text appropriately and converts it into audio data as needed. This stage includes the process of converting text data to audio data.

[1069] Step 10:

[1070] The server sends the generated response data to the terminal. This data includes responses in both text and audio formats.

[1071] Step 11:

[1072] The terminal receives response data from the server and provides it to the user using a response output mechanism. Specifically, it plays a voice message saying, "Thank you for ordering pizza. If you have any questions, we will be happy to assist you."

[1073] This series of steps is expected to improve the user experience of food delivery services by allowing users to receive appropriate responses tailored to their emotional state.

[1074] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1075] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1076] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[1077] [Third Embodiment]

[1078] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[1079] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1080] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1081] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[1082] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1083] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1084] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1085] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1086] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1087] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1088] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1089] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[1090] This invention relates to a system for responding quickly and accurately to a variety of user inquiries, and combines speech recognition technology, natural language processing technology, and generative artificial intelligence.

[1091] System-wide configuration

[1092] This system consists of a user, a terminal, and a server. The user makes a query through the terminal, which converts it into text data using speech recognition. The server then receives the text, analyzes its content using analysis tools, selects the optimal model to generate a response, and returns it to the terminal for the user to receive.

[1093] Server operation

[1094] 1. Inquiry reception

[1095] The server receives text data sent from the terminal.

[1096] The received text is passed to the analysis device, and the analysis begins.

[1097] 2. Analysis and Categorization

[1098] The server uses parsing tools to analyze the context of the text data and classify it into specific categories.

[1099] Based on the classified categories, an appropriate generative artificial intelligence model (e.g., fortune-telling, elderly care, housing, etc.) is selected.

[1100] 3. Response generation

[1101] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection method.

[1102] The response generated using the response generation means is formatted into an appropriate format and converted into audio data as needed.

[1103] 4. Send a response

[1104] The server sends the generated response to the terminal.

[1105] Terminal operation

[1106] 1. Speech recognition

[1107] The device receives voice input from the user.

[1108] The system uses speech recognition to convert speech into text data and then sends that text data to a server.

[1109] 2. Response reception and output

[1110] The terminal receives the response data sent from the server.

[1111] The system uses a response output mechanism to provide the user with the received data as text or audio.

[1112] User actions

[1113] 1. Inquiry

[1114] Users ask questions using a device (such as a public telephone, smartphone, or tablet).

[1115] By asking questions using voice, information can be easily obtained.

[1116] 2. Confirmation of response

[1117] The user checks the response provided by the device.

[1118] You can ask further questions or inquire about additional relevant information as needed.

[1119] Specific usage examples

[1120] Fortune-telling inquiries

[1121] User: "Please tell me my fortune for this month."

[1122] Terminal: Converts audio into text data and sends it to the server.

[1123] Server: Using analysis tools, the text "Please tell me my fortune for this month" is classified into the fortune-telling category, and a generative artificial intelligence model specifically for fortune-telling is selected.

[1124] Server: The fortune-telling model generates a response such as "Your fortune this month is good" and sends it to the terminal.

[1125] Terminal: Conveys received responses to the user via voice.

[1126] Caregiving inquiries

[1127] User: "I'm looking for a senior living facility."

[1128] Terminal: Converts audio into text data and sends it to the server.

[1129] Server: Using analysis tools, the text "I am looking for a facility for the elderly" is classified into the care category, and a generative artificial intelligence model specifically for care is selected.

[1130] Server: The caregiving model generates a response such as "The recommended elderly care facility is XX," and sends it to the terminal.

[1131] Terminal: Conveys received responses to the user via text or voice.

[1132] This embodiment of the present invention enables rapid response to diverse user needs and improves user satisfaction. The system provides advanced analysis and appropriate responses to various inquiries, thus enabling it to meet a wide range of user needs.

[1133] The following describes the processing flow.

[1134] Step 1:

[1135] The user speaks a specific question into the device's microphone (e.g., "Please tell me my fortune for this month").

[1136] Step 2:

[1137] The device uses speech recognition to convert the user's voice into text data.

[1138] Send the converted text data to the server.

[1139] Step 3:

[1140] The server receives text data sent from the terminal.

[1141] The received text data is passed to an analysis tool, which uses a natural language processing engine to analyze the context of the text.

[1142] Step 4:

[1143] The server uses analysis tools to classify text data into specific categories (e.g., fortune-telling, elderly care, housing).

[1144] Based on the analysis results, an appropriate generative artificial intelligence model will be selected.

[1145] Step 5:

[1146] The server generates an appropriate response to the query using a selected generative artificial intelligence model.

[1147] The generated response is formatted into an appropriate format using a response generation means.

[1148] Step 6:

[1149] The server sends the generated response data to the terminal.

[1150] Step 7:

[1151] The terminal receives the response data sent from the server.

[1152] The system uses a response output mechanism to provide the user with the received data in text or audio format.

[1153] Step 8:

[1154] The user checks the response provided by the device.

[1155] Ask additional questions as needed, then return to step 1.

[1156] (Example 1)

[1157] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1158] Traditional systems struggled to provide quick and accurate responses to user inquiries. Furthermore, generating responses that fully considered the context and history of the inquiry was difficult, making improving user satisfaction a challenge.

[1159] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1160] In this invention, the server includes: speech recognition means for receiving voice input from a user and converting the voice into text; analysis means for analyzing the text data using a natural language engine and classifying the inquiry into specific categories; model selection means for selecting a generative artificial intelligence engine that generates an appropriate response based on the analysis results; response generation means for generating a response to the inquiry using the selected generative artificial intelligence engine; and response output means for providing the generated response data to the user as text or voice. This enables the rapid and accurate provision of responses to user inquiries and the generation of sophisticated responses that take into account context and history.

[1161] A "speech recognition means" is a technological device that receives speech input from a user and converts that speech into text data.

[1162] A "natural language engine" is a technology that analyzes text data, understands its context and meaning, and classifies the content of inquiries into specific categories.

[1163] "Analysis means" refers to a system component that uses a natural language engine to analyze the context of text data and classify the query content into a specific category.

[1164] A "generative artificial intelligence engine" is an artificial intelligence technology component that generates appropriate responses based on a specific category.

[1165] A "model selection means" is a system component that selects an appropriate generative artificial intelligence engine based on the analysis results.

[1166] "Response generation means" refers to a technology that generates responses to inquiries using a selected generative artificial intelligence engine.

[1167] A "response output means" is a system component that provides the generated response data to the user as text or audio.

[1168] "History analysis means" refers to a technology that stores the user's inquiry history and uses that history data for analysis to generate a response that takes contextual information into account.

[1169] "Communication means" refers to the technology by which a terminal and a server send and receive data via a communication path.

[1170] This invention relates to a system that responds quickly and accurately to a variety of user inquiries, and is comprised of a combination of advanced speech recognition technology, natural language processing technology, and a generative artificial intelligence engine.

[1171] System-wide configuration

[1172] This system primarily consists of three elements: user, terminal, and server. The user makes inquiries via voice through the terminal, which converts them into text data using speech recognition technology. The text data is sent to the server, which analyzes the received text and uses an optimal generative artificial intelligence engine to provide a response to the user.

[1173] Server operation

[1174] 1. Inquiry reception

[1175] The server receives text data sent from the terminal.

[1176] The received text is passed to the analysis device, and the analysis begins.

[1177] 2. Analysis and Categorization

[1178] The server uses analysis tools to analyze the context of the text data and classify it into specific categories.

[1179] Specifically, a natural language engine is used to classify the data into categories such as "fortune telling," "elderly care," and "housing."

[1180] 3. Model Selection

[1181] The server selects an appropriate generative artificial intelligence engine (e.g., GPT-3, DALL-E, etc.) based on the analysis results.

[1182] For example, if a question like "Please tell me my fortune for this month" is classified under the fortune-telling category, a generative artificial intelligence engine specifically for fortune-telling will be selected.

[1183] 4. Response generation

[1184] The server inputs a prompt (e.g., "Please tell me my fortune for this month") into the selected generative artificial intelligence engine and generates a response.

[1185] Using a response generation tool, specific answers such as "Your luck this month is good" are created.

[1186] 5. Send a response

[1187] The server formats the generated response into the appropriate format and converts it into audio data if necessary.

[1188] Send the response to the terminal.

[1189] Terminal operation

[1190] 1. Speech recognition

[1191] The device receives voice input from the user.

[1192] Speech-to-Text API is used to convert speech into text data.

[1193] Specifically, the audio "Please tell me my fortune for this month" is converted into the text format "Please tell me my fortune for this month".

[1194] 2. Response reception and output

[1195] The terminal receives the response data sent from the server.

[1196] The system uses a response output mechanism to provide the user with the received data as text or audio.

[1197] For example, the system could tell the user via voice message, "Your fortune this month is good."

[1198] User actions

[1199] 1. Inquiry

[1200] Users ask questions using their devices (such as public telephones, smartphones, or tablets) via voice.

[1201] For example, the device receives voice input when the user says, "I'm looking for a senior care facility."

[1202] 2. Confirmation of response

[1203] The user checks the response provided by the device.

[1204] You can ask further questions or inquire about additional relevant information as needed.

[1205] Specific usage examples

[1206] Fortune-telling inquiries

[1207] User: "Please tell me my fortune for this month."

[1208] Terminal: Converts audio into text data and sends it to the server.

[1209] Server: Classifies text data into fortune-telling categories and selects a generative artificial intelligence engine specifically for fortune-telling.

[1210] Server: Uses a fortune-telling engine to generate a response such as "Your fortune this month is good" and sends it to the terminal.

[1211] Terminal: Communicates responses to the user via voice.

[1212] Caregiving inquiries

[1213] User: "I'm looking for a senior living facility."

[1214] Terminal: Converts audio into text data and sends it to the server.

[1215] Server: Classifies text data into caregiving categories and selects a generative artificial intelligence engine specifically for caregiving.

[1216] Server: Uses an engine designed for elderly care to generate responses such as "We recommend XX elderly care facility" and send them to the terminal.

[1217] Terminal: Communicates responses to the user via voice.

[1218] This system enables quick and accurate responses to a wide range of user inquiries, thereby improving user satisfaction. Furthermore, by taking past inquiry history into account, it can provide more personalized responses.

[1219] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1220] Program execution flow: Explained in steps.

[1221] Step 1: User voice input

[1222] Users make inquiries to their devices (such as smartphones and tablets) using voice commands.

[1223] For example, a user might say, "Please tell me my fortune for this month."

[1224] Input: User's voice data

[1225] Output: Audio data is input to the terminal.

[1226] Step 2: Voice recognition on the device

[1227] The device uses speech recognition technology (e.g., Google Speech-to-Text API) to convert the user's voice into text data.

[1228] The system processes the audio data and generates the text, "Please tell me my fortune for this month."

[1229] Input: User's voice data

[1230] Data processing: Converting audio data to text data.

[1231] Output: Text data "Please tell me my fortune for this month"

[1232] Step 3: Send text data

[1233] The terminal sends the converted text data to the server. HTTP or similar communication protocols are used.

[1234] The text data is sent to the server and prepared for analysis.

[1235] Input: Text data "Please tell me my fortune for this month"

[1236] Data processing: Packeting and preparing text data for transmission.

[1237] Output: Text data sent to the server

[1238] Step 4: Text message received

[1239] The server receives text data sent from the terminal.

[1240] The received text data is passed to the analysis tool.

[1241] Input: Text data sent from the device

[1242] Data processing: Preparing for analysis of received data.

[1243] Output: Text data to be analyzed

[1244] Step 5: Analyze the inquiry

[1245] The server uses a natural language engine to analyze the context of the text data and then parses the query content.

[1246] For example, the text "Please tell me my fortune for this month" would be classified under the fortune-telling category.

[1247] Input: Text data to be analyzed

[1248] Data processing: Categorization using natural language processing

[1249] Output: Category information (e.g., "Fortune Telling")

[1250] Step 6: Model Selection

[1251] The server selects an appropriate generative artificial intelligence engine based on the analysis results.

[1252] For example, if it is classified as a fortune-telling category, a generative artificial intelligence engine specifically for fortune-telling (e.g., a GPT-3 model for fortune-telling) will be selected.

[1253] Input: Category information (e.g., "Fortune Telling")

[1254] Data processing: Selection of an appropriate AI model

[1255] Output: Selected generative artificial intelligence engine

[1256] Step 7: Generating the response

[1257] The server inputs a prompt (e.g., "Please tell me my fortune for this month") into the selected generative artificial intelligence engine and generates an appropriate response.

[1258] For example, a response such as "Your luck this month is good" is generated.

[1259] Input: Selected generative artificial intelligence engine, prompt

[1260] Data processing: Response generation by generative AI

[1261] Output: Generated response data (e.g., "Your fortune this month is good")

[1262] Step 8: Sending a response

[1263] The server formats the generated response into the appropriate format and converts it into audio data as needed.

[1264] The response data is sent to the terminal.

[1265] Input: Generated response data

[1266] Data processing: Formatting of response data and conversion to audio data (if necessary).

[1267] Output: Response data ready to be sent to the terminal

[1268] Step 9: Receive response / correction

[1269] The terminal receives the response data sent from the server.

[1270] Input: Response data sent from the server

[1271] Data processing: Processing of received data

[1272] Output: Response data in a state that can be provided to the user.

[1273] Step 10: Provide a response to the user

[1274] The terminal uses a response output means to provide the user with the received data as text or audio.

[1275] For example, the system could tell the user via voice message, "Your fortune this month is good."

[1276] Input: Response data in a state that can be provided to the user.

[1277] Data calculation: Output of response data

[1278] Output: The response provided to the user (text or audio)

[1279] This processing flow enables quick and accurate responses to a wide range of user inquiries.

[1280] (Application Example 1)

[1281] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1282] In autonomous vehicles, it is crucial for users to quickly and accurately obtain various information while driving. However, existing systems often lack sufficient accuracy in speech recognition and natural language processing, sometimes providing only incorrect or incomplete information. Furthermore, their ability to appropriately acquire and respond to real-time updated information is insufficient, which can impair user convenience. The problem this invention aims to solve is to address these challenges and provide a system that allows users to always obtain the latest information appropriately.

[1283] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1284] In this invention, the server includes speech recognition means for converting speech to text, analysis means for analyzing text data and classifying the inquiry content into specific categories, and model selection means for selecting a generative artificial intelligence model based on the analysis results. This makes it possible to generate and provide a quick and accurate response when a user queries for information in real time by voice within an autonomous vehicle.

[1285] "Voice recognition means" refers to a device or software that converts voice data input by a user into text data.

[1286] "Analysis means" refers to a device or software that analyzes text data using a natural language processing engine and classifies the content of a query into a specific category.

[1287] "Model selection means" refers to a device or software that selects an appropriate generative artificial intelligence model based on the analysis results.

[1288] "Response generation means" refers to a device or software that generates a response to an inquiry using a selected generative artificial intelligence model.

[1289] "Response output means" refers to a device or software that provides the generated response to the user in text or voice.

[1290] "Communication means" refers to a device or software that connects to an online server via a communication system within a vehicle and transmits and receives data in real time.

[1291] "Voice output means" refers to a device or software that provides the user with a voice response generated using the acoustic equipment inside the vehicle.

[1292] "History analysis means" refers to a device or software that stores the user's inquiry history and uses that history data for analysis to generate a response that takes contextual information into account.

[1293] This invention provides a system that responds quickly and accurately to user inquiries made by voice within an autonomous vehicle. The following is a specific embodiment of this system.

[1294] Hardware and software configuration

[1295] Hardware configuration:

[1296] 1. Microphone: A device installed inside a vehicle to capture sound.

[1297] 2. In-vehicle communication system: This is a communication system that connects to the internet and sends and receives data with online servers.

[1298] 3. Speaker: A device for providing the generated voice response to the user.

[1299] Software configuration:

[1300] 1. Speech Recognition Method: The speech acquired from the microphone is converted into text data using the Python library speech_recognition.

[1301] 2. Analysis method: A natural language processing engine is used to analyze the context of the text data and classify the query content into specific categories.

[1302] 3. Model Selection Method: Based on the analysis results, an appropriate generative artificial intelligence model (generative AI model) is selected.

[1303] 4. Response generation means: A response to the query is generated using a selected generative AI model.

[1304] 5. Response output method: The generated response is converted into audio data using the Google Text-to-Speech (gTTS) library and provided to the user through the speaker.

[1305] Specific operation of the system

[1306] 1. Voice Recognition: The user asks a question into the microphone inside the car. For example, "Where is the nearest gas station?"

[1307] 2. Text conversion: The speech recognition system converts the user's speech into text data.

[1308] 3. Analysis: The analysis tool analyzes the text data using a natural language processing engine and classifies the sentence "Where is the nearest gas station?" into the traffic information category.

[1309] 4. Model Selection: The model selection method selects a generation AI model for traffic information.

[1310] 5. Response Generation: The response generation means uses a selected generative AI model to generate a response such as, "The nearest gas station is 2km ahead on your right."

[1311] 6. Response Output: The response output means converts the generated response into audio data and transmits it to the user through the speaker.

[1312] Specific example

[1313] For example, if a user asks "Where is the nearest gas station?" while inside an autonomous vehicle, the system will respond in the following order:

[1314] 1. The microphone captures audio data, and the speech recognition system converts it into text data such as "Where is the nearest gas station?".

[1315] 2. The analysis tool analyzes the text data and classifies it into traffic information categories.

[1316] 3. The model selection method selects an AI model for generating traffic information.

[1317] 4. The response generation means generates the response, "The nearest gas station is 2km ahead on your right."

[1318] 5. The response output means converts the response into audio data using the gTTS library and provides it to the user through the speaker.

[1319] Example of a prompt

[1320] Example of a user inquiry: "Where is the nearest gas station?"

[1321] Prompt text: "In response to the voice-recognized inquiry 'Where is the nearest gas station?', the system will provide the location of the best gas station."

[1322] Examples of input content for a generative AI model:

[1323] Inquiry: "Where is the nearest gas station?"

[1324] Reply: "The nearest gas station is 2km ahead on your right."

[1325] In this way, users can obtain quick and accurate information using voice commands within autonomous vehicles.

[1326] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1327] Step 1:

[1328] The user makes a voice inquiry into the microphone inside the vehicle. The input is the user's voice data, and the output is the voice data captured by the microphone.

[1329] Step 2:

[1330] The device's speech recognition system converts the voice data input by the user into text data. The input for this step is the captured voice data, and the output is the converted text data. Specifically, the Python library speech_recognition is used to analyze the voice data and convert it into string information.

[1331] Step 3:

[1332] The terminal sends the converted text data to the server via the in-vehicle communication system. The input for this step is the converted text data, and the output is the text data sent to the server. Communication means are used for sending and receiving data.

[1333] Step 4:

[1334] The server's analysis mechanism analyzes the received text data using a natural language processing engine. The input for this step is the received text data, and the output is the analyzed query content and its categorization. Natural language processing techniques are used for the analysis, understanding the context and classifying it into appropriate categories.

[1335] Step 5:

[1336] The server's model selection mechanism selects an appropriate generative AI model based on the analyzed results. The input to this step is the analysis results and their category classification, and the output is the selected generative AI model. The selection mechanism determines the optimal generative AI model for each query category.

[1337] Step 6:

[1338] The server's response generation mechanism generates an appropriate response using a selected generative AI model. The inputs to this step are the selected generative AI model and the analyzed query, while the output is the generated response data. This includes the process by which the generative AI model generates the optimal answer to the query.

[1339] Step 7:

[1340] The server sends the generated response data to the terminal. The input to this step is the generated response data, and the output is the response data sent to the terminal. Communication means are used for sending and receiving data.

[1341] Step 8:

[1342] The terminal's response output mechanism converts the received response data into speech data. The input for this step is the received response data, and the output is the converted speech data. Specifically, the Google Text-to-Speech (gTTS) library is used to convert the text data into speech data.

[1343] Step 9:

[1344] The device's speaker plays the converted audio data to the user. The input in this step is the converted audio data, and the output is the provision of audio information to the user. Through the speaker, the user can hear the generated response.

[1345] This series of processes will enable users to obtain quick and accurate information using voice commands within autonomous vehicles.

[1346] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1347] This invention relates to a system that recognizes a user's emotions and generates a response based on those emotions, and combines speech recognition technology, natural language processing technology, generative artificial intelligence, and an emotion engine.

[1348] System-wide configuration

[1349] This system consists of a user, a terminal, and a server. The user makes inquiries through the terminal, which converts them into text data using speech recognition. The server then receives the text data and speech sentiment data, analyzes them using an analysis tool, selects an appropriate model to generate a response, and returns it to the terminal for the user to receive.

[1350] Server operation

[1351] 1. Inquiry reception

[1352] The server receives text data and sentiment data sent from the terminal.

[1353] The received data is passed to the analysis device, and the analysis begins.

[1354] 2. Analysis and categorization of text data

[1355] The server uses parsing tools to analyze the context of text data and classify it into specific categories (e.g., fortune-telling, elderly care, housing).

[1356] Based on the classified categories, an appropriate generative artificial intelligence model is selected.

[1357] 3. Analysis of emotional data

[1358] The server uses an emotion engine to analyze emotional data obtained from the voice and recognize the user's emotional state.

[1359] The recognized emotion data is combined with the query content to adjust the content and tone of the response.

[1360] 4. Response generation

[1361] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection method.

[1362] Using a response generation means, the generated response is formatted into an appropriate format and converted into audio data as needed.

[1363] 5. Send a response

[1364] The server sends the generated response data to the terminal.

[1365] Terminal operation

[1366] 1. Speech recognition

[1367] The device receives voice input from the user.

[1368] The system uses speech recognition to convert speech into text data and then sends that text data to a server.

[1369] Simultaneously, an emotion engine is used to extract emotion data from the audio and send it to the server.

[1370] 2. Response reception and output

[1371] The terminal receives the response data sent from the server.

[1372] The system uses a response output mechanism to provide the user with the received data as text or audio.

[1373] User actions

[1374] 1. Inquiry

[1375] Users ask questions using a device (such as a public telephone, smartphone, or tablet).

[1376] By asking questions using voice, information can be easily obtained.

[1377] 2. Confirmation of response

[1378] The user checks the response provided by the device.

[1379] You can ask further questions or inquire about additional relevant information as needed.

[1380] Specific usage examples

[1381] Fortune-telling inquiries

[1382] User: "Please tell me my fortune for this month (in an excited voice)."

[1383] Terminal: Converts audio into text data, extracts emotional data from the audio, and sends it to the server.

[1384] Server: Using analysis tools, the text "Please tell me my fortune for this month" is classified into the fortune-telling category, and a generative artificial intelligence model specifically for fortune-telling is selected.

[1385] Server: Uses an emotion engine to analyze emotional data and recognizes the state of being "excited."

[1386] Server: Based on the recognized emotion data, it generates a response such as, "Your luck this month is good, especially your career luck!" and sends it to the terminal.

[1387] Terminal: Conveys received responses to the user via voice.

[1388] Caregiving inquiries

[1389] User: "I'm looking for a senior living facility (in a tired voice)."

[1390] Terminal: Converts audio into text data, extracts emotional data from the audio, and sends it to the server.

[1391] Server: Using analysis tools, the text "I am looking for a facility for the elderly" is classified into the care category, and a generative artificial intelligence model specifically for care is selected.

[1392] Server: Uses an emotion engine to analyze emotional data and recognize the state of being "tired".

[1393] Server: Based on recognized emotion data, it generates a response such as, "We recommend XX senior care facility. You seem tired, rest is important," and sends it to the terminal.

[1394] Terminal: Conveys received responses to the user via voice.

[1395] In this embodiment of the present invention, it is possible to recognize the user's emotional state and provide an appropriate response accordingly, thereby greatly improving user satisfaction. The system can combine advanced analysis and emotion recognition for various inquiries, enabling it to provide users with more personalized services.

[1396] The following describes the processing flow.

[1397] Step 1:

[1398] The user speaks a specific question into the device's microphone (e.g., "Please tell me my fortune for this month").

[1399] Step 2:

[1400] The device uses speech recognition to convert the user's voice into text data.

[1401] Send the converted text data to the server.

[1402] Step 3:

[1403] The device uses an emotion engine to extract emotional data from the audio.

[1404] The extracted emotion data is sent to the server.

[1405] Step 4:

[1406] The server receives text data and sentiment data sent from the terminal.

[1407] Step 5:

[1408] The server passes the text data to the analysis tool, which then uses a natural language processing engine to analyze the context.

[1409] Step 6:

[1410] The server uses analysis tools to classify text data into specific categories (e.g., fortune-telling, elderly care, housing).

[1411] Step 7:

[1412] The server selects an appropriate generative artificial intelligence model based on the analysis results.

[1413] Step 8:

[1414] The server uses an emotion engine to analyze emotional data and recognize the user's emotional state.

[1415] Step 9:

[1416] The server combines the recognized emotion data with the query content to adjust the content and tone of the response.

[1417] Step 10:

[1418] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection method.

[1419] Step 11:

[1420] The server formats the response generated using the response generation means into an appropriate format and converts it into audio data as needed.

[1421] Step 12:

[1422] The server sends the generated response data to the terminal.

[1423] Step 13:

[1424] The terminal receives the response data sent from the server.

[1425] Step 14:

[1426] The terminal uses a response output mechanism to provide the user with the received data in text or voice.

[1427] Step 15:

[1428] The user checks the response provided by the device.

[1429] Ask additional questions as needed, then return to step 1.

[1430] (Example 2)

[1431] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1432] Conventional speech recognition systems have been immature in generating responses that take user emotions into account, making it difficult to enhance user satisfaction. Furthermore, they often lack accuracy in analyzing inquiries and the appropriateness of their responses, particularly in providing personalized responses that reflect emotions. As a result, the user experience deteriorates, limiting the effectiveness of the system.

[1433] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1434] In this invention, the server includes emotion analysis means for extracting emotion data from speech, analysis means for analyzing text data using a natural language processing engine and classifying the inquiry content into specific categories, and model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on the analysis results. This makes it possible to generate personalized responses that take the user's emotions into consideration.

[1435] "Voice recognition means" refers to a technology or device for converting a user's voice input into text data.

[1436] "Emotional analysis means" refers to a technology or device for extracting emotional data from voice input and determining the user's emotional state.

[1437] A "natural language processing engine" is a technology or software that analyzes text data to understand its context and meaning.

[1438] "Analysis means" refers to a technique or device for classifying text data into specific categories.

[1439] A "generative artificial intelligence model" is an artificial intelligence model designed to generate appropriate responses based on input data.

[1440] "Model selection means" refers to a technique or device for selecting an appropriate generative artificial intelligence model based on analysis results.

[1441] "Response generation means" refers to a technology or apparatus that generates a response to a user's inquiry using a selected generative artificial intelligence model.

[1442] "Response output means" refers to a technology or device that provides the generated response to the user as text or audio.

[1443] "History analysis means" refers to a technology or device that analyzes user inquiry history data and generates a response that takes contextual information into account.

[1444] "Communication means" refers to the technology or device used by a terminal and a server to send and receive data via a communication line.

[1445] "Distributed processing means" refers to a technology or device that distributes processing to efficiently perform data analysis and response generation within a server.

[1446] This invention relates to a system that recognizes a user's emotions and generates a response based on those emotions. Specifically, it is a system that combines speech recognition technology, natural language processing technology, generative artificial intelligence, and emotion analysis technology.

[1447] System-wide configuration

[1448] This system consists of a user, a terminal, and a server. The user makes inquiries through the terminal, which converts them into text data using speech recognition. The server then receives the text data and speech emotion data, analyzes them using an analysis tool, selects an appropriate generative artificial intelligence model to generate a response, and returns it to the terminal for the user to receive.

[1449] Hardware and software used

[1450] Speech recognition means: The terminal uses a microphone to receive voice input from the user and a speech recognition service such as the Google Speech-to-Text API.

[1451] Sentiment analysis methods: IBM Watson Tone Analyzer and Microsoft Azure Emotion API are used for sentiment analysis.

[1452] Natural Language Processing Engine: Uses natural language processing libraries such as the BERT model, spaCy, and NLTK.

[1453] Generative AI Models: Advanced generative AI models such as GPT-4 are used.

[1454] Communication method: The terminal and server use the HTTP protocol over the internet to send and receive data.

[1455] Distributed processing methods: Cloud infrastructure (e.g., AWS, Google Cloud) is used to efficiently perform data analysis and response generation within the server.

[1456] Specific usage examples

[1457] Fortune-telling inquiries

[1458] Example: The user asks a voice question, "Please tell me my fortune for this month (in an excited voice)."

[1459] Terminal: Converts the audio into text data and sends the text "Please tell me my fortune for this month" to the server. At the same time, it extracts the emotion data "excited" from the audio and also sends this to the server.

[1460] Server: Uses analysis tools to classify text data into the "fortune-telling" category and selects a generative artificial intelligence model specifically for fortune-telling. Uses emotion analysis technology to recognize when the user is in an "excited" state.

[1461] Server: Based on recognized sentiment data and text data, it generates the response, "Your luck this month is good, especially your career luck!"

[1462] Terminal: Converts the generated response into speech data using speech synthesis technology and provides it to the user.

[1463] Caregiving inquiries

[1464] Example: The user asks a question via voice, "I'm looking for a senior living facility (in a tired voice)."

[1465] Terminal: Converts the voice into text data and sends the text "I am looking for a senior care facility" to the server. At the same time, it extracts emotion data, "I am tired," and also sends this to the server.

[1466] Server: Uses analysis tools to classify text data into the "caregiving" category and selects a generative artificial intelligence model specifically for caregiving. Uses emotion analysis technology to recognize that the user is in a "tired" state.

[1467] Server: Based on recognized sentiment data and text data, it generates a response such as, "We recommend XX senior living facility. You seem tired; rest is important."

[1468] Terminal: Converts the generated response into speech data using speech synthesis technology and provides it to the user.

[1469] This invention makes it possible to recognize a user's emotional state and provide an appropriate response accordingly. The system can combine advanced analysis and emotion recognition to provide users with more personalized services in response to various inquiries.

[1470] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1471] Step 1:

[1472] Receiving and preprocessing voice input

[1473] The device receives voice input from the user. For example, it captures voice data such as "Please tell me my fortune for this month" through the smartphone's microphone.

[1474] Input: User's voice data (e.g., "Please tell me my fortune for this month.")

[1475] Specific operation: The device uses noise cancellation technology to improve the quality of audio data.

[1476] Output: Clear audio data with noise removed.

[1477] Step 2:

[1478] Speech recognition and text conversion

[1479] The device converts speech data into text data using speech recognition technology. Specifically, it uses the Google Speech-to-Text API for speech recognition.

[1480] Input: Clear audio data (e.g., "Please tell me my fortune for this month")

[1481] Specific operation: Send audio data to the API and retrieve text data as a response.

[1482] Output: Text data (Example: "Please tell me my fortune for this month")

[1483] Step 3:

[1484] Emotion analysis

[1485] The terminal uses sentiment analysis tools to extract sentiment data from voice data. In this case, IBM Watson Tone Analyzer is used.

[1486] Input: Clear audio data (e.g., "Please tell me my fortune for this month")

[1487] Specific operation: Voice data is sent to the emotion analysis engine, and emotion data is obtained as a result of the analysis.

[1488] Output: Sentiment data (e.g., "excited")

[1489] Step 4:

[1490] Sending data

[1491] The device sends text data and sentiment data to the server.

[1492] Input: Text data and sentiment data (e.g., "Please tell me my horoscope for this month," "I'm excited")

[1493] Specific operation: Send data to the server using an HTTP request.

[1494] Output: The server receives text data and sentiment data.

[1495] Step 5:

[1496] Text data analysis and categorization

[1497] The server analyzes the context of text data using analytical tools and classifies it into specific categories (e.g., fortune-telling, elderly care, housing). Here, the BERT model and spaCy are used.

[1498] Input: Text data (Example: "Please tell me my fortune for this month.")

[1499] Specific operation: Text data is fed into a natural language processing engine, and categories are determined based on the analysis results.

[1500] Output: Category information (e.g., "Fortune Telling")

[1501] Step 6:

[1502] Category-based model selection

[1503] The server selects an appropriate generative artificial intelligence model based on the analysis results. In this case, a generative artificial intelligence model such as GPT-4 is selected for the fortune-telling category.

[1504] Input: Category information (e.g., "Fortune Telling")

[1505] Specific operation: Use the model selection method within the server to select a relevant artificial intelligence model.

[1506] Output: Generative artificial intelligence model (e.g., GPT-4)

[1507] Step 7:

[1508] Response generation

[1509] The server uses a selected generative artificial intelligence model to generate responses based on text data and sentiment data. Examples of prompts include "Please tell me my fortune for this month" and the state of being "excited."

[1510] Input: Text data, sentiment data, generative artificial intelligence model (e.g., "Please tell me my fortune for this month," "I'm excited")

[1511] Specific operation: A prompt is fed into a generative artificial intelligence model, and the generated response is retrieved.

[1512] Output: Generated response data (Example: "Your luck this month is good, especially your career luck!")

[1513] Step 8:

[1514] Response formatting and speech conversion

[1515] The server formats the generated response into an appropriate format (e.g., JSON) and uses speech synthesis technology to convert it into audio data. In this case, the Google Text-to-Speech API is used.

[1516] Input: Generated response data (Example: "Your luck this month is good, especially your career luck!")

[1517] Specific operation: Sends a text response to the speech synthesis engine and generates speech data.

[1518] Output: Audio data or formatted text data

[1519] Step 9:

[1520] Sending a response

[1521] The server sends the generated response data to the terminal.

[1522] Input: Voice data or formatted text data (e.g., "Your luck this month is good, especially your career luck!")

[1523] Specific operation: Send data to the terminal using an HTTP response.

[1524] Output: The terminal receives the response data.

[1525] Step 10:

[1526] Providing a response

[1527] The terminal receives response data sent from the server and provides it to the user.

[1528] Input: Voice data or formatted text data (e.g., "Your luck this month is good, especially your career luck!")

[1529] Specific operation: If the data is audio, it will be played as sound from the speaker; if the data is text, it will be displayed on the screen.

[1530] Output: Provides a response to the user.

[1531] The above outlines the specific program processing flow of this system. By describing the inputs and outputs at each step in detail, the overall operation of the system is made easier to understand.

[1532] (Application Example 2)

[1533] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1534] In modern food delivery services, users experience a range of emotions when placing an order, and a response tailored to those emotions is required. However, traditional systems lacked the ability to analyze user emotions and could only provide uniform responses, resulting in a poor user experience. Furthermore, there was a lack of technology to generate appropriate responses that considered context and emotions in response to user inquiries. As a result, gaining user trust was difficult, and improving the quality of service was essential.

[1535] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes: speech recognition means that receive voice input from the user and convert the voice into text; analysis means that analyze the text data using a natural language processing engine and classify the inquiry into specific categories; model selection means that select a generative artificial intelligence model that generates an appropriate response based on the analysis results; emotion analysis means that extract emotion data from the user's voice and adjust the response content based on that emotion; response generation means that generate a response to the inquiry using the selected generative artificial intelligence model; and response output means that provide the generated response to the user in text or voice. This makes it possible to provide an appropriate response according to the user's emotions and improve the user experience.

[1536] "A speech recognition means that receives voice input from a user and converts that voice into text" refers to a technology or device that converts a user's voice into digital data and generates text data from the voice.

[1537] "Analysis means for analyzing text data using a natural language processing engine and classifying query content into specific categories" refers to a technology or apparatus that performs a process of analyzing generated text data, identifying its content, and classifying it into specific categories.

[1538] "Model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on analysis results" refers to a technology or device that selects the most appropriate generative artificial intelligence model based on the analyzed data.

[1539] "Response generation means for generating a response to a query using a selected generative artificial intelligence model" refers to a technology or apparatus that uses a selected generative artificial intelligence model to generate an appropriate response to a query.

[1540] "An emotion analysis means that extracts emotion data from a user's voice and adjusts the response content based on that emotion" refers to a technology or device that analyzes a user's voice information to recognize their emotional state and adjusts the response content based on that emotional state.

[1541] "Response output means that provides the generated response to the user in text or audio" refers to a technology or device that conveys the generated response to the user in text or audio format.

[1542] This invention is a system that combines speech recognition technology, natural language processing technology, generative artificial intelligence models, and an emotion engine to provide customizable responses tailored to the user's emotions when using a food delivery service. Specific embodiments are described below.

[1543] System-wide configuration

[1544] The system of the present invention consists of a user, a terminal, and a server. The user provides voice input through the terminal, which converts it into text data using speech recognition means. Subsequently, the server receives the text data and sentiment data, analyzes them using analysis means, selects an appropriate generative artificial intelligence model to generate a response, and returns it to the terminal for the user to receive.

[1545] Server operation

[1546] Inquiry reception:

[1547] The server receives text data and sentiment data sent from the terminal. It then passes the received data to the analysis system and begins the analysis.

[1548] Text data analysis and categorization:

[1549] The server uses analysis tools to analyze the context of text data and classify it into specific categories (e.g., food categories). Based on the classified categories, it selects an appropriate generative artificial intelligence model.

[1550] Analysis of emotional data:

[1551] The server uses an emotion engine to analyze emotional data obtained from the voice and recognize the user's emotional state. Based on the recognized emotion, it adjusts the content and tone of the response.

[1552] Response generation:

[1553] The server generates an appropriate response using a generative artificial intelligence model selected by the model selection means. The response generation means then formats the generated response into an appropriate format and converts it into audio data as needed.

[1554] Send response:

[1555] The server sends the generated response data to the terminal.

[1556] Terminal operation

[1557] Speech recognition:

[1558] The terminal receives voice input from the user, converts the voice into text data using speech recognition technology, and sends that text data to the server. Simultaneously, it extracts emotion data from the voice using an emotion engine and sends it to the server.

[1559] Response reception and output:

[1560] The terminal receives response data sent from the server and uses a response output means to provide the received data to the user as text or audio.

[1561] User actions

[1562] inquiry:

[1563] The user asks questions using their device via voice. Example: "I'd like to order a pizza (in a worried voice)."

[1564] Response confirmation:

[1565] The user confirms the response provided by the device. Example: "Thank you for ordering a pizza. Please let us know if you have any questions."

[1566] Core technology

[1567] This system is built using the following technologies:

[1568] Speech recognition technology: Uses the speech_recognition library.

[1569] Natural language processing technology: Uses the natural_language_processing engine.

[1570] Generative artificial intelligence models: Generative AI models

[1571] Emotion engine: Uses emotion_recognition technology

[1572] Specific example

[1573] For example, if a user types "I'd like to order a pizza (in a worried voice)," the server converts the speech into text data and recognizes the emotion of "worry." Based on the recognized emotion, it generates a response such as "Thank you for ordering a pizza. Please let us know if you have any questions," and provides this to the user in voice.

[1574] Example of a prompt

[1575] Examples of prompts for a generative AI model include the following:

[1576] "Generate a reassuring response when the user is worried. For example, if the user's voice input is 'I'd like to order a pizza (in a worried voice),' create a response like, 'Thank you for ordering a pizza. Please let us know if you have any questions.'"

[1577] In this way, it becomes possible to automatically generate responses that respond to the user's emotions and improve the user experience.

[1578] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1579] Step 1:

[1580] The user uses their device to perform voice input. For example, they might say, "I'd like to order a pizza (in a worried voice)." This voice is the input.

[1581] Step 2:

[1582] The device receives voice input from the user and converts the voice into text data using speech recognition technology. The process of converting voice data into text data is performed, and the output is the text data "I want to order a pizza."

[1583] Step 3:

[1584] Simultaneously, the device uses emotion analysis to extract emotion data from the user's voice. The emotion in the voice data is analyzed, and the emotion data "worry" is output.

[1585] Step 4:

[1586] The terminal sends the text data from step 2 and the sentiment data from step 3 to the server. The server then receives the user's inquiry and sentiment state.

[1587] Step 5:

[1588] The server passes the received text data to a natural language processing engine, which then categorizes the query into a specific category. In this case, the text data "I want to order a pizza" is classified under the "Food Category."

[1589] Step 6:

[1590] The server uses a model selection mechanism to select an appropriate generative artificial intelligence model based on the analysis results of the text data. In this case, a generative artificial intelligence model specifically for food delivery is selected.

[1591] Step 7:

[1592] The server uses emotion analysis tools to re-analyze the user's emotional data and adjusts its response accordingly. For example, based on the emotional data of "worry," it provides a prompt to the AI ​​model to generate a reassuring response.

[1593] Step 8:

[1594] The server uses a selected generative artificial intelligence model to generate a tailored response. Here, the text response "Thank you for ordering pizza. Please let us know if you have any questions." is generated.

[1595] Step 9:

[1596] The server formats the generated response text appropriately and converts it into audio data as needed. This stage includes the process of converting text data to audio data.

[1597] Step 10:

[1598] The server sends the generated response data to the terminal. This data includes responses in both text and audio formats.

[1599] Step 11:

[1600] The terminal receives response data from the server and provides it to the user using a response output mechanism. Specifically, it plays a voice message saying, "Thank you for ordering pizza. If you have any questions, we will be happy to assist you."

[1601] This series of steps is expected to improve the user experience of food delivery services by allowing users to receive appropriate responses tailored to their emotional state.

[1602] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1603] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1604] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1605] [Fourth Embodiment]

[1606] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1607] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1608] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1609] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1610] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1611] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1612] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1613] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1614] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1615] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1616] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1617] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1618] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1619] This invention relates to a system for responding quickly and accurately to a variety of user inquiries, and combines speech recognition technology, natural language processing technology, and generative artificial intelligence.

[1620] System-wide configuration

[1621] This system consists of a user, a terminal, and a server. The user makes a query through the terminal, which converts it into text data using speech recognition. The server then receives the text, analyzes its content using analysis tools, selects the optimal model to generate a response, and returns it to the terminal for the user to receive.

[1622] Server operation

[1623] 1. Inquiry reception

[1624] The server receives text data sent from the terminal.

[1625] The received text is passed to the analysis device, and the analysis begins.

[1626] 2. Analysis and Categorization

[1627] The server uses parsing tools to analyze the context of the text data and classify it into specific categories.

[1628] Based on the classified categories, an appropriate generative artificial intelligence model (e.g., fortune-telling, elderly care, housing, etc.) is selected.

[1629] 3. Response generation

[1630] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection method.

[1631] The response generated using the response generation means is formatted into an appropriate format and converted into audio data as needed.

[1632] 4. Send a response

[1633] The server sends the generated response to the terminal.

[1634] Terminal operation

[1635] 1. Speech recognition

[1636] The device receives voice input from the user.

[1637] The system uses speech recognition to convert speech into text data and then sends that text data to a server.

[1638] 2. Response reception and output

[1639] The terminal receives the response data sent from the server.

[1640] The system uses a response output mechanism to provide the user with the received data as text or audio.

[1641] User actions

[1642] 1. Inquiry

[1643] Users ask questions using a device (such as a public telephone, smartphone, or tablet).

[1644] By asking questions using voice, information can be easily obtained.

[1645] 2. Confirmation of response

[1646] The user checks the response provided by the device.

[1647] You can ask further questions or inquire about additional relevant information as needed.

[1648] Specific usage examples

[1649] Fortune-telling inquiries

[1650] User: "Please tell me my fortune for this month."

[1651] Terminal: Converts audio into text data and sends it to the server.

[1652] Server: Using analysis tools, the text "Please tell me my fortune for this month" is classified into the fortune-telling category, and a generative artificial intelligence model specifically for fortune-telling is selected.

[1653] Server: The fortune-telling model generates a response such as "Your fortune this month is good" and sends it to the terminal.

[1654] Terminal: Conveys received responses to the user via voice.

[1655] Caregiving inquiries

[1656] User: "I'm looking for a senior living facility."

[1657] Terminal: Converts audio into text data and sends it to the server.

[1658] Server: Using analysis tools, the text "I am looking for a facility for the elderly" is classified into the care category, and a generative artificial intelligence model specifically for care is selected.

[1659] Server: The caregiving model generates a response such as "The recommended elderly care facility is XX," and sends it to the terminal.

[1660] Terminal: Conveys received responses to the user via text or voice.

[1661] This embodiment of the present invention enables rapid response to diverse user needs and improves user satisfaction. The system provides advanced analysis and appropriate responses to various inquiries, thus enabling it to meet a wide range of user needs.

[1662] The following describes the processing flow.

[1663] Step 1:

[1664] The user speaks a specific question into the device's microphone (e.g., "Please tell me my fortune for this month").

[1665] Step 2:

[1666] The device uses speech recognition to convert the user's voice into text data.

[1667] Send the converted text data to the server.

[1668] Step 3:

[1669] The server receives text data sent from the terminal.

[1670] The received text data is passed to an analysis tool, which uses a natural language processing engine to analyze the context of the text.

[1671] Step 4:

[1672] The server uses analysis tools to classify text data into specific categories (e.g., fortune-telling, elderly care, housing).

[1673] Based on the analysis results, an appropriate generative artificial intelligence model will be selected.

[1674] Step 5:

[1675] The server generates an appropriate response to the query using a selected generative artificial intelligence model.

[1676] The generated response is formatted into an appropriate format using a response generation means.

[1677] Step 6:

[1678] The server sends the generated response data to the terminal.

[1679] Step 7:

[1680] The terminal receives the response data sent from the server.

[1681] The system uses a response output mechanism to provide the user with the received data in text or audio format.

[1682] Step 8:

[1683] The user checks the response provided by the device.

[1684] Ask additional questions as needed, then return to step 1.

[1685] (Example 1)

[1686] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1687] Traditional systems struggled to provide quick and accurate responses to user inquiries. Furthermore, generating responses that fully considered the context and history of the inquiry was difficult, making improving user satisfaction a challenge.

[1688] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1689] In this invention, the server includes: speech recognition means for receiving voice input from a user and converting the voice into text; analysis means for analyzing the text data using a natural language engine and classifying the inquiry into specific categories; model selection means for selecting a generative artificial intelligence engine that generates an appropriate response based on the analysis results; response generation means for generating a response to the inquiry using the selected generative artificial intelligence engine; and response output means for providing the generated response data to the user as text or voice. This enables the rapid and accurate provision of responses to user inquiries and the generation of sophisticated responses that take into account context and history.

[1690] A "speech recognition means" is a technological device that receives speech input from a user and converts that speech into text data.

[1691] A "natural language engine" is a technology that analyzes text data, understands its context and meaning, and classifies the content of inquiries into specific categories.

[1692] "Analysis means" refers to a system component that uses a natural language engine to analyze the context of text data and classify the query content into a specific category.

[1693] A "generative artificial intelligence engine" is an artificial intelligence technology component that generates appropriate responses based on a specific category.

[1694] A "model selection means" is a system component that selects an appropriate generative artificial intelligence engine based on the analysis results.

[1695] "Response generation means" refers to a technology that generates responses to inquiries using a selected generative artificial intelligence engine.

[1696] A "response output means" is a system component that provides the generated response data to the user as text or audio.

[1697] "History analysis means" refers to a technology that stores the user's inquiry history and uses that history data for analysis to generate a response that takes contextual information into account.

[1698] "Communication means" refers to the technology by which a terminal and a server send and receive data via a communication path.

[1699] This invention relates to a system that responds quickly and accurately to a variety of user inquiries, and is comprised of a combination of advanced speech recognition technology, natural language processing technology, and a generative artificial intelligence engine.

[1700] System-wide configuration

[1701] This system primarily consists of three elements: user, terminal, and server. The user makes inquiries via voice through the terminal, which converts them into text data using speech recognition technology. The text data is sent to the server, which analyzes the received text and uses an optimal generative artificial intelligence engine to provide a response to the user.

[1702] Server operation

[1703] 1. Inquiry reception

[1704] The server receives text data sent from the terminal.

[1705] The received text is passed to the analysis device, and the analysis begins.

[1706] 2. Analysis and Categorization

[1707] The server uses analysis tools to analyze the context of the text data and classify it into specific categories.

[1708] Specifically, a natural language engine is used to classify the data into categories such as "fortune telling," "elderly care," and "housing."

[1709] 3. Model Selection

[1710] The server selects an appropriate generative artificial intelligence engine (e.g., GPT-3, DALL-E, etc.) based on the analysis results.

[1711] For example, if a question like "Please tell me my fortune for this month" is classified under the fortune-telling category, a generative artificial intelligence engine specifically for fortune-telling will be selected.

[1712] 4. Response generation

[1713] The server inputs a prompt (e.g., "Please tell me my fortune for this month") into the selected generative artificial intelligence engine and generates a response.

[1714] Using a response generation tool, specific answers such as "Your luck this month is good" are created.

[1715] 5. Send a response

[1716] The server formats the generated response into the appropriate format and converts it into audio data if necessary.

[1717] Send the response to the terminal.

[1718] Terminal operation

[1719] 1. Speech recognition

[1720] The device receives voice input from the user.

[1721] Speech-to-Text API is used to convert speech into text data.

[1722] Specifically, the audio "Please tell me my fortune for this month" is converted into the text format "Please tell me my fortune for this month".

[1723] 2. Response reception and output

[1724] The terminal receives the response data sent from the server.

[1725] The system uses a response output mechanism to provide the user with the received data as text or audio.

[1726] For example, the system could tell the user via voice message, "Your fortune this month is good."

[1727] User actions

[1728] 1. Inquiry

[1729] Users ask questions using their devices (such as public telephones, smartphones, or tablets) via voice.

[1730] For example, the device receives voice input when the user says, "I'm looking for a senior care facility."

[1731] 2. Confirmation of response

[1732] The user checks the response provided by the device.

[1733] You can ask further questions or inquire about additional relevant information as needed.

[1734] Specific usage examples

[1735] Fortune-telling inquiries

[1736] User: "Please tell me my fortune for this month."

[1737] Terminal: Converts audio into text data and sends it to the server.

[1738] Server: Classifies text data into fortune-telling categories and selects a generative artificial intelligence engine specifically for fortune-telling.

[1739] Server: Uses a fortune-telling engine to generate a response such as "Your fortune this month is good" and sends it to the terminal.

[1740] Terminal: Communicates responses to the user via voice.

[1741] Caregiving inquiries

[1742] User: "I'm looking for a senior living facility."

[1743] Terminal: Converts audio into text data and sends it to the server.

[1744] Server: Classifies text data into caregiving categories and selects a generative artificial intelligence engine specifically for caregiving.

[1745] Server: Uses an engine designed for elderly care to generate responses such as "We recommend XX elderly care facility" and send them to the terminal.

[1746] Terminal: Communicates responses to the user via voice.

[1747] This system enables quick and accurate responses to a wide range of user inquiries, thereby improving user satisfaction. Furthermore, by taking past inquiry history into account, it can provide more personalized responses.

[1748] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1749] Program execution flow: Explained in steps.

[1750] Step 1: User voice input

[1751] Users make inquiries to their devices (such as smartphones and tablets) using voice commands.

[1752] For example, a user might say, "Please tell me my fortune for this month."

[1753] Input: User's voice data

[1754] Output: Audio data is input to the terminal.

[1755] Step 2: Voice recognition on the device

[1756] The device uses speech recognition technology (e.g., Google Speech-to-Text API) to convert the user's voice into text data.

[1757] The system processes the audio data and generates the text, "Please tell me my fortune for this month."

[1758] Input: User's voice data

[1759] Data processing: Converting audio data to text data.

[1760] Output: Text data "Please tell me my fortune for this month"

[1761] Step 3: Send text data

[1762] The terminal sends the converted text data to the server. HTTP or similar communication protocols are used.

[1763] The text data is sent to the server and prepared for analysis.

[1764] Input: Text data "Please tell me my fortune for this month"

[1765] Data processing: Packeting and preparing text data for transmission.

[1766] Output: Text data sent to the server

[1767] Step 4: Text message received

[1768] The server receives text data sent from the terminal.

[1769] The received text data is passed to the analysis tool.

[1770] Input: Text data sent from the device

[1771] Data processing: Preparing for analysis of received data.

[1772] Output: Text data to be analyzed

[1773] Step 5: Analyze the inquiry

[1774] The server uses a natural language engine to analyze the context of the text data and then parses the query content.

[1775] For example, the text "Please tell me my fortune for this month" would be classified under the fortune-telling category.

[1776] Input: Text data to be analyzed

[1777] Data processing: Categorization using natural language processing

[1778] Output: Category information (e.g., "Fortune Telling")

[1779] Step 6: Model Selection

[1780] The server selects an appropriate generative artificial intelligence engine based on the analysis results.

[1781] For example, if it is classified as a fortune-telling category, a generative artificial intelligence engine specifically for fortune-telling (e.g., a GPT-3 model for fortune-telling) will be selected.

[1782] Input: Category information (e.g., "Fortune Telling")

[1783] Data processing: Selection of an appropriate AI model

[1784] Output: Selected generative artificial intelligence engine

[1785] Step 7: Generating the response

[1786] The server inputs a prompt (e.g., "Please tell me my fortune for this month") into the selected generative artificial intelligence engine and generates an appropriate response.

[1787] For example, a response such as "Your luck this month is good" is generated.

[1788] Input: Selected generative artificial intelligence engine, prompt

[1789] Data processing: Response generation by generative AI

[1790] Output: Generated response data (e.g., "Your fortune this month is good")

[1791] Step 8: Sending a response

[1792] The server formats the generated response into the appropriate format and converts it into audio data as needed.

[1793] The response data is sent to the terminal.

[1794] Input: Generated response data

[1795] Data processing: Formatting of response data and conversion to audio data (if necessary).

[1796] Output: Response data ready to be sent to the terminal

[1797] Step 9: Receive response / correction

[1798] The terminal receives the response data sent from the server.

[1799] Input: Response data sent from the server

[1800] Data processing: Processing of received data

[1801] Output: Response data in a state that can be provided to the user.

[1802] Step 10: Provide a response to the user

[1803] The terminal uses a response output means to provide the user with the received data as text or audio.

[1804] For example, the system could tell the user via voice message, "Your fortune this month is good."

[1805] Input: Response data in a state that can be provided to the user.

[1806] Data calculation: Output of response data

[1807] Output: The response provided to the user (text or audio)

[1808] This processing flow enables quick and accurate responses to a wide range of user inquiries.

[1809] (Application Example 1)

[1810] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1811] In autonomous vehicles, it is crucial for users to quickly and accurately obtain various information while driving. However, existing systems often lack sufficient accuracy in speech recognition and natural language processing, sometimes providing only incorrect or incomplete information. Furthermore, their ability to appropriately acquire and respond to real-time updated information is insufficient, which can impair user convenience. The problem this invention aims to solve is to address these challenges and provide a system that allows users to always obtain the latest information appropriately.

[1812] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1813] In this invention, the server includes speech recognition means for converting speech to text, analysis means for analyzing text data and classifying the inquiry content into specific categories, and model selection means for selecting a generative artificial intelligence model based on the analysis results. This makes it possible to generate and provide a quick and accurate response when a user queries for information in real time by voice within an autonomous vehicle.

[1814] "Voice recognition means" refers to a device or software that converts voice data input by a user into text data.

[1815] "Analysis means" refers to a device or software that analyzes text data using a natural language processing engine and classifies the content of a query into a specific category.

[1816] "Model selection means" refers to a device or software that selects an appropriate generative artificial intelligence model based on the analysis results.

[1817] "Response generation means" refers to a device or software that generates a response to an inquiry using a selected generative artificial intelligence model.

[1818] "Response output means" refers to a device or software that provides the generated response to the user in text or voice.

[1819] "Communication means" refers to a device or software that connects to an online server via a communication system within a vehicle and transmits and receives data in real time.

[1820] "Voice output means" refers to a device or software that provides the user with a voice response generated using the acoustic equipment inside the vehicle.

[1821] "History analysis means" refers to a device or software that stores the user's inquiry history and uses that history data for analysis to generate a response that takes contextual information into account.

[1822] This invention provides a system that responds quickly and accurately to user inquiries made by voice within an autonomous vehicle. The following is a specific embodiment of this system.

[1823] Hardware and software configuration

[1824] Hardware configuration:

[1825] 1. Microphone: A device installed inside a vehicle to capture sound.

[1826] 2. In-vehicle communication system: This is a communication system that connects to the internet and sends and receives data with online servers.

[1827] 3. Speaker: A device for providing the generated voice response to the user.

[1828] Software configuration:

[1829] 1. Speech Recognition Method: The speech acquired from the microphone is converted into text data using the Python library speech_recognition.

[1830] 2. Analysis method: A natural language processing engine is used to analyze the context of the text data and classify the query content into specific categories.

[1831] 3. Model Selection Method: Based on the analysis results, an appropriate generative artificial intelligence model (generative AI model) is selected.

[1832] 4. Response generation means: A response to the query is generated using a selected generative AI model.

[1833] 5. Response output method: The generated response is converted into audio data using the Google Text-to-Speech (gTTS) library and provided to the user through the speaker.

[1834] Specific operation of the system

[1835] 1. Voice Recognition: The user asks a question into the microphone inside the car. For example, "Where is the nearest gas station?"

[1836] 2. Text conversion: The speech recognition system converts the user's speech into text data.

[1837] 3. Analysis: The analysis tool analyzes the text data using a natural language processing engine and classifies the sentence "Where is the nearest gas station?" into the traffic information category.

[1838] 4. Model Selection: The model selection method selects a generation AI model for traffic information.

[1839] 5. Response Generation: The response generation means uses a selected generative AI model to generate a response such as, "The nearest gas station is 2km ahead on your right."

[1840] 6. Response Output: The response output means converts the generated response into audio data and transmits it to the user through the speaker.

[1841] Specific example

[1842] For example, if a user asks "Where is the nearest gas station?" while inside an autonomous vehicle, the system will respond in the following order:

[1843] 1. The microphone captures audio data, and the speech recognition system converts it into text data such as "Where is the nearest gas station?".

[1844] 2. The analysis tool analyzes the text data and classifies it into traffic information categories.

[1845] 3. The model selection method selects an AI model for generating traffic information.

[1846] 4. The response generation means generates the response, "The nearest gas station is 2km ahead on your right."

[1847] 5. The response output means converts the response into audio data using the gTTS library and provides it to the user through the speaker.

[1848] Example of a prompt

[1849] Example of a user inquiry: "Where is the nearest gas station?"

[1850] Prompt text: "In response to the voice-recognized inquiry 'Where is the nearest gas station?', the system will provide the location of the best gas station."

[1851] Examples of input content for a generative AI model:

[1852] Inquiry: "Where is the nearest gas station?"

[1853] Reply: "The nearest gas station is 2km ahead on your right."

[1854] In this way, users can obtain quick and accurate information using voice commands within autonomous vehicles.

[1855] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1856] Step 1:

[1857] The user makes a voice inquiry into the microphone inside the vehicle. The input is the user's voice data, and the output is the voice data captured by the microphone.

[1858] Step 2:

[1859] The device's speech recognition system converts the voice data input by the user into text data. The input for this step is the captured voice data, and the output is the converted text data. Specifically, the Python library speech_recognition is used to analyze the voice data and convert it into string information.

[1860] Step 3:

[1861] The terminal sends the converted text data to the server via the in-vehicle communication system. The input for this step is the converted text data, and the output is the text data sent to the server. Communication means are used for sending and receiving data.

[1862] Step 4:

[1863] The server's analysis mechanism analyzes the received text data using a natural language processing engine. The input for this step is the received text data, and the output is the analyzed query content and its categorization. Natural language processing techniques are used for the analysis, understanding the context and classifying it into appropriate categories.

[1864] Step 5:

[1865] The server's model selection mechanism selects an appropriate generative AI model based on the analyzed results. The input to this step is the analysis results and their category classification, and the output is the selected generative AI model. The selection mechanism determines the optimal generative AI model for each query category.

[1866] Step 6:

[1867] The server's response generation mechanism generates an appropriate response using a selected generative AI model. The inputs to this step are the selected generative AI model and the analyzed query, while the output is the generated response data. This includes the process by which the generative AI model generates the optimal answer to the query.

[1868] Step 7:

[1869] The server sends the generated response data to the terminal. The input to this step is the generated response data, and the output is the response data sent to the terminal. Communication means are used for sending and receiving data.

[1870] Step 8:

[1871] The terminal's response output mechanism converts the received response data into speech data. The input for this step is the received response data, and the output is the converted speech data. Specifically, the Google Text-to-Speech (gTTS) library is used to convert the text data into speech data.

[1872] Step 9:

[1873] The device's speaker plays the converted audio data to the user. The input in this step is the converted audio data, and the output is the provision of audio information to the user. Through the speaker, the user can hear the generated response.

[1874] This series of processes will enable users to obtain quick and accurate information using voice commands within autonomous vehicles.

[1875] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1876] This invention relates to a system that recognizes a user's emotions and generates a response based on those emotions, and combines speech recognition technology, natural language processing technology, generative artificial intelligence, and an emotion engine.

[1877] System-wide configuration

[1878] This system consists of a user, a terminal, and a server. The user makes inquiries through the terminal, which converts them into text data using speech recognition. The server then receives the text data and speech sentiment data, analyzes them using an analysis tool, selects an appropriate model to generate a response, and returns it to the terminal for the user to receive.

[1879] Server operation

[1880] 1. Inquiry reception

[1881] The server receives text data and sentiment data sent from the terminal.

[1882] The received data is passed to the analysis device, and the analysis begins.

[1883] 2. Analysis and categorization of text data

[1884] The server uses parsing tools to analyze the context of text data and classify it into specific categories (e.g., fortune-telling, elderly care, housing).

[1885] Based on the classified categories, an appropriate generative artificial intelligence model is selected.

[1886] 3. Analysis of emotional data

[1887] The server uses an emotion engine to analyze emotional data obtained from the voice and recognize the user's emotional state.

[1888] The recognized emotion data is combined with the query content to adjust the content and tone of the response.

[1889] 4. Response generation

[1890] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection method.

[1891] Using a response generation means, the generated response is formatted into an appropriate format and converted into audio data as needed.

[1892] 5. Send a response

[1893] The server sends the generated response data to the terminal.

[1894] Terminal operation

[1895] 1. Speech recognition

[1896] The device receives voice input from the user.

[1897] The system uses speech recognition to convert speech into text data and then sends that text data to a server.

[1898] Simultaneously, an emotion engine is used to extract emotion data from the audio and send it to the server.

[1899] 2. Response reception and output

[1900] The terminal receives the response data sent from the server.

[1901] The system uses a response output mechanism to provide the user with the received data as text or audio.

[1902] User actions

[1903] 1. Inquiry

[1904] Users ask questions using a device (such as a public telephone, smartphone, or tablet).

[1905] By asking questions using voice, information can be easily obtained.

[1906] 2. Confirmation of response

[1907] The user checks the response provided by the device.

[1908] You can ask further questions or inquire about additional relevant information as needed.

[1909] Specific usage examples

[1910] Fortune-telling inquiries

[1911] User: "Please tell me my fortune for this month (in an excited voice)."

[1912] Terminal: Converts audio into text data, extracts emotional data from the audio, and sends it to the server.

[1913] Server: Using analysis tools, the text "Please tell me my fortune for this month" is classified into the fortune-telling category, and a generative artificial intelligence model specifically for fortune-telling is selected.

[1914] Server: Uses an emotion engine to analyze emotional data and recognizes the state of being "excited."

[1915] Server: Based on the recognized emotion data, it generates a response such as, "Your luck this month is good, especially your career luck!" and sends it to the terminal.

[1916] Terminal: Conveys received responses to the user via voice.

[1917] Caregiving inquiries

[1918] User: "I'm looking for a senior living facility (in a tired voice)."

[1919] Terminal: Converts audio into text data, extracts emotional data from the audio, and sends it to the server.

[1920] Server: Using analysis tools, the text "I am looking for a facility for the elderly" is classified into the care category, and a generative artificial intelligence model specifically for care is selected.

[1921] Server: Uses an emotion engine to analyze emotional data and recognize the state of being "tired".

[1922] Server: Based on recognized emotion data, it generates a response such as, "We recommend XX senior care facility. You seem tired, rest is important," and sends it to the terminal.

[1923] Terminal: Conveys received responses to the user via voice.

[1924] In this embodiment of the present invention, it is possible to recognize the user's emotional state and provide an appropriate response accordingly, thereby greatly improving user satisfaction. The system can combine advanced analysis and emotion recognition for various inquiries, enabling it to provide users with more personalized services.

[1925] The following describes the processing flow.

[1926] Step 1:

[1927] The user speaks a specific question into the device's microphone (e.g., "Please tell me my fortune for this month").

[1928] Step 2:

[1929] The device uses speech recognition to convert the user's voice into text data.

[1930] Send the converted text data to the server.

[1931] Step 3:

[1932] The device uses an emotion engine to extract emotional data from the audio.

[1933] The extracted emotion data is sent to the server.

[1934] Step 4:

[1935] The server receives text data and sentiment data sent from the terminal.

[1936] Step 5:

[1937] The server passes the text data to the analysis tool, which then uses a natural language processing engine to analyze the context.

[1938] Step 6:

[1939] The server uses analysis tools to classify text data into specific categories (e.g., fortune-telling, elderly care, housing).

[1940] Step 7:

[1941] The server selects an appropriate generative artificial intelligence model based on the analysis results.

[1942] Step 8:

[1943] The server uses an emotion engine to analyze emotional data and recognize the user's emotional state.

[1944] Step 9:

[1945] The server combines the recognized emotion data with the query content to adjust the content and tone of the response.

[1946] Step 10:

[1947] The server generates an appropriate response using the generative artificial intelligence model selected by the model selection method.

[1948] Step 11:

[1949] The server formats the response generated using the response generation means into an appropriate format and converts it into audio data as needed.

[1950] Step 12:

[1951] The server sends the generated response data to the terminal.

[1952] Step 13:

[1953] The terminal receives the response data sent from the server.

[1954] Step 14:

[1955] The terminal uses a response output mechanism to provide the user with the received data in text or voice.

[1956] Step 15:

[1957] The user checks the response provided by the device.

[1958] Ask additional questions as needed, then return to step 1.

[1959] (Example 2)

[1960] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1961] Conventional speech recognition systems have been immature in generating responses that take user emotions into account, making it difficult to enhance user satisfaction. Furthermore, they often lack accuracy in analyzing inquiries and the appropriateness of their responses, particularly in providing personalized responses that reflect emotions. As a result, the user experience deteriorates, limiting the effectiveness of the system.

[1962] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1963] In this invention, the server includes emotion analysis means for extracting emotion data from speech, analysis means for analyzing text data using a natural language processing engine and classifying the inquiry content into specific categories, and model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on the analysis results. This makes it possible to generate personalized responses that take the user's emotions into consideration.

[1964] "Voice recognition means" refers to a technology or device for converting a user's voice input into text data.

[1965] "Emotional analysis means" refers to a technology or device for extracting emotional data from voice input and determining the user's emotional state.

[1966] A "natural language processing engine" is a technology or software that analyzes text data to understand its context and meaning.

[1967] "Analysis means" refers to a technique or device for classifying text data into specific categories.

[1968] A "generative artificial intelligence model" is an artificial intelligence model designed to generate appropriate responses based on input data.

[1969] "Model selection means" refers to a technique or device for selecting an appropriate generative artificial intelligence model based on analysis results.

[1970] "Response generation means" refers to a technology or apparatus that generates a response to a user's inquiry using a selected generative artificial intelligence model.

[1971] "Response output means" refers to a technology or device that provides the generated response to the user as text or audio.

[1972] "History analysis means" refers to a technology or device that analyzes user inquiry history data and generates a response that takes contextual information into account.

[1973] "Communication means" refers to the technology or device used by a terminal and a server to send and receive data via a communication line.

[1974] "Distributed processing means" refers to a technology or device that distributes processing to efficiently perform data analysis and response generation within a server.

[1975] This invention relates to a system that recognizes a user's emotions and generates a response based on those emotions. Specifically, it is a system that combines speech recognition technology, natural language processing technology, generative artificial intelligence, and emotion analysis technology.

[1976] System-wide configuration

[1977] This system consists of a user, a terminal, and a server. The user makes inquiries through the terminal, which converts them into text data using speech recognition. The server then receives the text data and speech emotion data, analyzes them using an analysis tool, selects an appropriate generative artificial intelligence model to generate a response, and returns it to the terminal for the user to receive.

[1978] Hardware and software used

[1979] Speech recognition means: The terminal uses a microphone to receive voice input from the user and a speech recognition service such as the Google Speech-to-Text API.

[1980] Sentiment analysis methods: IBM Watson Tone Analyzer and Microsoft Azure Emotion API are used for sentiment analysis.

[1981] Natural Language Processing Engine: Uses natural language processing libraries such as the BERT model, spaCy, and NLTK.

[1982] Generative AI Models: Advanced generative AI models such as GPT-4 are used.

[1983] Communication method: The terminal and server use the HTTP protocol over the internet to send and receive data.

[1984] Distributed processing methods: Cloud infrastructure (e.g., AWS, Google Cloud) is used to efficiently perform data analysis and response generation within the server.

[1985] Specific usage examples

[1986] Fortune-telling inquiries

[1987] Example: The user asks a voice question, "Please tell me my fortune for this month (in an excited voice)."

[1988] Terminal: Converts the audio into text data and sends the text "Please tell me my fortune for this month" to the server. At the same time, it extracts the emotion data "excited" from the audio and also sends this to the server.

[1989] Server: Uses analysis tools to classify text data into the "fortune-telling" category and selects a generative artificial intelligence model specifically for fortune-telling. Uses emotion analysis technology to recognize when the user is in an "excited" state.

[1990] Server: Based on recognized sentiment data and text data, it generates the response, "Your luck this month is good, especially your career luck!"

[1991] Terminal: Converts the generated response into speech data using speech synthesis technology and provides it to the user.

[1992] Caregiving inquiries

[1993] Example: The user asks a question via voice, "I'm looking for a senior living facility (in a tired voice)."

[1994] Terminal: Converts the voice into text data and sends the text "I am looking for a senior care facility" to the server. At the same time, it extracts emotion data, "I am tired," and also sends this to the server.

[1995] Server: Uses analysis tools to classify text data into the "caregiving" category and selects a generative artificial intelligence model specifically for caregiving. Uses emotion analysis technology to recognize that the user is in a "tired" state.

[1996] Server: Based on recognized sentiment data and text data, it generates a response such as, "We recommend XX senior living facility. You seem tired; rest is important."

[1997] Terminal: Converts the generated response into speech data using speech synthesis technology and provides it to the user.

[1998] This invention makes it possible to recognize a user's emotional state and provide an appropriate response accordingly. The system can combine advanced analysis and emotion recognition to provide users with more personalized services in response to various inquiries.

[1999] The flow of the specific processing in Example 2 will be explained using Figure 13.

[2000] Step 1:

[2001] Receiving and preprocessing voice input

[2002] The device receives voice input from the user. For example, it captures voice data such as "Please tell me my fortune for this month" through the smartphone's microphone.

[2003] Input: User's voice data (e.g., "Please tell me my fortune for this month.")

[2004] Specific operation: The device uses noise cancellation technology to improve the quality of audio data.

[2005] Output: Clear audio data with noise removed.

[2006] Step 2:

[2007] Speech recognition and text conversion

[2008] The device converts speech data into text data using speech recognition technology. Specifically, it uses the Google Speech-to-Text API for speech recognition.

[2009] Input: Clear audio data (e.g., "Please tell me my fortune for this month")

[2010] Specific operation: Send audio data to the API and retrieve text data as a response.

[2011] Output: Text data (Example: "Please tell me my fortune for this month")

[2012] Step 3:

[2013] Emotion analysis

[2014] The terminal uses sentiment analysis tools to extract sentiment data from voice data. In this case, IBM Watson Tone Analyzer is used.

[2015] Input: Clear audio data (e.g., "Please tell me my fortune for this month")

[2016] Specific operation: Voice data is sent to the emotion analysis engine, and emotion data is obtained as a result of the analysis.

[2017] Output: Sentiment data (e.g., "excited")

[2018] Step 4:

[2019] Sending data

[2020] The device sends text data and sentiment data to the server.

[2021] Input: Text data and sentiment data (e.g., "Please tell me my horoscope for this month," "I'm excited")

[2022] Specific operation: Send data to the server using an HTTP request.

[2023] Output: The server receives text data and sentiment data.

[2024] Step 5:

[2025] Text data analysis and categorization

[2026] The server analyzes the context of text data using analytical tools and classifies it into specific categories (e.g., fortune-telling, elderly care, housing). Here, the BERT model and spaCy are used.

[2027] Input: Text data (Example: "Please tell me my fortune for this month.")

[2028] Specific operation: Text data is fed into a natural language processing engine, and categories are determined based on the analysis results.

[2029] Output: Category information (e.g., "Fortune Telling")

[2030] Step 6:

[2031] Category-based model selection

[2032] The server selects an appropriate generative artificial intelligence model based on the analysis results. In this case, a generative artificial intelligence model such as GPT-4 is selected for the fortune-telling category.

[2033] Input: Category information (e.g., "Fortune Telling")

[2034] Specific operation: Use the model selection method within the server to select a relevant artificial intelligence model.

[2035] Output: Generative artificial intelligence model (e.g., GPT-4)

[2036] Step 7:

[2037] Response generation

[2038] The server uses a selected generative artificial intelligence model to generate responses based on text data and sentiment data. Examples of prompts include "Please tell me my fortune for this month" and the state of being "excited."

[2039] Input: Text data, sentiment data, generative artificial intelligence model (e.g., "Please tell me my fortune for this month," "I'm excited")

[2040] Specific operation: A prompt is fed into a generative artificial intelligence model, and the generated response is retrieved.

[2041] Output: Generated response data (Example: "Your luck this month is good, especially your career luck!")

[2042] Step 8:

[2043] Response formatting and speech conversion

[2044] The server formats the generated response into an appropriate format (e.g., JSON) and uses speech synthesis technology to convert it into audio data. In this case, the Google Text-to-Speech API is used.

[2045] Input: Generated response data (Example: "Your luck this month is good, especially your career luck!")

[2046] Specific operation: Sends a text response to the speech synthesis engine and generates speech data.

[2047] Output: Audio data or formatted text data

[2048] Step 9:

[2049] Sending a response

[2050] The server sends the generated response data to the terminal.

[2051] Input: Voice data or formatted text data (e.g., "Your luck this month is good, especially your career luck!")

[2052] Specific operation: Send data to the terminal using an HTTP response.

[2053] Output: The terminal receives the response data.

[2054] Step 10:

[2055] Providing a response

[2056] The terminal receives response data sent from the server and provides it to the user.

[2057] Input: Voice data or formatted text data (e.g., "Your luck this month is good, especially your career luck!")

[2058] Specific operation: If the data is audio, it will be played as sound from the speaker; if the data is text, it will be displayed on the screen.

[2059] Output: Provides a response to the user.

[2060] The above outlines the specific program processing flow of this system. By describing the inputs and outputs at each step in detail, the overall operation of the system is made easier to understand.

[2061] (Application Example 2)

[2062] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[2063] In modern food delivery services, users experience a range of emotions when placing an order, and a response tailored to those emotions is required. However, traditional systems lacked the ability to analyze user emotions and could only provide uniform responses, resulting in a poor user experience. Furthermore, there was a lack of technology to generate appropriate responses that considered context and emotions in response to user inquiries. As a result, gaining user trust was difficult, and improving the quality of service was essential.

[2064] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes: speech recognition means that receive voice input from the user and convert the voice into text; analysis means that analyze the text data using a natural language processing engine and classify the inquiry into specific categories; model selection means that select a generative artificial intelligence model that generates an appropriate response based on the analysis results; emotion analysis means that extract emotion data from the user's voice and adjust the response content based on that emotion; response generation means that generate a response to the inquiry using the selected generative artificial intelligence model; and response output means that provide the generated response to the user in text or voice. This makes it possible to provide an appropriate response according to the user's emotions and improve the user experience.

[2065] "A speech recognition means that receives voice input from a user and converts that voice into text" refers to a technology or device that converts a user's voice into digital data and generates text data from the voice.

[2066] "Analysis means for analyzing text data using a natural language processing engine and classifying query content into specific categories" refers to a technology or apparatus that performs a process of analyzing generated text data, identifying its content, and classifying it into specific categories.

[2067] "Model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on analysis results" refers to a technology or device that selects the most appropriate generative artificial intelligence model based on the analyzed data.

[2068] "Response generation means for generating a response to a query using a selected generative artificial intelligence model" refers to a technology or apparatus that uses a selected generative artificial intelligence model to generate an appropriate response to a query.

[2069] "An emotion analysis means that extracts emotion data from a user's voice and adjusts the response content based on that emotion" refers to a technology or device that analyzes a user's voice information to recognize their emotional state and adjusts the response content based on that emotional state.

[2070] "Response output means that provides the generated response to the user in text or audio" refers to a technology or device that conveys the generated response to the user in text or audio format.

[2071] This invention is a system that combines speech recognition technology, natural language processing technology, generative artificial intelligence models, and an emotion engine to provide customizable responses tailored to the user's emotions when using a food delivery service. Specific embodiments are described below.

[2072] System-wide configuration

[2073] The system of the present invention consists of a user, a terminal, and a server. The user provides voice input through the terminal, which converts it into text data using speech recognition means. Subsequently, the server receives the text data and sentiment data, analyzes them using analysis means, selects an appropriate generative artificial intelligence model to generate a response, and returns it to the terminal for the user to receive.

[2074] Server operation

[2075] Inquiry reception:

[2076] The server receives text data and sentiment data sent from the terminal. It then passes the received data to the analysis system and begins the analysis.

[2077] Text data analysis and categorization:

[2078] The server uses analysis tools to analyze the context of text data and classify it into specific categories (e.g., food categories). Based on the classified categories, it selects an appropriate generative artificial intelligence model.

[2079] Analysis of emotional data:

[2080] The server uses an emotion engine to analyze emotional data obtained from the voice and recognize the user's emotional state. Based on the recognized emotion, it adjusts the content and tone of the response.

[2081] Response generation:

[2082] The server generates an appropriate response using a generative artificial intelligence model selected by the model selection means. The response generation means then formats the generated response into an appropriate format and converts it into audio data as needed.

[2083] Send response:

[2084] The server sends the generated response data to the terminal.

[2085] Terminal operation

[2086] Speech recognition:

[2087] The terminal receives voice input from the user, converts the voice into text data using speech recognition technology, and sends that text data to the server. Simultaneously, it extracts emotion data from the voice using an emotion engine and sends it to the server.

[2088] Response reception and output:

[2089] The terminal receives response data sent from the server and uses a response output means to provide the received data to the user as text or audio.

[2090] User actions

[2091] inquiry:

[2092] The user asks questions using their device via voice. Example: "I'd like to order a pizza (in a worried voice)."

[2093] Response confirmation:

[2094] The user confirms the response provided by the device. Example: "Thank you for ordering a pizza. Please let us know if you have any questions."

[2095] Core technology

[2096] This system is built using the following technologies:

[2097] Speech recognition technology: Uses the speech_recognition library.

[2098] Natural language processing technology: Uses the natural_language_processing engine.

[2099] Generative artificial intelligence models: Generative AI models

[2100] Emotion engine: Uses emotion_recognition technology

[2101] Specific example

[2102] For example, if a user types "I'd like to order a pizza (in a worried voice)," the server converts the speech into text data and recognizes the emotion of "worry." Based on the recognized emotion, it generates a response such as "Thank you for ordering a pizza. Please let us know if you have any questions," and provides this to the user in voice.

[2103] Example of a prompt

[2104] Examples of prompts for a generative AI model include the following:

[2105] "Generate a reassuring response when the user is worried. For example, if the user's voice input is 'I'd like to order a pizza (in a worried voice),' create a response like, 'Thank you for ordering a pizza. Please let us know if you have any questions.'"

[2106] In this way, it becomes possible to automatically generate responses that respond to the user's emotions and improve the user experience.

[2107] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[2108] Step 1:

[2109] The user uses their device to perform voice input. For example, they might say, "I'd like to order a pizza (in a worried voice)." This voice is the input.

[2110] Step 2:

[2111] The device receives voice input from the user and converts the voice into text data using speech recognition technology. The process of converting voice data into text data is performed, and the output is the text data "I want to order a pizza."

[2112] Step 3:

[2113] Simultaneously, the device uses emotion analysis to extract emotion data from the user's voice. The emotion in the voice data is analyzed, and the emotion data "worry" is output.

[2114] Step 4:

[2115] The terminal sends the text data from step 2 and the sentiment data from step 3 to the server. The server then receives the user's inquiry and sentiment state.

[2116] Step 5:

[2117] The server passes the received text data to a natural language processing engine, which then categorizes the query into a specific category. In this case, the text data "I want to order a pizza" is classified under the "Food Category."

[2118] Step 6:

[2119] The server uses a model selection mechanism to select an appropriate generative artificial intelligence model based on the analysis results of the text data. In this case, a generative artificial intelligence model specifically for food delivery is selected.

[2120] Step 7:

[2121] The server uses emotion analysis tools to re-analyze the user's emotional data and adjusts its response accordingly. For example, based on the emotional data of "worry," it provides a prompt to the AI ​​model to generate a reassuring response.

[2122] Step 8:

[2123] The server uses a selected generative artificial intelligence model to generate a tailored response. Here, the text response "Thank you for ordering pizza. Please let us know if you have any questions." is generated.

[2124] Step 9:

[2125] The server formats the generated response text appropriately and converts it into audio data as needed. This stage includes the process of converting text data to audio data.

[2126] Step 10:

[2127] The server sends the generated response data to the terminal. This data includes responses in both text and audio formats.

[2128] Step 11:

[2129] The terminal receives response data from the server and provides it to the user using a response output mechanism. Specifically, it plays a voice message saying, "Thank you for ordering pizza. If you have any questions, we will be happy to assist you."

[2130] This series of steps is expected to improve the user experience of food delivery services by allowing users to receive appropriate responses tailored to their emotional state.

[2131] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[2132] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[2133] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[2134] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[2135] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[2136] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[2137] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[2138] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[2139] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[2140] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[2141] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[2142] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[2143] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[2144] 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.

[2145] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[2146] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[2147] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[2148] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[2149] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[2150] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[2151] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted as being incorporated by reference.

[2152] The following is further disclosed regarding the embodiments described above.

[2153] (Claim 1)

[2154] A speech recognition means that receives voice input from a user and converts that voice into text,

[2155] An analysis method that analyzes text data using a natural language processing engine and classifies the content of inquiries into specific categories,

[2156] A model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on the analysis results,

[2157] A response generation means that generates a response to a query using a selected generative artificial intelligence model,

[2158] A response output means that provides the generated response to the user in text or voice,

[2159] A system that includes this.

[2160] (Claim 2)

[2161] The system according to claim 1, further comprising a history analysis means for storing the user's inquiry history and using the history data for analysis to generate a response that takes contextual information into account.

[2162] (Claim 3)

[2163] The system according to claim 1, comprising communication means for a terminal and a server to send and receive data via a communication line.

[2164] "Example 1"

[2165] (Claim 1)

[2166] A speech recognition means that receives voice input from a user and converts that voice into text,

[2167] An analysis method that analyzes text data using a natural language engine and classifies the content of inquiries into specific categories,

[2168] A model selection means for selecting a generative artificial intelligence engine that generates an appropriate response based on the analysis results,

[2169] A response generation means that generates a response to an inquiry using a selected generative artificial intelligence engine,

[2170] A response output means that provides the generated response data to the user as text or audio,

[2171] A system that includes this.

[2172] (Claim 2)

[2173] The system according to claim 1, further comprising a history analysis means for storing the user's inquiry history and using the history data for analysis to generate a response that takes contextual information into account.

[2174] (Claim 3)

[2175] The system according to claim 1, comprising communication means for a terminal and a server to send and receive data via a communication path.

[2176] "Application Example 1"

[2177] (Claim 1)

[2178] A speech recognition means that receives voice input from a user and converts that voice into text,

[2179] An analysis method that analyzes text data using a natural language processing engine and classifies the content of inquiries into specific categories,

[2180] A model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on the analysis results,

[2181] A response generation means that generates a response to a query using a selected generative artificial intelligence model,

[2182] A response output means that provides the generated response to the user in text or voice,

[2183] A communication method that connects to an online server via the vehicle's communication system and performs real-time data transmission and reception,

[2184] A system that includes this.

[2185] (Claim 2)

[2186] The system according to claim 1, further comprising an audio output means for providing a voice response generated using an acoustic device inside the vehicle to the user.

[2187] (Claim 3)

[2188] The system according to claim 1, further comprising a history analysis means for storing the user's inquiry history and using the history data for analysis to generate a response that takes contextual information into account.

[2189] "Example 2 of combining an emotion engine"

[2190] (Claim 1)

[2191] A speech recognition means that receives voice input from a user and converts that voice into text,

[2192] A means of emotion analysis that extracts emotional data from voice,

[2193] An analysis method that analyzes text data using a natural language processing engine and classifies the content of inquiries into specific categories,

[2194] A model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on the analysis results,

[2195] A response generation means that generates a response to a query using a selected generative artificial intelligence model,

[2196] A response output means that provides the generated response to the user in text or voice,

[2197] A system that includes this.

[2198] (Claim 2)

[2199] The system according to claim 1, further comprising a history analysis means for storing the user's inquiry history and using the history data for analysis to generate a response that takes contextual information into account.

[2200] (Claim 3)

[2201] The system according to claim 1, comprising communication means for a terminal and a server to send and receive data via a communication line, and distributed processing means for performing data analysis and response generation within the server.

[2202] "Application example 2 when combining with an emotional engine"

[2203] (Claim 1)

[2204] A speech recognition means that receives voice input from a user and converts that voice into text,

[2205] An analysis method that analyzes text data using a natural language processing engine and classifies the content of inquiries into specific categories,

[2206] A model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on the analysis results,

[2207] A response generation means that generates a response to a query using a selected generative artificial intelligence model,

[2208] An emotion analysis means that extracts emotion data from the user's voice and adjusts the response content based on that emotion,

[2209] A response output means that provides the generated response to the user in text or voice,

[2210] A system that includes this.

[2211] (Claim 2)

[2212] The system according to claim 1, further comprising a history analysis means for storing the user's inquiry history and using the history data for analysis to generate a response that takes contextual information into account.

[2213] (Claim 3)

[2214] The system according to claim 1, comprising communication means for a terminal and a server to send and receive data via a communication line. [Explanation of Symbols]

[2215] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A speech recognition means that receives voice input from a user and converts that voice into text, An analysis method that analyzes text data using a natural language processing engine and classifies the content of inquiries into specific categories, A model selection means for selecting a generative artificial intelligence model that generates an appropriate response based on the analysis results, A response generation means that generates a response to a query using a selected generative artificial intelligence model, A response output means that provides the generated response to the user in text or voice, A system that includes this.

2. The system according to claim 1, further comprising a history analysis means for storing the user's inquiry history and using the history data for analysis to generate a response that takes contextual information into account.

3. The system according to claim 1, which includes communication means for a terminal and a server to send and receive data via a communication line.

Images