system

Abstract

Provide a system. 【Solution means】 Means for capturing the user's voice, Means for converting the captured voice into string data, Means for analyzing the string data to identify the user's request, Means for searching for information based on the identified user's request, Means for generating a response text based on the searched information, Means for converting the generated response text into voice data, Means for playing back the converted voice data to the user, A system including the above.

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 conventional car navigation systems and voice recognition systems, it is often impossible to accurately recognize a user's utterance, and the user may feel frustrated when obtaining a destination or information. In addition, due to the lack of multi-language support and responses tailored to the personality of the character, the user experience has been limited. Furthermore, it has been difficult to smoothly perform a translation function and complex information search. For this reason, there has been a demand for a system that allows users to obtain a smooth and enjoyable driving experience.

Means for Solving the Problems

[0005] The present invention provides means for accurately capturing a user's voice and converting it into text data. It also solves the above-mentioned problems by providing a system that includes means for analyzing the text data to identify the user's request, means for searching for information based on the identified request, means for generating response text based on the search results, means for converting the response text into audio data, and means for playing the audio data to the user. Furthermore, it provides a system that improves the user experience by including means for translating the searched information into multiple languages ​​and means for editing the generated response text to suit the personality of a specific character.

[0006] "Means for capturing audio" refers to a device that includes microphones and sensors for recording user speech as digital audio data.

[0007] "Means for converting speech to text data" refers to a device that includes speech recognition software or algorithms that analyze captured speech data and generate corresponding text data.

[0008] "Means of analyzing string data to identify user requests" refers to software or modules that use natural language processing (NLP) techniques to understand text data and extract user intent and requests.

[0009] "Means of retrieving information" refers to software, including search engines and APIs, that have the function of retrieving relevant information from the internet or databases based on the requests of a specific user.

[0010] "Means for generating response text" refers to a device that includes a text generation algorithm and software for generating a response sentence to the user based on search results and analysis results.

[0011] "Means for converting response text into audio data" refers to a device that includes text-to-speech (TTS) software or algorithms that convert the generated text into synthesized speech.

[0012] "Means for playing audio data to a user" refers to a device that includes hardware and software for delivering converted audio data to the user via speakers or other audio output devices.

[0013] "Means of multilingual translation" refers to software, including translation engines and APIs, for translating retrieved information or generated text into different languages.

[0014] "Means for editing to match the personality of a specific character" refers to a device that includes text editing algorithms and modules for editing generated response text to match the personality and speaking style of a predefined character. [Brief explanation of the drawing]

[0015] [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] Shows an emotion map to which a plurality of emotions are mapped. [Figure 11] It is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Example 2 when an emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when an emotion engine is combined.

Mode for Carrying Out the Invention

[0016] 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.

[0017] First, the terms used in the following description will be described.

[0018] In the following embodiments, a numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of a plurality of arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of a plurality of 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.

[0019] 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.

[0020] In the following embodiments, the signed storage is one or more non-volatile storage devices that store various programs and various parameters. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes.

[0021] 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).

[0022] 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."

[0023] [First Embodiment]

[0024] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0025] 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.

[0026] 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).

[0027] 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.

[0028] 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.

[0029] 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.

[0030] 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.

[0031] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0032] 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.

[0033] 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.

[0034] 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.

[0035] 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".

[0036] This invention relates to a system that accurately captures a user's voice and uses that voice for navigation and information retrieval. The following describes a detailed configuration for implementing this system.

[0037] System Overview

[0038] This system performs a series of processes: capturing the user's voice, converting it into text data, analyzing it to identify the request, searching for information to generate response text, converting it back into audio data, and playing it back to the user. Furthermore, it includes functions to translate information into multiple languages ​​and to edit response text to match the personality of a specific character.

[0039] Program processing flow

[0040] 1. Voice input (device)

[0041] The user says, "Find a nearby cafe."

[0042] The device uses the microphone to capture the user's voice and saves it as audio data.

[0043] 2. Voice recognition (device)

[0044] The device uses a speech recognition module to analyze the voice data and convert it into text, "Find a nearby cafe."

[0045] 3. Sending a request (from the terminal)

[0046] The terminal generates a request containing the converted text data and the user's current location information, and sends it to the server.

[0047] 4. Request parsing (server)

[0048] The server receives the request data, uses a natural language processing (NLP) module to analyze the text, and identifies the request as "Find a cafe."

[0049] 5. Information Retrieval (Server)

[0050] The server searches for relevant cafe information from the internet and databases based on the specified request. For example, it might use a map API to identify nearby cafes.

[0051] 6. Response generation (server)

[0052] The server generates response text for the user based on the search results. For example, it might generate, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0053] 7. Multilingual support and character editing (server)

[0054] The server translates response text into multiple languages ​​as needed. It also edits response text to suit the personality of specific characters.

[0055] 8. Sending a response (server)

[0056] The server sends the generated response text to the terminal.

[0057] 9. Speech synthesis (device)

[0058] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[0059] 10. Response playback (terminal)

[0060] The device plays audio data through its speaker and responds to the user, saying, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[0061] Specific example

[0062] Situation: "Look for a nearby cafe."

[0063] 1. The user says, "Find a nearby cafe."

[0064] 2. The device captures the audio and saves it as audio data.

[0065] 3. The device's voice recognition module converts the voice data into "Find a nearby cafe."

[0066] 4. The device generates a request containing the converted text and current location information and sends it to the server.

[0067] 5. The server receives the request, parses it with the NLP module, and identifies the request as "Find a cafe".

[0068] 6. The server searches for information about cafes related to it using a map API.

[0069] 7. The server generates the response text "There are 3 cafes nearby. We recommend 'Cafe Latte Street'." based on the search results.

[0070] 8. If necessary, the server will translate the response text into multiple languages ​​and edit it to suit the character's personality.

[0071] 9. The server sends a response text to the terminal.

[0072] 10. The terminal converts the response text into speech using a speech synthesis module.

[0073] 11. The device plays a voice message through its speaker, responding to the user with, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0074] In this way, the system allows users to enjoy conversations with their favorite characters while obtaining the information they need.

[0075] The following describes the processing flow.

[0076] Step 1:

[0077] The user says, "Find a nearby cafe."

[0078] The device uses the microphone to capture this audio and saves it as audio data.

[0079] Step 2:

[0080] The device analyzes the voice data captured using its speech recognition module and converts it into text. For example, it might convert it to text like, "Find a nearby cafe."

[0081] Step 3:

[0082] The terminal generates request data that includes generated text data and user location information. This request data contains the user's request and location information.

[0083] Step 4:

[0084] The device sends the request data to the server via the internet.

[0085] Step 5:

[0086] The server receives the request data. The request data includes text and location information from the user.

[0087] Step 6:

[0088] The server uses a natural language processing (NLP) module to analyze the text portion of the request and identify the user's request. For example, it might extract the phrase "find a cafe."

[0089] Step 7:

[0090] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[0091] Step 8:

[0092] The server organizes the search results and generates response text for the user. For example, it might generate a sentence like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0093] Step 9:

[0094] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[0095] Step 10:

[0096] The server sends the generated and edited response text to the terminal.

[0097] Step 11:

[0098] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[0099] Step 12:

[0100] The device plays audio data through its speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[0101] In this way, the system allows users to enjoy conversations with their favorite characters while smoothly obtaining the information they need.

[0102] (Example 1)

[0103] 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."

[0104] Conventional speech recognition systems have had the problem of difficulty in accurately capturing user voices and providing information that meets the user's requests. Furthermore, they are limited to single-language support and cannot provide responses tailored to the personality of a character, resulting in a limited user experience. This invention aims to solve these problems and provide a more interactive and multi-functional information provision system.

[0105] 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.

[0106] In this invention, the server includes means for analyzing a request to identify the user's request, means for retrieving information based on the identified user's request, and means for generating response text based on the retrieved information. This makes it possible to analyze the user's request with high accuracy and provide appropriate information quickly.

[0107] "User" refers to a person who uses a system or an end-user.

[0108] "Audio capture" is the process of recording audio signals using a microphone.

[0109] "Converting to string data" refers to converting audio data into text format using speech recognition technology.

[0110] A "request" refers to a data packet containing a user's request or command, which is sent to the server.

[0111] "Location information" refers to geographical coordinate information obtained using GPS or other positioning technologies.

[0112] A "server" refers to a computer system that processes data and provides services to clients via a network.

[0113] "Natural language processing" refers to the technology that enables computers to understand, generate, and analyze human language.

[0114] "Information retrieval" is the process of obtaining relevant information from databases or the internet based on user requests.

[0115] "Response text" refers to a text-based representation of a user's request or answer to a user's request.

[0116] "Translating into multiple languages" is the process of converting text written in one language into another language.

[0117] "Editing to match the character's personality" refers to modifying text to reflect the characteristics and speaking style of a specific character.

[0118] "Converting to audio data" refers to the process of changing text information into computer-generated speech.

[0119] "Audio playback" refers to the physical output of audio data using sound devices such as speakers.

[0120] This invention is a system that captures user voice and uses it for navigation and information retrieval. The following describes a detailed configuration for implementing this system.

[0121] General overview

[0122] This system includes a voice input device, a speech recognition module, a data communication module, an information retrieval module, a natural language processing module, a response generation module, and a speech synthesis module. Through a series of processes, it is possible to analyze and respond to user voice requests. In addition, it also provides multilingual translation and character-based response generation capabilities.

[0123] Hardware and software to be used

[0124] Device: Mobile information terminals such as smartphones and tablets

[0125] Microphone: Built-in microphone or external microphone device

[0126] Server: Cloud-based servers, such as AWS® or Google® Cloud

[0127] Speech recognition module: Google Speech-to-Text API

[0128] Natural language processing modules: SpaCy and Google NLP API

[0129] Information retrieval module: Google Maps API or any database

[0130] Multilingual translation module: Google Translate API

[0131] Speech synthesis module: Amazon Polly

[0132] Processing flow

[0133] 1. The user speaks through a voice input device, for example, saying, "Find a nearby cafe."

[0134] 2. The device's built-in microphone captures this audio and saves it as audio data. This data is temporarily stored in the device's storage.

[0135] 3. The device uses a speech recognition module to convert this speech data into text data. The converted text will read, "Find a nearby cafe."

[0136] 4. The terminal generates the converted text and location information as request data and sends it to the server. This includes precise location information using GPS and other positioning technologies.

[0137] 5. The server receives the request data and parses the text using a natural language processing module. This analysis identifies the user's request (e.g., "find a cafe").

[0138] 6. The server uses the Google Maps API to search for nearby cafes based on the user's request. The retrieved information includes details such as the cafe's name, address, and rating.

[0139] 7. The server generates response text based on the search results. For example, it might generate something like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0140] 8. If necessary, the server will use a multilingual translation module to translate response text. It may also edit the text to suit the character's personality.

[0141] 9. The server sends the edited and translated text to the terminal.

[0142] 10. The terminal converts the received text into speech data using a speech synthesis module.

[0143] 11. Finally, the device plays this audio data to the user, providing the desired information verbally.

[0144] Specific example

[0145] Let's take the situation "finding a nearby cafe" as an example.

[0146] The user says, "Find a nearby cafe."

[0147] The device captures the audio and converts it to text using a speech recognition module.

[0148] The device sends the converted text and location information to the server.

[0149] The server uses an NLP module to analyze the request and identify the request to search for a cafe.

[0150] The server uses the Google Maps API to search for cafe information and generates a response text.

[0151] The server performs multilingual translation and character editing, and then sends the text to the terminal.

[0152] The device converts text into speech using its speech synthesis module.

[0153] The device plays the generated audio and provides the user with cafe information.

[0154] This process allows users to quickly and accurately obtain information through voice input. Furthermore, multilingual support and character-based responses provide a more user-friendly experience.

[0155] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0156] Step 1:

[0157] The user says, "Find a nearby cafe."

[0158] Input: User's spoken voice

[0159] Specific action: The user speaks the desired information (in this case, information about a cafe) into the voice input device.

[0160] Step 2:

[0161] The device uses its built-in microphone to capture the user's voice and saves it as audio data.

[0162] Input: User's spoken voice

[0163] Output: Audio data file (WAV or MP3 format)

[0164] Specific operation: The device's microphone acquires audio as a digital signal, which is then converted and saved as an audio data file.

[0165] Step 3:

[0166] The device uses a speech recognition module (e.g., Google Speech-to-Text API) to convert the audio data into text data.

[0167] Input: Audio data file

[0168] Output: Text data "Find a nearby cafe"

[0169] Specific operation: The speech recognition module receives an audio data file as input, performs speech analysis, and converts it into text. The converted text data is saved to the device's memory.

[0170] Step 4:

[0171] The device generates request data based on the converted text data and location information obtained from its GPS function, and sends it to the server.

[0172] Input: Text data "Find a nearby cafe" and location information

[0173] Output: Request data (API request format)

[0174] Specific operation: Generate a request combining the converted text and location information, and send this request to the server using the HTTPS protocol.

[0175] Step 5:

[0176] The server receives the request data, uses a natural language processing (NLP) module to parse the text data, and identifies the user's request.

[0177] Input: Request data (text and location information)

[0178] Output: Analysis results (User request: Find a cafe)

[0179] Specific operation: The server analyzes the request and uses an NLP module to identify the request as "Find a cafe." This information is stored in the server's memory.

[0180] Step 6:

[0181] The server uses the Google Maps API based on the user's request to search for cafe information based on the user's current location.

[0182] Input: Analysis results (user request) and location information

[0183] Output: Cafe information list (cafe name, address, rating, etc.)

[0184] Specific operation: Call the Google Maps API to retrieve cafe information based on the current location. The retrieved information is stored in the server's memory.

[0185] Step 7:

[0186] The server generates a response text based on the cafe information it has acquired.

[0187] Input: Cafe information list

[0188] Output: Response text "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0189] Specific operation: Analyzes cafe information and generates response text in a user-friendly format.

[0190] Step 8:

[0191] The server translates the response text using a multilingual translation module as needed and edits the response text to suit the personality of the specific character.

[0192] Input: Response text

[0193] Output: Translated and edited response text

[0194] Specific actions: Use the Google Translate API to translate into multiple languages ​​and edit the text to suit the character's personality.

[0195] Step 9:

[0196] The server sends the edited and translated response text to the terminal.

[0197] Input: Response text (after translation and editing)

[0198] Output: Transmitted data

[0199] Specific operation: The edited and translated text data is sent to the terminal using the HTTPS protocol.

[0200] Step 10:

[0201] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[0202] Input: Response text

[0203] Output: Audio data file (MP3 format, etc.)

[0204] Specific operation: Text data is converted into speech data using a speech synthesis module such as Amazon Polly. The converted speech data is stored in the device's storage.

[0205] Step 11:

[0206] The device plays audio data through its speaker and responds to the user.

[0207] Input: Audio data file

[0208] Output: Physical audio output

[0209] Specific operation: The device's speaker is used to play the generated audio data, and the user is responded with, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0210] (Application Example 1)

[0211] 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."

[0212] In modern customer service, there is a demand for immediate answers to customer questions and provision of appropriate information within the store. However, conventional systems have difficulty providing information via voice input, and furthermore, multilingual support and responses using specific character voices are not adequately provided. As a result, the customer experience is limited, and improving service quality remains a challenge.

[0213] 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.

[0214] In this invention, the server includes means for capturing the user's voice, means for converting the captured voice into string data, means for analyzing the string data to identify the user's request, means for retrieving information based on the identified user's request, means for generating response text based on the retrieved information, means for converting the generated response text into audio data, means for playing the converted audio data back to the user, means for converting the generated response text into the voice of a specific virtual character, means for acquiring the user's current location information, and means for providing information about features within the store based on the acquired location information and the user's request. This enables immediate customer service within the store, and significantly improves the customer experience by enabling multilingual support and responses by specific characters.

[0215] "Means for capturing user voice" refers to a device for acquiring the voice emitted by a user as a digital signal using an audio input device such as a microphone.

[0216] "Methods for converting captured audio into text data" refers to the process of converting acquired audio data into corresponding text data using speech recognition technology.

[0217] "Means for analyzing the string data to identify the user's request" refers to a mechanism that uses natural language processing technology to understand and identify the user's intent and request from the string data.

[0218] "Means for retrieving information based on the requests of a specified user" refers to means for searching and extracting information from databases or the internet that meets the user's requirements.

[0219] "Means for generating response text based on searched information" refers to the process of creating an appropriate response to the user in text format based on search results.

[0220] "Means for converting generated response text into audio data" refers to a system that converts text into audio data using speech synthesis technology.

[0221] "Means for playing back converted audio data to the user" refers to a device for presenting the generated audio data to the user through a speaker or ear-hook type device.

[0222] "Means for converting generated response text into the voice of a specific virtual character" refers to the process of synthesizing generated text into speech using a pre-configured character and voice.

[0223] "Means for obtaining the user's current location information" refers to technologies that use GPS or other location information services to determine the user's current location.

[0224] "Means for providing information about in-store features based on acquired location information and user requests" refers to a system that provides specific product placement and guidance information within a store in accordance with user requests and location information.

[0225] System Overview

[0226] This invention is a system that retrieves information from a user's voice and returns a voice response. The system consists of various modules for realizing voice capture, speech recognition, natural language processing, information retrieval, response generation, speech synthesis, location information acquisition, and voice responses in the voice of a specific virtual character. The system uses a wearable device such as smart glasses to provide customer service in physical stores.

[0227] Hardware and software to be used

[0228] Hardware:

[0229] Smart glasses (voice input device, speaker)

[0230] GPS or location acquisition module

[0231] software:

[0232] Speech recognition modules (e.g., Google Speech-to-Text, Amazon Transcribe)

[0233] Natural Language Processing (NLP) modules (e.g., Google NLP API, OpenAI® GPT-3®)

[0234] Speech synthesis modules (e.g., Microsoft® Azure® TTS, Amazon Polly)

[0235] Data processing and data calculation

[0236] 1. Audio capture:

[0237] When a user speaks into the microphone attached to the smart glasses, audio data is captured. This data is stored digitally on the smart glasses.

[0238] 2. Speech recognition:

[0239] The captured audio data is sent to the speech recognition module in the smart glasses and converted into text data. This converts the audio into text format.

[0240] 3. Submit the request:

[0241] The converted text data is sent to the server along with the user's current location information. Here, the user's location information is obtained using a GPS module.

[0242] 4. Request Analysis:

[0243] The server analyzes the received request data and uses a natural language processing (NLP) module to identify the user's request.

[0244] 5. Information Retrieval:

[0245] Based on the specified request, the server searches for relevant information from databases and the internet.

[0246] 6. Response generation:

[0247] The server generates appropriate response text based on the search results. Furthermore, this response text can be translated into multiple languages ​​and edited to sound like a specific virtual character.

[0248] 7. Send a reply:

[0249] The response text is sent from the server to the smart glasses.

[0250] 8. Speech synthesis:

[0251] The speech synthesis module within the smart glasses converts the received response text into speech data. If the voice is converted to that of a specific character, a voice that reflects the character's characteristics will be generated.

[0252] 9. Audio Playback:

[0253] The generated audio data is played back to the user through the smart glasses' speaker.

[0254] Specific example

[0255] Situation: "Can you recommend a wine?"

[0256] 1. The user speaks to the smart glasses and says, "Please recommend a wine."

[0257] 2. The smart glasses capture the audio data and convert it into text data.

[0258] 3. The converted text data and current location information are sent to the server.

[0259] 4. The server analyzes the received data and identifies the request, "Please recommend a wine."

[0260] 5. The server retrieves recommended wine information from the database.

[0261] 6. Based on the information obtained by the server, it generates the response text: "The recommended wine is 'a specific wine'. It is available for purchase on the left side of the liquor section."

[0262] 7. Convert the generated text into the voice of a specific virtual character.

[0263] 8. The server sends the generated response text to the smart glasses.

[0264] 9. The smart glasses convert the speech into voice using a speech synthesis module and play it back to the user.

[0265] Example of a prompt:

[0266] When a user asks, "Can you recommend a wine?", the response text will be generated based on the following text:

[0267] Prompt: The customer asked, "Can you recommend a wine?" Please create a response text following the format below.

[0268] 1. Product Name

[0269] 2. A brief description of the sales location

[0270] Example response: "Our recommended wine is 'a specific wine.' It's located on the left side of the liquor section."

[0271] Text: "Please recommend a wine."

[0272] Thus, the invention provides a concrete means for implementing voice-based customer service in physical stores.

[0273] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0274] Step 1:

[0275] When a user speaks into the smart glasses, the smart glasses' microphone captures the voice and saves it as digital audio data. For example, the user might say, "Tell me your recommended wine." The input is an audio signal, and the output is digital audio data.

[0276] Step 2:

[0277] The speech recognition module on the device performs speech recognition using the captured audio data and converts it into text data. Specifically, the speech recognition module (e.g., Google Speech-to-Text) generates the text "Tell me your recommended wine" from the audio data. The input is audio data, and the output is text data.

[0278] Step 3:

[0279] The terminal acquires the converted text data and the user's current location information (such as GPS data), and sends a request containing these data to the server. As a specific operation, the location information acquisition module detects the current location and incorporates it into the transmission packet together with the text data. The input is the text data and the location information, and the output is the request data.

[0280] Step 4:

[0281] The server analyzes the received request data and uses a natural language processing module (e.g., Google NLP API) to identify the user's request. As a specific operation, it analyzes a request such as "Tell me some recommended wines" and identifies a request such as "Wine recommendation". The input is the request data, and the output is the analyzed request.

[0282] Step 5:

[0283] Based on the analyzed request, the server searches for relevant information from the specified database or the Internet. As a specific operation, it executes a database query to obtain information related to "recommended wines". The input is the analyzed request, and the output is the search result.

[0284] Step 6:

[0285] Based on the search result, the server generates a response text. Furthermore, if necessary, the generated text is translated into multiple languages and edited into the voice of a specific virtual character. As a specific operation, it generates a text such as "The recommended wine is 'Specific Wine'. The sales location is on the left side of the liquor corner." and edits it into the character voice. The input is the search result, and the output is the response text.

[0286] Step 7:

[0287] The server sends the generated response text to the terminal. As a specific operation, it converts the response text into a packet and sends it to the terminal. The input is the response text, and the output is the transmission packet.

[0288] Step 8:

[0289] The speech synthesis module on the terminal (e.g., Microsoft Azure TTS) converts the received response text into speech data. As a specific operation, it generates speech data from the text data. The input is the response text, and the output is the speech data.

[0290] Step 9:

[0291] The terminal plays the speech data and provides a response to the user through the speaker of the smart glasses. As a specific operation, it passes the generated speech data to the playback device and outputs the speech. The input is the speech data, and the output is the speech information provided to the user.

[0292] Furthermore, an emotion engine for estimating the user's emotion may be combined. That is, the specific processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform specific processing using the user's emotion.

[0293] System Overview

[0294] The present invention is a system that executes a series of processes of capturing the user's voice, converting it into string data, analyzing it to identify a request, searching for information, generating a response text, converting it into speech data, and playing it back to the user. Furthermore, the present invention incorporates an emotion engine for recognizing the user's emotion and provides a function for improving the user experience. It also includes a function of translating the retrieved information into multiple languages and editing it according to the personality of a specific character.

[0295] Program Processing Flow

[0296] 1. Voice Input (Device)

[0297] The user says, "Find a nearby café."

[0298] The device uses the microphone to capture the user's voice and saves it as voice data.

[0299] 2. Voice Recognition (Device)

[0300] The device analyzes the voice data captured using the voice recognition module and converts it into text. For example, it converts it into the text "Find a nearby café."

[0301] 3. Emotion Recognition (Device)

[0302] The device uses the emotion engine to recognize the user's emotions (such as joy, anger, sadness, etc.) from the voice data and text data.

[0303] 4. Request Sending (Device) <The device generates request data including the converted text data, the user's location information, and the recognized emotion data, and sends it to the server.

[0304]

[0305] 5. Request Analysis (Server)

[0306] The server receives the request data. The request data includes the user's text, location information, and emotion data.

[0307] 6. Information Search (Server)

[0308] The server uses the natural language processing (NLP) module to analyze the text part of the request and identify the user's request. For example, it identifies the request "Find a café."

[0309] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[0310] 7. Response generation (server)

[0311] The server generates response text for the user based on the search results. For example, it might generate a sentence like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0312] The server adjusts the generated response text based on recognized emotion data. For example, if the user is angry, the tone of the response will be softened.

[0313] 8. Multilingual support and character editing (server)

[0314] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[0315] 9. Sending a response (server)

[0316] The server sends the generated and edited response text to the terminal.

[0317] 10. Speech synthesis (device)

[0318] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[0319] 11. Response playback (terminal)

[0320] The device plays audio data through its speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[0321] Specific example

[0322] Situation: "Look for a nearby cafe."

[0323] 1. The user says, "Find a nearby cafe."

[0324] 2. The device captures the audio and saves it as audio data.

[0325] 3. The device's voice recognition module converts the voice data into "Find a nearby cafe."

[0326] 4. The device uses an emotion engine to recognize the user's emotions and determines that they are "excited."

[0327] 5. The device generates a request containing the converted text, location information, and sentiment data, and sends it to the server.

[0328] 6. The server receives the request, parses it with the NLP module, and identifies the request as "Find a cafe".

[0329] 7. The server uses a map API to search for information about nearby cafes.

[0330] 8. The server generates the response text "There are 3 cafes nearby. We recommend 'Cafe Latte Street'." based on the search results.

[0331] 9. Based on the recognized emotion data (excited), the server adjusts the response text and generates it in a brighter tone.

[0332] 10. If necessary, the server will translate the response text into multiple languages ​​and edit it to suit the character's personality.

[0333] 11. The server sends a response text to the terminal.

[0334] 12. The terminal converts the response text into speech using a speech synthesis module.

[0335] 13. The device plays a voice message through its speaker, responding to the user in a cheerful tone, saying, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0336] In this way, the system allows users to enjoy conversations with their favorite characters while smoothly obtaining the necessary information. Furthermore, responses adapted to the user's emotions provide a more personalized experience.

[0337] The following describes the processing flow.

[0338] Step 1:

[0339] The user says, "Find a nearby cafe."

[0340] The device uses the microphone to capture audio and saves it as audio data.

[0341] Step 2:

[0342] The device uses a speech recognition module to analyze the captured audio data and convert it into text data such as "Find a nearby cafe."

[0343] Step 3:

[0344] The device uses an emotion engine to recognize the user's emotions from voice data and converted text data. For example, it might determine that the user is "excited" based on the tone and speed of their voice.

[0345] Step 4:

[0346] The device generates request data that includes converted text data, user location information, and recognized sentiment data. This request data contains the user's request, location information, and sentiment information.

[0347] Step 5:

[0348] The device sends the request data to the server via the internet.

[0349] Step 6:

[0350] The server receives the request data, which includes the user's text, location information, and sentiment data.

[0351] Step 7:

[0352] The server uses a natural language processing (NLP) module to analyze the text portion of the request and identify the user's request. For example, it might extract the request "Find a cafe."

[0353] Step 8:

[0354] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[0355] Step 9:

[0356] The server organizes the search results and generates response text for the user. For example, it might generate a response like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0357] Step 10:

[0358] The server adjusts the response text based on recognized sentiment data. For example, if the user is excited, the response tone becomes brighter and positive words are added.

[0359] Step 11:

[0360] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[0361] Step 12:

[0362] The server sends the generated and formatted response text to the terminal.

[0363] Step 13:

[0364] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[0365] Step 14:

[0366] The device plays audio data through its speaker and responds to the user in a cheerful tone, saying, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[0367] Through this series of processes, users can enjoy conversations with their favorite characters while quickly and effectively obtaining the information they need. Furthermore, emotion recognition provides more personalized responses, improving the user experience.

[0368] (Example 2)

[0369] 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".

[0370] Conventional voice dialogue systems struggled to generate flexible responses that reflected user emotions or to provide responses tailored to the individuality of specific characters. As a result, the user experience was limited, and it was difficult to address individual requests. Furthermore, multilingual support was insufficient, making it difficult to serve a global user base.

[0371] 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.

[0372] In this invention, the server includes means for acquiring user voice, means for converting the acquired voice into text data, means for analyzing the converted text data to identify the user's request, means for retrieving information based on the identified user's request, means for generating response data based on the retrieved information, means for converting the generated response data into voice data, means for providing the converted voice data to the user, means for recognizing the user's emotions, and means for adjusting the response data based on the recognized emotions. This enables flexible response generation based on the user's emotions, provision of responses tailored to specific characters, and multilingual support.

[0373] "Means of acquiring user voice" refers to devices and technologies that capture user voice and acquire it as data.

[0374] "Means of converting acquired audio into text data" refers to technologies and algorithms that convert audio data into string data.

[0375] "Means of analyzing converted text data to identify user requests" refers to technologies and methods that analyze text data to identify what the user wants.

[0376] "Means of retrieving information based on the requests of identified users" refers to technologies and systems that retrieve relevant information from databases and external services based on user requests.

[0377] "Means for generating response data based on retrieved information" refers to algorithms and technologies that generate appropriate responses for the user based on retrieved information.

[0378] "Means for converting generated response data into audio data" refers to technologies and devices that convert text-based response data into audio data.

[0379] "Means of providing the converted audio data to the user" refers to speakers or other audio output devices for playing the audio data to the user.

[0380] "Means of recognizing user emotions" refers to technologies and algorithms that identify emotions from user voice and text data.

[0381] "Means of adjusting response data based on recognized emotions" refers to technologies and methods that adjust the tone and content of responses according to the user's emotions.

[0382] "Means of translating into multiple languages" refers to technologies and systems that translate text data into different languages.

[0383] "Methods for editing to suit the characteristics of a specific character" refers to techniques and methods for modifying response data to match the personality and manner of speaking of a particular character.

[0384] This invention is a system that performs a series of processes: acquiring user voice, converting it into text data, analyzing it to identify requests, retrieving information, generating response data, converting it back into voice data, and providing it to the user. Furthermore, this invention provides a function to recognize the user's emotions and adjust the response data accordingly. It also includes a function to translate the retrieved information into multiple languages ​​and edit it to suit a specific character.

[0385] Hardware and software to be used

[0386] This system primarily uses the following hardware and software.

[0387] hardware

[0388] Device: A device equipped with a microphone and speaker for audio capture and playback (e.g., smartphone, tablet, smart speaker)

[0389] Server: A remote computer used for data processing and management.

[0390] software

[0391] Speech recognition module: An API for converting speech to text (e.g., Google Cloud Speech-to-Text API)

[0392] Emotion recognition module: A service for analyzing user emotions (e.g., Amazon Comprehend)

[0393] Natural Language Processing Module: A model for analyzing user requests (e.g., OpenAI GPT-3)

[0394] Map API: A service for searching for nearby information based on the user's location (e.g., Google Maps API).

[0395] Speech synthesis module: A technology for converting text into speech (e.g., Amazon Polly)

[0396] Translation Module: An API for translating response text into multiple languages ​​(e.g., Microsoft Translator).

[0397] Specific example

[0398] The following are examples of specific methods for implementing this invention.

[0399] Situation: "Looking for a nearby cafe"

[0400] 1. The user says, "Find a nearby cafe."

[0401] 2. The device uses the microphone to capture audio and saves it as audio data.

[0402] 3. The device uses a speech recognition module (Google Cloud Speech-to-Text API) to convert the captured audio into text: "Find a nearby cafe."

[0403] 4. The device uses an emotion recognition module (Amazon Comprehend) to analyze the voice and text data and determine that the user is "excited."

[0404] 5. The device generates a request containing text data, location information (GPS data), and sentiment data, and sends it to the server.

[0405] 6. The server receives the request, parses it using the NLP module (OpenAI GPT-3), and identifies the request as "Find a cafe."

[0406] 7. The server uses a map API (Google Maps API) to search for cafe information near the user's current location.

[0407] 8. Based on the search results, the server generates response data such as, "There are 3 cafes nearby. We recommend 'Cafe Latte Street'."

[0408] 9. Based on the recognized emotion data "excited," the server generates a response text in a cheerful tone.

[0409] 10. If necessary, the server will translate the response text into multiple languages ​​using a translation module (Microsoft Translator) and edit it to suit the characteristics of the specific character.

[0410] 11. The server sends the generated and edited response data to the terminal.

[0411] 12. The device uses a speech synthesis module (Amazon Polly) to convert the received response text into speech data.

[0412] 13. The device responds through the speaker, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0413] Example of a prompt

[0414] "Generate responses based on the user's emotions when they are looking for a nearby cafe."

[0415] "Generate a response for when the user is angry."

[0416] As described above, this invention is a system that efficiently executes a series of processes, starting with user voice input, going through text conversion and emotion recognition, searching for appropriate information, and generating and providing responses. As a result, users can obtain flexible responses that respond to their emotions, multilingual support, and personalized experiences tailored to their characters.

[0417] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0418] Program processing flow

[0419] Step 1: Voice input (on the device)

[0420] input:

[0421] The user says, "Find a nearby cafe."

[0422] Specific operation:

[0423] The device's microphone captures the audio and saves it as audio data.

[0424] output:

[0425] The acquired audio data is saved.

[0426] Step 2: Voice Recognition (Device)

[0427] input:

[0428] Captured audio data.

[0429] Specific operation:

[0430] The device calls the Google Cloud Speech-to-Text API and uploads the audio data. The API analyzes the audio and retrieves the text "Find a nearby cafe."

[0431] output:

[0432] Text data converted from audio data: "Find a nearby cafe."

[0433] Step 3: Emotion Recognition (Device)

[0434] input:

[0435] Audio and text data: "Find a nearby cafe."

[0436] Specific operation:

[0437] The device uses Amazon Comprehend to analyze voice and text data. From the analysis results, it extracts emotions and determines that the user is "excited."

[0438] output:

[0439] The recognized emotion data is "excited".

[0440] Step 4: Sending the request (on the device)

[0441] input:

[0442] Text data "Looking for a nearby cafe," sentiment data "Excited," and current location obtained from GPS.

[0443] Specific operation:

[0444] The terminal generates request data containing this information and sends it to the server.

[0445] output:

[0446] Request data sent to the server.

[0447] Step 5: Request parsing (server)

[0448] input:

[0449] Request data received by the server.

[0450] Specific operation:

[0451] The server logs the request data to a log file. A natural language processing (NLP) module (OpenAI GPT-3) is used to analyze the text data "Find a nearby cafe". The user's request "Looking for a cafe" is identified from the analysis results.

[0452] output:

[0453] User request: "I'm looking for a cafe."

[0454] Step 6: Information Retrieval (Server)

[0455] input:

[0456] The user's request is "I'm looking for a cafe" and their location information is provided.

[0457] Specific operation:

[0458] The server accesses the Google Maps API to search for nearby cafes based on location information. It then parses the response from the API to retrieve relevant cafe information.

[0459] output:

[0460] The retrieved cafe information (e.g., 3 nearby cafes including "Cafe Latte Street").

[0461] Step 7: Response generation (server)

[0462] input:

[0463] The acquired cafe information and emotion data indicated "excited."

[0464] Specific operation:

[0465] The server generates a text response based on cafe information: "There are three cafes nearby. We recommend 'Cafe Latte Street'." Based on recognized sentiment data, the tone of the response text is adjusted to be brighter.

[0466] output:

[0467] Adjusted response text: "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0468] Step 8: Multilingual support and character editing (server)

[0469] input:

[0470] The generated response text reads, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0471] Specific operation:

[0472] If necessary, the server translates the response text into multiple languages ​​using a translation module. It then edits it to suit the characteristics of the specific character.

[0473] output:

[0474] Translated into multiple languages, response text tailored to the character's characteristics.

[0475] Step 9: Send response (server)

[0476] input:

[0477] Edited response text.

[0478] Specific operation:

[0479] The server sends the edited response text to the terminal.

[0480] output:

[0481] The response text sent to the terminal.

[0482] Step 10: Speech synthesis (device)

[0483] input:

[0484] The received response text.

[0485] Specific operation:

[0486] The device uses a speech synthesis module (Amazon Polly) to convert the response text into speech data.

[0487] output:

[0488] Audio data.

[0489] Step 11: Response playback (device)

[0490] input:

[0491] Generated audio data.

[0492] Specific operation:

[0493] The device plays audio data through its built-in speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[0494] output:

[0495] Voice response to the user.

[0496] (Application Example 2)

[0497] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0498] Conventional voice assistant systems are limited to simple information retrieval and responses to user requests, making it difficult to provide personalized responses that take into account the user's emotions and circumstances. Furthermore, they lack multilingual support and the ability to edit responses to match the character's personality, posing a challenge in providing an optimal user experience, particularly in virtual stores. Additionally, the system's ability to provide information considering the user's location is underdeveloped, making it difficult to deliver appropriate information to users in real time.

[0499] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0500] In this invention, the server includes means for recognizing the user's emotions, means for acquiring the user's location information, and means for searching relevant data based on the acquired user location information and generating response text. This makes it possible to provide personalized responses in real time that are tailored to the user's emotions and current situation. In addition, by incorporating multilingual support and response editing functions that match character personalities, a more natural and engaging user experience can be achieved.

[0501] A "user" refers to a person who uses the system.

[0502] "Means for capturing speech" refers to a device or technology for acquiring a user's speech as audio data.

[0503] "Means of converting to string data" refers to technology or equipment for converting audio data into text data.

[0504] "Means of analysis to identify user requests" refers to technology or devices that process converted text data to identify what the user is asking for.

[0505] "Means of recognizing user emotions" refers to technology or devices that determine a user's emotions (e.g., joy, anger, sadness, etc.) from their voice data or speech content.

[0506] "Means of retrieving information" refers to technologies or devices for finding appropriate information based on the user's needs and feelings.

[0507] "Means for generating response text" refers to technology or equipment for creating text that serves as a response to a user based on retrieved information.

[0508] "Means of converting to audio data" refers to technology or equipment for converting generated response text into audio output.

[0509] "Means for playing audio data" refers to the technology or device used to allow the user to listen to the converted audio data.

[0510] "Means of translation into multiple languages" refers to technologies or devices that translate generated response text into different languages.

[0511] "Means of editing to match a character's personality" refers to techniques or devices that adjust generated response text to suit a specific character.

[0512] "Means for acquiring location information" refers to technology or devices used to determine the user's current location.

[0513] "Means for searching relevant data and generating response text" refers to a technology or device that searches for appropriate information based on acquired location information and user requests and creates response text.

[0514] This invention relates to a system that captures user voice, converts it into text data, analyzes it to identify requests, retrieves information, generates response text, converts it back into audio data, and plays it back to the user. Furthermore, this system has the capability to recognize user emotions and provides personalized responses according to the user's emotions. The detailed configuration and operation of the system are described below.

[0515] Hardware and software configuration

[0516] This system primarily consists of user terminals and servers. User terminals include devices such as smartphones, smart glasses, and head-mounted displays (HMDs), and these devices include the following software modules.

[0517] 1. Voice Capture Module: Captures the user's voice using the device's microphone.

[0518] 2. Speech Recognition Module: Converts speech data into text data. Specifically, it uses the Google Cloud Speech-to-Text API.

[0519] 3. Emotion Recognition Module: Recognizes emotions from the user's voice and text data. Specifically, it uses the Microsoft Azure Emotion API.

[0520] 4. Location Information Acquisition Module: Acquires the current location information using the GPS function of the user's device.

[0521] The server side includes the following software modules:

[0522] 1. Natural Language Processing (NLP) Module: Uses the Google Cloud Natural Language API to analyze and identify user requests.

[0523] 2. Information Retrieval Module: Searches for appropriate information based on user requests and location information.

[0524] 3. Response Generation Module: Generates response text based on the searched information and the user's sentiment.

[0525] 4. Multilingual Translation Module: Translates response text into multiple languages ​​as needed.

[0526] 5. Character Editing Module: Edit response text to match the personality of a specific character.

[0527] System Operation Description

[0528] When a user says, "Tell me your recommended shampoo," the smart device's microphone captures the voice and saves it as audio data. Then, a speech recognition module (Google Cloud Speech-to-Text API) is used to convert the captured audio data into text, "Tell me your recommended shampoo." At the same time, an emotion recognition module (Microsoft Azure Emotion API) recognizes the user's emotion and determines that they are "interested."

[0529] Request data is generated, containing the converted text data, user location information, and recognized sentiment data, and sent to the server. The server uses a natural language processing (NLP) module to parse the request and identify the request for "shampoo recommendations."

[0530] The server uses an information retrieval module to search the database for relevant data and generates the response text "We recommend 'Shampoo Excellence'." Based on the recognized sentiment data, the response text is edited and presented in a user-friendly tone. Additionally, a multilingual translation module is used to translate the text as needed, and a character editing module is used to edit it to match the character's personality.

[0531] The device receives the response text and converts it into speech data using a text-to-speech (TTS) module (Amazon Polly). It then plays the audio through the smart device's speaker, telling the user, "We recommend 'Shampoo Excellence'."

[0532] Specific examples and prompt statements

[0533] A concrete example of this scenario would be a series of actions taken by a smart device in response to a user's request, such as "Tell me your recommended shampoo." The following is an example of a prompt to a generative AI model.

[0534] Example of a prompt:

[0535] You are a customer service assistant in a virtual store. Generate a response when a user asks for a "recommended shampoo." Respond in a friendly and positive tone, especially if the user shows interest. Remember to mention specific product names.

[0536] In this way, the present invention can provide personalized information in response to user requests and improve the user experience.

[0537] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0538] Step 1:

[0539] Voice input

[0540] User: "Can you recommend a shampoo?"

[0541] Device: Captures the user's voice using the microphone and saves it as audio data.

[0542] Input: User's speech (audio data)

[0543] Output: Audio data

[0544] Step 2:

[0545] Speech recognition

[0546] Terminal: Analyzes captured audio data using a speech recognition module (Google Cloud Speech-to-Text API) and converts it to text.

[0547] Input: Audio data

[0548] Output: Text data "Please recommend a shampoo"

[0549] Step 3:

[0550] emotion recognition

[0551] Terminal: Uses an emotion recognition module (Microsoft Azure Emotion API) to recognize the user's emotions from voice and text data and determine that they are "interested."

[0552] Input: Audio data, text data

[0553] Output: Sentiment data "Interested"

[0554] Step 4:

[0555] Send Request

[0556] Terminal: Generates request data including converted text data, user location information, and recognized sentiment data, and sends it to the server.

[0557] Input: Text data "Please recommend a shampoo," location information, sentiment data "Interested"

[0558] Output: Request data

[0559] Step 5:

[0560] Request Analysis

[0561] Server: Receives request data, uses a natural language processing (NLP) module to parse the text portion and identify the user's request. Identifies the request as "shampoo recommendation".

[0562] Input: Request data

[0563] Output: User request "Shampoo recommendation"

[0564] Step 6:

[0565] Information Retrieval

[0566] Server: Retrieves information on relevant products from the database based on the identified request and location information.

[0567] Input: User request "Shampoo recommendation", location information

[0568] Output: Search result "Recommended shampoo: 'Shampoo Excellence'"

[0569] Step 7:

[0570] Response generation

[0571] Server: Based on the obtained search results and sentiment data, it generates response text for the user. It generates "We recommend 'Shampoo Excellence'."

[0572] Input: Search results, sentiment data "Interested"

[0573] Output: Response text: "We recommend 'Shampoo Excellence'."

[0574] Step 8:

[0575] Multilingual support and character editing

[0576] Server: Translates response text into multiple languages ​​as needed and edits it to suit the character's personality.

[0577] Input: Response text: "My recommendation is 'Shampoo Excellence'."

[0578] Output: Edited response text

[0579] Step 9:

[0580] Send response

[0581] Server: Sends the edited response text to the terminal.

[0582] Input: Edited response text

[0583] Output: Response text received by the terminal

[0584] Step 10:

[0585] Speech synthesis

[0586] Terminal: Converts the received response text into speech data using a text-to-speech (TTS) module (Amazon Polly).

[0587] Input: Response text: "My recommendation is 'Shampoo Excellence'."

[0588] Output: Audio data

[0589] Step 11:

[0590] Response playback

[0591] Device: Plays audio data through the speaker and tells the user, "We recommend 'Shampoo Excellence'."

[0592] Input: Audio data

[0593] Output: Voice response to the user: "We recommend 'Shampoo Excellence'."

[0594] 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.

[0595] 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.

[0596] 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.

[0597] [Second Embodiment]

[0598] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0599] 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.

[0600] 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).

[0601] 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.

[0602] 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.

[0603] 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).

[0604] 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.

[0605] 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.

[0606] 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.

[0607] 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.

[0608] 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.

[0609] 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".

[0610] This invention relates to a system that accurately captures a user's voice and uses that voice for navigation and information retrieval. The following describes a detailed configuration for implementing this system.

[0611] System Overview

[0612] This system performs a series of processes: capturing the user's voice, converting it into text data, analyzing it to identify the request, searching for information to generate response text, converting it back into audio data, and playing it back to the user. Furthermore, it includes functions to translate information into multiple languages ​​and to edit response text to match the personality of a specific character.

[0613] Program processing flow

[0614] 1. Voice input (device)

[0615] The user says, "Find a nearby cafe."

[0616] The device uses the microphone to capture the user's voice and saves it as audio data.

[0617] 2. Voice recognition (device)

[0618] The device uses a speech recognition module to analyze the voice data and convert it into text, "Find a nearby cafe."

[0619] 3. Sending a request (from the terminal)

[0620] The terminal generates a request containing the converted text data and the user's current location information, and sends it to the server.

[0621] 4. Request parsing (server)

[0622] The server receives the request data, uses a natural language processing (NLP) module to analyze the text, and identifies the request as "Find a cafe."

[0623] 5. Information Retrieval (Server)

[0624] The server searches for relevant cafe information from the internet and databases based on the specified request. For example, it might use a map API to identify nearby cafes.

[0625] 6. Response generation (server)

[0626] The server generates response text for the user based on the search results. For example, it might generate, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0627] 7. Multilingual support and character editing (server)

[0628] The server translates response text into multiple languages ​​as needed. It also edits response text to suit the personality of specific characters.

[0629] 8. Sending a response (server)

[0630] The server sends the generated response text to the terminal.

[0631] 9. Speech synthesis (device)

[0632] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[0633] 10. Response playback (terminal)

[0634] The device plays audio data through its speaker and responds to the user, saying, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[0635] Specific example

[0636] Situation: "Look for a nearby cafe."

[0637] 1. The user says, "Find a nearby cafe."

[0638] 2. The device captures the audio and saves it as audio data.

[0639] 3. The device's voice recognition module converts the voice data into "Find a nearby cafe."

[0640] 4. The device generates a request containing the converted text and current location information and sends it to the server.

[0641] 5. The server receives the request, parses it with the NLP module, and identifies the request as "Find a cafe".

[0642] 6. The server searches for information about cafes related to it using a map API.

[0643] 7. The server generates the response text "There are 3 cafes nearby. We recommend 'Cafe Latte Street'." based on the search results.

[0644] 8. If necessary, the server will translate the response text into multiple languages ​​and edit it to suit the character's personality.

[0645] 9. The server sends a response text to the terminal.

[0646] 10. The terminal converts the response text into speech using a speech synthesis module.

[0647] 11. The device plays a voice message through its speaker, responding to the user with, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0648] In this way, the system allows users to enjoy conversations with their favorite characters while obtaining the information they need.

[0649] The following describes the processing flow.

[0650] Step 1:

[0651] The user says, "Find a nearby cafe."

[0652] The device uses the microphone to capture this audio and saves it as audio data.

[0653] Step 2:

[0654] The device analyzes the voice data captured using its speech recognition module and converts it into text. For example, it might convert it to text like, "Find a nearby cafe."

[0655] Step 3:

[0656] The terminal generates request data that includes generated text data and user location information. This request data contains the user's request and location information.

[0657] Step 4:

[0658] The device sends the request data to the server via the internet.

[0659] Step 5:

[0660] The server receives the request data. The request data includes text and location information from the user.

[0661] Step 6:

[0662] The server uses a natural language processing (NLP) module to analyze the text portion of the request and identify the user's request. For example, it might extract the phrase "find a cafe."

[0663] Step 7:

[0664] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[0665] Step 8:

[0666] The server organizes the search results and generates response text for the user. For example, it might generate a sentence like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0667] Step 9:

[0668] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[0669] Step 10:

[0670] The server sends the generated and edited response text to the terminal.

[0671] Step 11:

[0672] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[0673] Step 12:

[0674] The device plays audio data through its speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[0675] In this way, the system allows users to enjoy conversations with their favorite characters while smoothly obtaining the information they need.

[0676] (Example 1)

[0677] 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".

[0678] Conventional speech recognition systems have had the problem of difficulty in accurately capturing user voices and providing information that meets the user's requests. Furthermore, they are limited to single-language support and cannot provide responses tailored to the personality of a character, resulting in a limited user experience. This invention aims to solve these problems and provide a more interactive and multi-functional information provision system.

[0679] 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.

[0680] In this invention, the server includes means for analyzing a request to identify the user's request, means for retrieving information based on the identified user's request, and means for generating response text based on the retrieved information. This makes it possible to analyze the user's request with high accuracy and provide appropriate information quickly.

[0681] "User" refers to a person who uses a system or an end-user.

[0682] "Audio capture" is the process of recording audio signals using a microphone.

[0683] "Converting to string data" refers to converting audio data into text format using speech recognition technology.

[0684] A "request" refers to a data packet containing a user's request or command, which is sent to the server.

[0685] "Location information" refers to geographical coordinate information obtained using GPS or other positioning technologies.

[0686] A "server" refers to a computer system that processes data and provides services to clients via a network.

[0687] "Natural language processing" refers to the technology that enables computers to understand, generate, and analyze human language.

[0688] "Information retrieval" is the process of obtaining relevant information from databases or the internet based on user requests.

[0689] "Response text" refers to a text-based representation of a user's request or answer to a user's request.

[0690] "Translating into multiple languages" is the process of converting text written in one language into another language.

[0691] "Editing to match the character's personality" refers to modifying text to reflect the characteristics and speaking style of a specific character.

[0692] "Converting to audio data" refers to the process of changing text information into computer-generated speech.

[0693] "Audio playback" refers to the physical output of audio data using sound devices such as speakers.

[0694] This invention is a system that captures user voice and uses it for navigation and information retrieval. The following describes a detailed configuration for implementing this system.

[0695] General overview

[0696] This system includes a voice input device, a speech recognition module, a data communication module, an information retrieval module, a natural language processing module, a response generation module, and a speech synthesis module. Through a series of processes, it is possible to analyze and respond to user voice requests. In addition, it also provides multilingual translation and character-based response generation capabilities.

[0697] Hardware and software to be used

[0698] Device: Mobile information terminals such as smartphones and tablets

[0699] Microphone: Built-in microphone or external microphone device

[0700] Servers: Cloud-based servers, such as AWS or Google Cloud

[0701] Speech recognition module: Google Speech-to-Text API

[0702] Natural language processing modules: SpaCy and Google NLP API

[0703] Information retrieval module: Google Maps API or any database

[0704] Multilingual translation module: Google Translate API

[0705] Speech synthesis module: Amazon Polly

[0706] Processing flow

[0707] 1. The user speaks through a voice input device, for example, saying, "Find a nearby cafe."

[0708] 2. The device's built-in microphone captures this audio and saves it as audio data. This data is temporarily stored in the device's storage.

[0709] 3. The device uses a speech recognition module to convert this speech data into text data. The converted text will read, "Find a nearby cafe."

[0710] 4. The terminal generates the converted text and location information as request data and sends it to the server. This includes precise location information using GPS and other positioning technologies.

[0711] 5. The server receives the request data and parses the text using a natural language processing module. This analysis identifies the user's request (e.g., "find a cafe").

[0712] 6. The server uses the Google Maps API to search for nearby cafes based on the user's request. The retrieved information includes details such as the cafe's name, address, and rating.

[0713] 7. The server generates response text based on the search results. For example, it might generate something like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0714] 8. If necessary, the server will use a multilingual translation module to translate response text. It may also edit the text to suit the character's personality.

[0715] 9. The server sends the edited and translated text to the terminal.

[0716] 10. The terminal converts the received text into speech data using a speech synthesis module.

[0717] 11. Finally, the device plays this audio data to the user, providing the desired information verbally.

[0718] Specific example

[0719] Let's take the situation "finding a nearby cafe" as an example.

[0720] The user says, "Find a nearby cafe."

[0721] The device captures the audio and converts it to text using a speech recognition module.

[0722] The device sends the converted text and location information to the server.

[0723] The server uses an NLP module to analyze the request and identify the request to search for a cafe.

[0724] The server uses the Google Maps API to search for cafe information and generates a response text.

[0725] The server performs multilingual translation and character editing, and then sends the text to the terminal.

[0726] The device converts text into speech using its speech synthesis module.

[0727] The device plays the generated audio and provides the user with cafe information.

[0728] This process allows users to quickly and accurately obtain information through voice input. Furthermore, multilingual support and character-based responses provide a more user-friendly experience.

[0729] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0730] Step 1:

[0731] The user says, "Find a nearby cafe."

[0732] Input: User's spoken voice

[0733] Specific action: The user speaks the desired information (in this case, information about a cafe) into the voice input device.

[0734] Step 2:

[0735] The device uses its built-in microphone to capture the user's voice and saves it as audio data.

[0736] Input: User's spoken voice

[0737] Output: Audio data file (WAV or MP3 format)

[0738] Specific operation: The device's microphone acquires audio as a digital signal, which is then converted and saved as an audio data file.

[0739] Step 3:

[0740] The device uses a speech recognition module (e.g., Google Speech-to-Text API) to convert the audio data into text data.

[0741] Input: Audio data file

[0742] Output: Text data "Find a nearby cafe"

[0743] Specific operation: The speech recognition module receives an audio data file as input, performs speech analysis, and converts it into text. The converted text data is saved to the device's memory.

[0744] Step 4:

[0745] The device generates request data based on the converted text data and location information obtained from its GPS function, and sends it to the server.

[0746] Input: Text data "Find a nearby cafe" and location information

[0747] Output: Request data (API request format)

[0748] Specific operation: Generate a request combining the converted text and location information, and send this request to the server using the HTTPS protocol.

[0749] Step 5:

[0750] The server receives the request data, uses a natural language processing (NLP) module to parse the text data, and identifies the user's request.

[0751] Input: Request data (text and location information)

[0752] Output: Analysis results (User request: Find a cafe)

[0753] Specific operation: The server analyzes the request and uses an NLP module to identify the request as "Find a cafe." This information is stored in the server's memory.

[0754] Step 6:

[0755] The server uses the Google Maps API based on the user's request to search for cafe information based on the user's current location.

[0756] Input: Analysis results (user request) and location information

[0757] Output: Cafe information list (cafe name, address, rating, etc.)

[0758] Specific operation: Call the Google Maps API to retrieve cafe information based on the current location. The retrieved information is stored in the server's memory.

[0759] Step 7:

[0760] The server generates a response text based on the cafe information it has acquired.

[0761] Input: Cafe information list

[0762] Output: Response text "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0763] Specific operation: Analyzes cafe information and generates response text in a user-friendly format.

[0764] Step 8:

[0765] The server translates the response text using a multilingual translation module as needed and edits the response text to suit the personality of the specific character.

[0766] Input: Response text

[0767] Output: Translated and edited response text

[0768] Specific actions: Use the Google Translate API to translate into multiple languages ​​and edit the text to suit the character's personality.

[0769] Step 9:

[0770] The server sends the edited and translated response text to the terminal.

[0771] Input: Response text (after translation and editing)

[0772] Output: Transmitted data

[0773] Specific operation: The edited and translated text data is sent to the terminal using the HTTPS protocol.

[0774] Step 10:

[0775] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[0776] Input: Response text

[0777] Output: Audio data file (MP3 format, etc.)

[0778] Specific operation: Text data is converted into speech data using a speech synthesis module such as Amazon Polly. The converted speech data is stored in the device's storage.

[0779] Step 11:

[0780] The device plays audio data through its speaker and responds to the user.

[0781] Input: Audio data file

[0782] Output: Physical audio output

[0783] Specific operation: The device's speaker is used to play the generated audio data, and the user is responded with, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0784] (Application Example 1)

[0785] 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."

[0786] In modern customer service, there is a demand for immediate answers to customer questions and provision of appropriate information within the store. However, conventional systems have difficulty providing information via voice input, and furthermore, multilingual support and responses using specific character voices are not adequately provided. As a result, the customer experience is limited, and improving service quality remains a challenge.

[0787] 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.

[0788] In this invention, the server includes means for capturing the user's voice, means for converting the captured voice into string data, means for analyzing the string data to identify the user's request, means for retrieving information based on the identified user's request, means for generating response text based on the retrieved information, means for converting the generated response text into audio data, means for playing the converted audio data back to the user, means for converting the generated response text into the voice of a specific virtual character, means for acquiring the user's current location information, and means for providing information about features within the store based on the acquired location information and the user's request. This enables immediate customer service within the store, and significantly improves the customer experience by enabling multilingual support and responses by specific characters.

[0789] "Means for capturing user voice" refers to a device for acquiring the voice emitted by a user as a digital signal using an audio input device such as a microphone.

[0790] "Methods for converting captured audio into text data" refers to the process of converting acquired audio data into corresponding text data using speech recognition technology.

[0791] "Means for analyzing the string data to identify the user's request" refers to a mechanism that uses natural language processing technology to understand and identify the user's intent and request from the string data.

[0792] "Means for retrieving information based on the requests of a specified user" refers to means for searching and extracting information from databases or the internet that meets the user's requirements.

[0793] "Means for generating response text based on searched information" refers to the process of creating an appropriate response to the user in text format based on search results.

[0794] "Means for converting generated response text into audio data" refers to a system that converts text into audio data using speech synthesis technology.

[0795] "Means for playing back converted audio data to the user" refers to a device for presenting the generated audio data to the user through a speaker or ear-hook type device.

[0796] "Means for converting generated response text into the voice of a specific virtual character" refers to the process of synthesizing generated text into speech using a pre-configured character and voice.

[0797] "Means for obtaining the user's current location information" refers to technologies that use GPS or other location information services to determine the user's current location.

[0798] "Means for providing information about in-store features based on acquired location information and user requests" refers to a system that provides specific product placement and guidance information within a store in accordance with user requests and location information.

[0799] System Overview

[0800] This invention is a system that retrieves information from a user's voice and returns a voice response. The system consists of various modules for realizing voice capture, speech recognition, natural language processing, information retrieval, response generation, speech synthesis, location information acquisition, and voice responses in the voice of a specific virtual character. The system uses a wearable device such as smart glasses to provide customer service in physical stores.

[0801] Hardware and software to be used

[0802] Hardware:

[0803] Smart glasses (voice input device, speaker)

[0804] GPS or location acquisition module

[0805] software:

[0806] Speech recognition modules (e.g., Google Speech-to-Text, Amazon Transcribe)

[0807] Natural Language Processing (NLP) modules (e.g., Google NLP API, OpenAI GPT-3)

[0808] Speech synthesis modules (e.g., Microsoft Azure TTS, Amazon Polly)

[0809] Data processing and data calculation

[0810] 1. Audio capture:

[0811] When a user speaks into the microphone attached to the smart glasses, audio data is captured. This data is stored digitally on the smart glasses.

[0812] 2. Speech recognition:

[0813] The captured audio data is sent to the speech recognition module in the smart glasses and converted into text data. This converts the audio into text format.

[0814] 3. Submit the request:

[0815] The converted text data is sent to the server along with the user's current location information. Here, the user's location information is obtained using a GPS module.

[0816] 4. Request Analysis:

[0817] The server analyzes the received request data and uses a natural language processing (NLP) module to identify the user's request.

[0818] 5. Information Retrieval:

[0819] Based on the specified request, the server searches for relevant information from databases and the internet.

[0820] 6. Response generation:

[0821] The server generates appropriate response text based on the search results. Furthermore, this response text can be translated into multiple languages ​​and edited to sound like a specific virtual character.

[0822] 7. Send a reply:

[0823] The response text is sent from the server to the smart glasses.

[0824] 8. Speech synthesis:

[0825] The speech synthesis module within the smart glasses converts the received response text into speech data. If the voice is converted to that of a specific character, a voice that reflects the character's characteristics will be generated.

[0826] 9. Audio Playback:

[0827] The generated audio data is played back to the user through the smart glasses' speaker.

[0828] Specific example

[0829] Situation: "Can you recommend a wine?"

[0830] 1. The user speaks to the smart glasses and says, "Please recommend a wine."

[0831] 2. The smart glasses capture the audio data and convert it into text data.

[0832] 3. The converted text data and current location information are sent to the server.

[0833] 4. The server analyzes the received data and identifies the request, "Please recommend a wine."

[0834] 5. The server retrieves recommended wine information from the database.

[0835] 6. Based on the information obtained by the server, it generates the response text: "The recommended wine is 'a specific wine'. It is available for purchase on the left side of the liquor section."

[0836] 7. Convert the generated text into the voice of a specific virtual character.

[0837] 8. The server sends the generated response text to the smart glasses.

[0838] 9. The smart glasses convert the speech into voice using a speech synthesis module and play it back to the user.

[0839] Example of a prompt:

[0840] When a user asks, "Can you recommend a wine?", the response text will be generated based on the following text:

[0841] Prompt: The customer asked, "Can you recommend a wine?" Please create a response text following the format below.

[0842] 1. Product Name

[0843] 2. A brief description of the sales location

[0844] Example response: "Our recommended wine is 'a specific wine.' It's located on the left side of the liquor section."

[0845] Text: "Please recommend a wine."

[0846] Thus, the invention provides a concrete means for implementing voice-based customer service in physical stores.

[0847] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0848] Step 1:

[0849] When a user speaks into the smart glasses, the smart glasses' microphone captures the voice and saves it as digital audio data. For example, the user might say, "Tell me your recommended wine." The input is an audio signal, and the output is digital audio data.

[0850] Step 2:

[0851] The speech recognition module on the device performs speech recognition using the captured audio data and converts it into text data. Specifically, the speech recognition module (e.g., Google Speech-to-Text) generates the text "Tell me your recommended wine" from the audio data. The input is audio data, and the output is text data.

[0852] Step 3:

[0853] The terminal obtains the converted text data and the user's current location information (such as GPS data), and sends a request containing this data to the server. Specifically, a location information acquisition module detects the current location and incorporates it into the transmission packet along with the text data. The input is text data and location information, and the output is request data.

[0854] Step 4:

[0855] The server analyzes the received request data and uses a natural language processing module (e.g., Google NLP API) to identify the user's request. Specifically, it analyzes the request "Tell me a wine recommendation" and identifies the request as "Wine Recommendations". The input is the request data, and the output is the analyzed request.

[0856] Step 5:

[0857] Based on the parsed request, the server searches for relevant information from the specified database and the internet. Specifically, it executes a database query to retrieve information about "recommended wines." The input is the parsed request, and the output is the search result.

[0858] Step 6:

[0859] The server generates response text based on the search results. Furthermore, it translates the generated text into multiple languages ​​as needed and edits it to the voice of a specific virtual character. For example, it might generate the text, "Our recommended wine is 'Specific Wine'. It's available on the left side of the liquor section," and edit it to the character's voice. The input is the search results, and the output is the response text.

[0860] Step 7:

[0861] The server sends the generated response text to the terminal. Specifically, it converts the response text into a packet and sends it to the terminal. The input is the response text, and the output is the transmitted packet.

[0862] Step 8:

[0863] The speech synthesis module on the terminal (e.g., Microsoft Azure TTS) converts the received response text into speech data. Specifically, it generates speech data from text data. The input is the response text, and the output is speech data.

[0864] Step 9:

[0865] The device plays audio data and provides a response to the user through the smart glasses' speaker. Specifically, it passes the generated audio data to the playback device and outputs the audio. The input is audio data, and the output is audio information provided to the user.

[0866] 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.

[0867] System Overview

[0868] This invention is a system that performs a series of processes, including capturing user voice, converting it into text data, analyzing it to identify requests, retrieving information, generating response text, and converting it back into audio data for playback to the user. Furthermore, this invention incorporates an emotion engine that recognizes user emotions, providing a function to improve the user experience. It also includes the ability to translate retrieved information into multiple languages ​​and edit it to suit the personality of a specific character.

[0869] Program processing flow

[0870] 1. Voice input (device)

[0871] The user says, "Find a nearby cafe."

[0872] The device uses the microphone to capture the user's voice and saves it as audio data.

[0873] 2. Voice recognition (device)

[0874] The device analyzes the voice data captured using its speech recognition module and converts it into text. For example, it might convert it to text like, "Find a nearby cafe."

[0875] 3. Emotion Recognition (Device)

[0876] The device uses an emotion engine to recognize the user's emotions (joy, anger, sadness, etc.) from voice and text data.

[0877] 4. Sending a request (from the terminal)

[0878] The device generates request data containing converted text data, user location information, and recognized sentiment data, and sends it to the server.

[0879] 5. Request parsing (server)

[0880] The server receives the request data. The request data includes the user's text, location information, and sentiment data.

[0881] 6. Information Retrieval (Server)

[0882] The server uses a natural language processing (NLP) module to parse the text portion of the request and identify the user's request. For example, it identifies the request "Find a cafe."

[0883] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[0884] 7. Response generation (server)

[0885] The server generates response text for the user based on the search results. For example, it might generate a sentence like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0886] The server adjusts the generated response text based on recognized emotion data. For example, if the user is angry, the tone of the response will be softened.

[0887] 8. Multilingual support and character editing (server)

[0888] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[0889] 9. Sending a response (server)

[0890] The server sends the generated and edited response text to the terminal.

[0891] 10. Speech synthesis (device)

[0892] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[0893] 11. Response playback (terminal)

[0894] The device plays audio data through its speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[0895] Specific example

[0896] Situation: "Look for a nearby cafe."

[0897] 1. The user says, "Find a nearby cafe."

[0898] 2. The device captures the audio and saves it as audio data.

[0899] 3. The device's voice recognition module converts the voice data into "Find a nearby cafe."

[0900] 4. The device uses an emotion engine to recognize the user's emotions and determines that they are "excited."

[0901] 5. The device generates a request containing the converted text, location information, and sentiment data, and sends it to the server.

[0902] 6. The server receives the request, parses it with the NLP module, and identifies the request as "Find a cafe".

[0903] 7. The server uses a map API to search for information about nearby cafes.

[0904] 8. The server generates the response text "There are 3 cafes nearby. We recommend 'Cafe Latte Street'." based on the search results.

[0905] 9. Based on the recognized emotion data (excited), the server adjusts the response text and generates it in a brighter tone.

[0906] 10. If necessary, the server will translate the response text into multiple languages ​​and edit it to suit the character's personality.

[0907] 11. The server sends a response text to the terminal.

[0908] 12. The terminal converts the response text into speech using a speech synthesis module.

[0909] 13. The device plays a voice message through its speaker, responding to the user in a cheerful tone, saying, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0910] In this way, the system allows users to enjoy conversations with their favorite characters while smoothly obtaining the necessary information. Furthermore, responses adapted to the user's emotions provide a more personalized experience.

[0911] The following describes the processing flow.

[0912] Step 1:

[0913] The user says, "Find a nearby cafe."

[0914] The device uses the microphone to capture audio and saves it as audio data.

[0915] Step 2:

[0916] The device uses a speech recognition module to analyze the captured audio data and convert it into text data such as "Find a nearby cafe."

[0917] Step 3:

[0918] The device uses an emotion engine to recognize the user's emotions from voice data and converted text data. For example, it might determine that the user is "excited" based on the tone and speed of their voice.

[0919] Step 4:

[0920] The device generates request data that includes converted text data, user location information, and recognized sentiment data. This request data contains the user's request, location information, and sentiment information.

[0921] Step 5:

[0922] The device sends the request data to the server via the internet.

[0923] Step 6:

[0924] The server receives the request data, which includes the user's text, location information, and sentiment data.

[0925] Step 7:

[0926] The server uses a natural language processing (NLP) module to analyze the text portion of the request and identify the user's request. For example, it might extract the request "Find a cafe."

[0927] Step 8:

[0928] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[0929] Step 9:

[0930] The server organizes the search results and generates response text for the user. For example, it might generate a response like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0931] Step 10:

[0932] The server adjusts the response text based on recognized sentiment data. For example, if the user is excited, the response tone becomes brighter and positive words are added.

[0933] Step 11:

[0934] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[0935] Step 12:

[0936] The server sends the generated and formatted response text to the terminal.

[0937] Step 13:

[0938] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[0939] Step 14:

[0940] The device plays audio data through its speaker and responds to the user in a cheerful tone, saying, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[0941] Through this series of processes, users can enjoy conversations with their favorite characters while quickly and effectively obtaining the information they need. Furthermore, emotion recognition provides more personalized responses, improving the user experience.

[0942] (Example 2)

[0943] 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".

[0944] Conventional voice dialogue systems struggled to generate flexible responses that reflected user emotions or to provide responses tailored to the individuality of specific characters. As a result, the user experience was limited, and it was difficult to address individual requests. Furthermore, multilingual support was insufficient, making it difficult to serve a global user base.

[0945] 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.

[0946] In this invention, the server includes means for acquiring user voice, means for converting the acquired voice into text data, means for analyzing the converted text data to identify the user's request, means for retrieving information based on the identified user's request, means for generating response data based on the retrieved information, means for converting the generated response data into voice data, means for providing the converted voice data to the user, means for recognizing the user's emotions, and means for adjusting the response data based on the recognized emotions. This enables flexible response generation based on the user's emotions, provision of responses tailored to specific characters, and multilingual support.

[0947] "Means of acquiring user voice" refers to devices and technologies that capture user voice and acquire it as data.

[0948] "Means of converting acquired audio into text data" refers to technologies and algorithms that convert audio data into string data.

[0949] "Means of analyzing converted text data to identify user requests" refers to technologies and methods that analyze text data to identify what the user wants.

[0950] "Means of retrieving information based on the requests of identified users" refers to technologies and systems that retrieve relevant information from databases and external services based on user requests.

[0951] "Means for generating response data based on retrieved information" refers to algorithms and technologies that generate appropriate responses for the user based on retrieved information.

[0952] "Means for converting generated response data into audio data" refers to technologies and devices that convert text-based response data into audio data.

[0953] "Means of providing the converted audio data to the user" refers to speakers or other audio output devices for playing the audio data to the user.

[0954] "Means of recognizing user emotions" refers to technologies and algorithms that identify emotions from user voice and text data.

[0955] "Means of adjusting response data based on recognized emotions" refers to technologies and methods that adjust the tone and content of responses according to the user's emotions.

[0956] "Means of translating into multiple languages" refers to technologies and systems that translate text data into different languages.

[0957] "Methods for editing to suit the characteristics of a specific character" refers to techniques and methods for modifying response data to match the personality and manner of speaking of a particular character.

[0958] This invention is a system that performs a series of processes: acquiring user voice, converting it into text data, analyzing it to identify requests, retrieving information, generating response data, converting it back into voice data, and providing it to the user. Furthermore, this invention provides a function to recognize the user's emotions and adjust the response data accordingly. It also includes a function to translate the retrieved information into multiple languages ​​and edit it to suit a specific character.

[0959] Hardware and software to be used

[0960] This system primarily uses the following hardware and software.

[0961] hardware

[0962] Device: A device equipped with a microphone and speaker for audio capture and playback (e.g., smartphone, tablet, smart speaker)

[0963] Server: A remote computer used for data processing and management.

[0964] software

[0965] Speech recognition module: An API for converting speech to text (e.g., Google Cloud Speech-to-Text API)

[0966] Emotion recognition module: A service for analyzing user emotions (e.g., Amazon Comprehend)

[0967] Natural Language Processing Module: A model for analyzing user requests (e.g., OpenAI GPT-3)

[0968] Map API: A service for searching for nearby information based on the user's location (e.g., Google Maps API).

[0969] Speech synthesis module: A technology for converting text into speech (e.g., Amazon Polly)

[0970] Translation Module: An API for translating response text into multiple languages ​​(e.g., Microsoft Translator).

[0971] Specific example

[0972] The following are examples of specific methods for implementing this invention.

[0973] Situation: "Looking for a nearby cafe"

[0974] 1. The user says, "Find a nearby cafe."

[0975] 2. The device uses the microphone to capture audio and saves it as audio data.

[0976] 3. The device uses a speech recognition module (Google Cloud Speech-to-Text API) to convert the captured audio into text: "Find a nearby cafe."

[0977] 4. The device uses an emotion recognition module (Amazon Comprehend) to analyze the voice and text data and determine that the user is "excited."

[0978] 5. The device generates a request containing text data, location information (GPS data), and sentiment data, and sends it to the server.

[0979] 6. The server receives the request, parses it using the NLP module (OpenAI GPT-3), and identifies the request as "Find a cafe."

[0980] 7. The server uses a map API (Google Maps API) to search for cafe information near the user's current location.

[0981] 8. Based on the search results, the server generates response data such as, "There are 3 cafes nearby. We recommend 'Cafe Latte Street'."

[0982] 9. Based on the recognized emotion data "excited," the server generates a response text in a cheerful tone.

[0983] 10. If necessary, the server will translate the response text into multiple languages ​​using a translation module (Microsoft Translator) and edit it to suit the characteristics of the specific character.

[0984] 11. The server sends the generated and edited response data to the terminal.

[0985] 12. The device uses a speech synthesis module (Amazon Polly) to convert the received response text into speech data.

[0986] 13. The device responds through the speaker, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[0987] Example of a prompt

[0988] "Generate responses based on the user's emotions when they are looking for a nearby cafe."

[0989] "Generate a response for when the user is angry."

[0990] As described above, this invention is a system that efficiently executes a series of processes, starting with user voice input, going through text conversion and emotion recognition, searching for appropriate information, and generating and providing responses. As a result, users can obtain flexible responses that respond to their emotions, multilingual support, and personalized experiences tailored to their characters.

[0991] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0992] Program processing flow

[0993] Step 1: Voice input (on the device)

[0994] input:

[0995] The user says, "Find a nearby cafe."

[0996] Specific operation:

[0997] The device's microphone captures the audio and saves it as audio data.

[0998] output:

[0999] The acquired audio data is saved.

[1000] Step 2: Voice Recognition (Device)

[1001] input:

[1002] Captured audio data.

[1003] Specific operation:

[1004] The device calls the Google Cloud Speech-to-Text API and uploads the audio data. The API analyzes the audio and retrieves the text "Find a nearby cafe."

[1005] output:

[1006] Text data converted from audio data: "Find a nearby cafe."

[1007] Step 3: Emotion Recognition (Device)

[1008] input:

[1009] Audio and text data: "Find a nearby cafe."

[1010] Specific operation:

[1011] The device uses Amazon Comprehend to analyze voice and text data. From the analysis results, it extracts emotions and determines that the user is "excited."

[1012] output:

[1013] The recognized emotion data is "excited".

[1014] Step 4: Sending the request (on the device)

[1015] input:

[1016] Text data "Looking for a nearby cafe," sentiment data "Excited," and current location obtained from GPS.

[1017] Specific operation:

[1018] The terminal generates request data containing this information and sends it to the server.

[1019] output:

[1020] Request data sent to the server.

[1021] Step 5: Request parsing (server)

[1022] input:

[1023] Request data received by the server.

[1024] Specific operation:

[1025] The server logs the request data to a log file. A natural language processing (NLP) module (OpenAI GPT-3) is used to analyze the text data "Find a nearby cafe". The user's request "Looking for a cafe" is identified from the analysis results.

[1026] output:

[1027] User request: "I'm looking for a cafe."

[1028] Step 6: Information Retrieval (Server)

[1029] input:

[1030] The user's request is "I'm looking for a cafe" and their location information is provided.

[1031] Specific operation:

[1032] The server accesses the Google Maps API to search for nearby cafes based on location information. It then parses the response from the API to retrieve relevant cafe information.

[1033] output:

[1034] The retrieved cafe information (e.g., 3 nearby cafes including "Cafe Latte Street").

[1035] Step 7: Response generation (server)

[1036] input:

[1037] The acquired cafe information and emotion data indicated "excited."

[1038] Specific operation:

[1039] The server generates a text response based on cafe information: "There are three cafes nearby. We recommend 'Cafe Latte Street'." Based on recognized sentiment data, the tone of the response text is adjusted to be brighter.

[1040] output:

[1041] Adjusted response text: "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1042] Step 8: Multilingual support and character editing (server)

[1043] input:

[1044] The generated response text reads, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1045] Specific operation:

[1046] If necessary, the server translates the response text into multiple languages ​​using a translation module. It then edits it to suit the characteristics of the specific character.

[1047] output:

[1048] Translated into multiple languages, response text tailored to the character's characteristics.

[1049] Step 9: Send response (server)

[1050] input:

[1051] Edited response text.

[1052] Specific operation:

[1053] The server sends the edited response text to the terminal.

[1054] output:

[1055] The response text sent to the terminal.

[1056] Step 10: Speech synthesis (device)

[1057] input:

[1058] The received response text.

[1059] Specific operation:

[1060] The device uses a speech synthesis module (Amazon Polly) to convert the response text into speech data.

[1061] output:

[1062] Audio data.

[1063] Step 11: Response playback (device)

[1064] input:

[1065] Generated audio data.

[1066] Specific operation:

[1067] The device plays audio data through its built-in speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[1068] output:

[1069] Voice response to the user.

[1070] (Application Example 2)

[1071] 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."

[1072] Conventional voice assistant systems are limited to simple information retrieval and responses to user requests, making it difficult to provide personalized responses that take into account the user's emotions and circumstances. Furthermore, they lack multilingual support and the ability to edit responses to match the character's personality, posing a challenge in providing an optimal user experience, particularly in virtual stores. Additionally, the system's ability to provide information considering the user's location is underdeveloped, making it difficult to deliver appropriate information to users in real time.

[1073] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1074] In this invention, the server includes means for recognizing the user's emotions, means for acquiring the user's location information, and means for searching relevant data based on the acquired user location information and generating response text. This makes it possible to provide personalized responses in real time that are tailored to the user's emotions and current situation. In addition, by incorporating multilingual support and response editing functions that match character personalities, a more natural and engaging user experience can be achieved.

[1075] A "user" refers to a person who uses the system.

[1076] "Means for capturing speech" refers to a device or technology for acquiring a user's speech as audio data.

[1077] "Means of converting to string data" refers to technology or equipment for converting audio data into text data.

[1078] "Means of analysis to identify user requests" refers to technology or devices that process converted text data to identify what the user is asking for.

[1079] "Means of recognizing user emotions" refers to technology or devices that determine a user's emotions (e.g., joy, anger, sadness, etc.) from their voice data or speech content.

[1080] "Means of retrieving information" refers to technologies or devices for finding appropriate information based on the user's needs and feelings.

[1081] "Means for generating response text" refers to technology or equipment for creating text that serves as a response to a user based on retrieved information.

[1082] "Means of converting to audio data" refers to technology or equipment for converting generated response text into audio output.

[1083] "Means for playing audio data" refers to the technology or device used to allow the user to listen to the converted audio data.

[1084] "Means of translation into multiple languages" refers to technologies or devices that translate generated response text into different languages.

[1085] "Means of editing to match a character's personality" refers to techniques or devices that adjust generated response text to suit a specific character.

[1086] "Means for acquiring location information" refers to technology or devices used to determine the user's current location.

[1087] "Means for searching relevant data and generating response text" refers to a technology or device that searches for appropriate information based on acquired location information and user requests and creates response text.

[1088] This invention relates to a system that captures user voice, converts it into text data, analyzes it to identify requests, retrieves information, generates response text, converts it back into audio data, and plays it back to the user. Furthermore, this system has the capability to recognize user emotions and provides personalized responses according to the user's emotions. The detailed configuration and operation of the system are described below.

[1089] Hardware and software configuration

[1090] This system primarily consists of user terminals and servers. User terminals include devices such as smartphones, smart glasses, and head-mounted displays (HMDs), and these devices include the following software modules.

[1091] 1. Voice Capture Module: Captures the user's voice using the device's microphone.

[1092] 2. Speech Recognition Module: Converts speech data into text data. Specifically, it uses the Google Cloud Speech-to-Text API.

[1093] 3. Emotion Recognition Module: Recognizes emotions from the user's voice and text data. Specifically, it uses the Microsoft Azure Emotion API.

[1094] 4. Location Information Acquisition Module: Acquires the current location information using the GPS function of the user's device.

[1095] The server side includes the following software modules:

[1096] 1. Natural Language Processing (NLP) Module: Uses the Google Cloud Natural Language API to analyze and identify user requests.

[1097] 2. Information Retrieval Module: Searches for appropriate information based on user requests and location information.

[1098] 3. Response Generation Module: Generates response text based on the searched information and the user's sentiment.

[1099] 4. Multilingual Translation Module: Translates response text into multiple languages ​​as needed.

[1100] 5. Character Editing Module: Edit response text to match the personality of a specific character.

[1101] System Operation Description

[1102] When a user says, "Tell me your recommended shampoo," the smart device's microphone captures the voice and saves it as audio data. Then, a speech recognition module (Google Cloud Speech-to-Text API) is used to convert the captured audio data into text, "Tell me your recommended shampoo." At the same time, an emotion recognition module (Microsoft Azure Emotion API) recognizes the user's emotion and determines that they are "interested."

[1103] Request data is generated, containing the converted text data, user location information, and recognized sentiment data, and sent to the server. The server uses a natural language processing (NLP) module to parse the request and identify the request for "shampoo recommendations."

[1104] The server uses an information retrieval module to search the database for relevant data and generates the response text "We recommend 'Shampoo Excellence'." Based on the recognized sentiment data, the response text is edited and presented in a user-friendly tone. Additionally, a multilingual translation module is used to translate the text as needed, and a character editing module is used to edit it to match the character's personality.

[1105] The device receives the response text and converts it into speech data using a text-to-speech (TTS) module (Amazon Polly). It then plays the audio through the smart device's speaker, telling the user, "We recommend 'Shampoo Excellence'."

[1106] Specific examples and prompt statements

[1107] A concrete example of this scenario would be a series of actions taken by a smart device in response to a user's request, such as "Tell me your recommended shampoo." The following is an example of a prompt to a generative AI model.

[1108] Example of a prompt:

[1109] You are a customer service assistant in a virtual store. Generate a response when a user asks for a "recommended shampoo." Respond in a friendly and positive tone, especially if the user shows interest. Remember to mention specific product names.

[1110] In this way, the present invention can provide personalized information in response to user requests and improve the user experience.

[1111] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1112] Step 1:

[1113] Voice input

[1114] User: "Can you recommend a shampoo?"

[1115] Device: Captures the user's voice using the microphone and saves it as audio data.

[1116] Input: User's speech (audio data)

[1117] Output: Audio data

[1118] Step 2:

[1119] Speech recognition

[1120] Terminal: Analyzes captured audio data using a speech recognition module (Google Cloud Speech-to-Text API) and converts it to text.

[1121] Input: Audio data

[1122] Output: Text data "Please recommend a shampoo"

[1123] Step 3:

[1124] emotion recognition

[1125] Terminal: Uses an emotion recognition module (Microsoft Azure Emotion API) to recognize the user's emotions from voice and text data and determine that they are "interested."

[1126] Input: Audio data, text data

[1127] Output: Sentiment data "Interested"

[1128] Step 4:

[1129] Send Request

[1130] Terminal: Generates request data including converted text data, user location information, and recognized sentiment data, and sends it to the server.

[1131] Input: Text data "Please recommend a shampoo," location information, sentiment data "Interested"

[1132] Output: Request data

[1133] Step 5:

[1134] Request Analysis

[1135] Server: Receives request data, uses a natural language processing (NLP) module to parse the text portion and identify the user's request. Identifies the request as "shampoo recommendation".

[1136] Input: Request data

[1137] Output: User request "Shampoo recommendation"

[1138] Step 6:

[1139] Information Retrieval

[1140] Server: Retrieves information on relevant products from the database based on the identified request and location information.

[1141] Input: User request "Shampoo recommendation", location information

[1142] Output: Search result "Recommended shampoo: 'Shampoo Excellence'"

[1143] Step 7:

[1144] Response generation

[1145] Server: Based on the obtained search results and sentiment data, it generates response text for the user. It generates "We recommend 'Shampoo Excellence'."

[1146] Input: Search results, sentiment data "Interested"

[1147] Output: Response text: "We recommend 'Shampoo Excellence'."

[1148] Step 8:

[1149] Multilingual support and character editing

[1150] Server: Translates response text into multiple languages ​​as needed and edits it to suit the character's personality.

[1151] Input: Response text: "My recommendation is 'Shampoo Excellence'."

[1152] Output: Edited response text

[1153] Step 9:

[1154] Send response

[1155] Server: Sends the edited response text to the terminal.

[1156] Input: Edited response text

[1157] Output: Response text received by the terminal

[1158] Step 10:

[1159] Speech synthesis

[1160] Terminal: Converts the received response text into speech data using a text-to-speech (TTS) module (Amazon Polly).

[1161] Input: Response text: "My recommendation is 'Shampoo Excellence'."

[1162] Output: Audio data

[1163] Step 11:

[1164] Response playback

[1165] Device: Plays audio data through the speaker and tells the user, "We recommend 'Shampoo Excellence'."

[1166] Input: Audio data

[1167] Output: Voice response to the user: "We recommend 'Shampoo Excellence'."

[1168] 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.

[1169] 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.

[1170] 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.

[1171] [Third Embodiment]

[1172] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[1173] 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.

[1174] 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).

[1175] 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.

[1176] 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.

[1177] 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).

[1178] 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.

[1179] 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.

[1180] 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.

[1181] 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.

[1182] 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.

[1183] 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".

[1184] This invention relates to a system that accurately captures a user's voice and uses that voice for navigation and information retrieval. The following describes a detailed configuration for implementing this system.

[1185] System Overview

[1186] This system performs a series of processes: capturing the user's voice, converting it into text data, analyzing it to identify the request, searching for information to generate response text, converting it back into audio data, and playing it back to the user. Furthermore, it includes functions to translate information into multiple languages ​​and to edit response text to match the personality of a specific character.

[1187] Program processing flow

[1188] 1. Voice input (device)

[1189] The user says, "Find a nearby cafe."

[1190] The device uses the microphone to capture the user's voice and saves it as audio data.

[1191] 2. Voice recognition (device)

[1192] The device uses a speech recognition module to analyze the voice data and convert it into text, "Find a nearby cafe."

[1193] 3. Sending a request (from the terminal)

[1194] The terminal generates a request containing the converted text data and the user's current location information, and sends it to the server.

[1195] 4. Request parsing (server)

[1196] The server receives the request data, uses a natural language processing (NLP) module to analyze the text, and identifies the request as "Find a cafe."

[1197] 5. Information Retrieval (Server)

[1198] The server searches for relevant cafe information from the internet and databases based on the specified request. For example, it might use a map API to identify nearby cafes.

[1199] 6. Response generation (server)

[1200] The server generates response text for the user based on the search results. For example, it might generate, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1201] 7. Multilingual support and character editing (server)

[1202] The server translates response text into multiple languages ​​as needed. It also edits response text to suit the personality of specific characters.

[1203] 8. Sending a response (server)

[1204] The server sends the generated response text to the terminal.

[1205] 9. Speech synthesis (device)

[1206] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[1207] 10. Response playback (terminal)

[1208] The device plays audio data through its speaker and responds to the user, saying, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[1209] Specific example

[1210] Situation: "Look for a nearby cafe."

[1211] 1. The user says, "Find a nearby cafe."

[1212] 2. The device captures the audio and saves it as audio data.

[1213] 3. The device's voice recognition module converts the voice data into "Find a nearby cafe."

[1214] 4. The device generates a request containing the converted text and current location information and sends it to the server.

[1215] 5. The server receives the request, parses it with the NLP module, and identifies the request as "Find a cafe".

[1216] 6. The server searches for information about cafes related to it using a map API.

[1217] 7. The server generates the response text "There are 3 cafes nearby. We recommend 'Cafe Latte Street'." based on the search results.

[1218] 8. If necessary, the server will translate the response text into multiple languages ​​and edit it to suit the character's personality.

[1219] 9. The server sends a response text to the terminal.

[1220] 10. The terminal converts the response text into speech using a speech synthesis module.

[1221] 11. The device plays a voice message through its speaker, responding to the user with, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1222] In this way, the system allows users to enjoy conversations with their favorite characters while obtaining the information they need.

[1223] The following describes the processing flow.

[1224] Step 1:

[1225] The user says, "Find a nearby cafe."

[1226] The device uses the microphone to capture this audio and saves it as audio data.

[1227] Step 2:

[1228] The device analyzes the voice data captured using its speech recognition module and converts it into text. For example, it might convert it to text like, "Find a nearby cafe."

[1229] Step 3:

[1230] The terminal generates request data that includes generated text data and user location information. This request data contains the user's request and location information.

[1231] Step 4:

[1232] The device sends the request data to the server via the internet.

[1233] Step 5:

[1234] The server receives the request data. The request data includes text and location information from the user.

[1235] Step 6:

[1236] The server uses a natural language processing (NLP) module to analyze the text portion of the request and identify the user's request. For example, it might extract the phrase "find a cafe."

[1237] Step 7:

[1238] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[1239] Step 8:

[1240] The server organizes the search results and generates response text for the user. For example, it might generate a sentence like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1241] Step 9:

[1242] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[1243] Step 10:

[1244] The server sends the generated and edited response text to the terminal.

[1245] Step 11:

[1246] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[1247] Step 12:

[1248] The device plays audio data through its speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[1249] In this way, the system allows users to enjoy conversations with their favorite characters while smoothly obtaining the information they need.

[1250] (Example 1)

[1251] 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."

[1252] Conventional speech recognition systems have had the problem of difficulty in accurately capturing user voices and providing information that meets the user's requests. Furthermore, they are limited to single-language support and cannot provide responses tailored to the personality of a character, resulting in a limited user experience. This invention aims to solve these problems and provide a more interactive and multi-functional information provision system.

[1253] 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.

[1254] In this invention, the server includes means for analyzing a request to identify the user's request, means for retrieving information based on the identified user's request, and means for generating response text based on the retrieved information. This makes it possible to analyze the user's request with high accuracy and provide appropriate information quickly.

[1255] "User" refers to a person who uses a system or an end-user.

[1256] "Audio capture" is the process of recording audio signals using a microphone.

[1257] "Converting to string data" refers to converting audio data into text format using speech recognition technology.

[1258] A "request" refers to a data packet containing a user's request or command, which is sent to the server.

[1259] "Location information" refers to geographical coordinate information obtained using GPS or other positioning technologies.

[1260] A "server" refers to a computer system that processes data and provides services to clients via a network.

[1261] "Natural language processing" refers to the technology that enables computers to understand, generate, and analyze human language.

[1262] "Information retrieval" is the process of obtaining relevant information from databases or the internet based on user requests.

[1263] "Response text" refers to a text-based representation of a user's request or answer to a user's request.

[1264] "Translating into multiple languages" is the process of converting text written in one language into another language.

[1265] "Editing to match the character's personality" refers to modifying text to reflect the characteristics and speaking style of a specific character.

[1266] "Converting to audio data" refers to the process of changing text information into computer-generated speech.

[1267] "Audio playback" refers to the physical output of audio data using sound devices such as speakers.

[1268] This invention is a system that captures user voice and uses it for navigation and information retrieval. The following describes a detailed configuration for implementing this system.

[1269] General overview

[1270] This system includes a voice input device, a speech recognition module, a data communication module, an information retrieval module, a natural language processing module, a response generation module, and a speech synthesis module. Through a series of processes, it is possible to analyze and respond to user voice requests. In addition, it also provides multilingual translation and character-based response generation capabilities.

[1271] Hardware and software to be used

[1272] Device: Mobile information terminals such as smartphones and tablets

[1273] Microphone: Built-in microphone or external microphone device

[1274] Servers: Cloud-based servers, such as AWS or Google Cloud

[1275] Speech recognition module: Google Speech-to-Text API

[1276] Natural language processing modules: SpaCy and Google NLP API

[1277] Information retrieval module: Google Maps API or any database

[1278] Multilingual translation module: Google Translate API

[1279] Speech synthesis module: Amazon Polly

[1280] Processing flow

[1281] 1. The user speaks through a voice input device, for example, saying, "Find a nearby cafe."

[1282] 2. The device's built-in microphone captures this audio and saves it as audio data. This data is temporarily stored in the device's storage.

[1283] 3. The device uses a speech recognition module to convert this speech data into text data. The converted text will read, "Find a nearby cafe."

[1284] 4. The terminal generates the converted text and location information as request data and sends it to the server. This includes precise location information using GPS and other positioning technologies.

[1285] 5. The server receives the request data and parses the text using a natural language processing module. This analysis identifies the user's request (e.g., "find a cafe").

[1286] 6. The server uses the Google Maps API to search for nearby cafes based on the user's request. The retrieved information includes details such as the cafe's name, address, and rating.

[1287] 7. The server generates response text based on the search results. For example, it might generate something like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1288] 8. If necessary, the server will use a multilingual translation module to translate response text. It may also edit the text to suit the character's personality.

[1289] 9. The server sends the edited and translated text to the terminal.

[1290] 10. The terminal converts the received text into speech data using a speech synthesis module.

[1291] 11. Finally, the device plays this audio data to the user, providing the desired information verbally.

[1292] Specific example

[1293] Let's take the situation "finding a nearby cafe" as an example.

[1294] The user says, "Find a nearby cafe."

[1295] The device captures the audio and converts it to text using a speech recognition module.

[1296] The device sends the converted text and location information to the server.

[1297] The server uses an NLP module to analyze the request and identify the request to search for a cafe.

[1298] The server uses the Google Maps API to search for cafe information and generates a response text.

[1299] The server performs multilingual translation and character editing, and then sends the text to the terminal.

[1300] The device converts text into speech using its speech synthesis module.

[1301] The device plays the generated audio and provides the user with cafe information.

[1302] This process allows users to quickly and accurately obtain information through voice input. Furthermore, multilingual support and character-based responses provide a more user-friendly experience.

[1303] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1304] Step 1:

[1305] The user says, "Find a nearby cafe."

[1306] Input: User's spoken voice

[1307] Specific action: The user speaks the desired information (in this case, information about a cafe) into the voice input device.

[1308] Step 2:

[1309] The device uses its built-in microphone to capture the user's voice and saves it as audio data.

[1310] Input: User's spoken voice

[1311] Output: Audio data file (WAV or MP3 format)

[1312] Specific operation: The device's microphone acquires audio as a digital signal, which is then converted and saved as an audio data file.

[1313] Step 3:

[1314] The device uses a speech recognition module (e.g., Google Speech-to-Text API) to convert the audio data into text data.

[1315] Input: Audio data file

[1316] Output: Text data "Find a nearby cafe"

[1317] Specific operation: The speech recognition module receives an audio data file as input, performs speech analysis, and converts it into text. The converted text data is saved to the device's memory.

[1318] Step 4:

[1319] The device generates request data based on the converted text data and location information obtained from its GPS function, and sends it to the server.

[1320] Input: Text data "Find a nearby cafe" and location information

[1321] Output: Request data (API request format)

[1322] Specific operation: Generate a request combining the converted text and location information, and send this request to the server using the HTTPS protocol.

[1323] Step 5:

[1324] The server receives the request data, uses a natural language processing (NLP) module to parse the text data, and identifies the user's request.

[1325] Input: Request data (text and location information)

[1326] Output: Analysis results (User request: Find a cafe)

[1327] Specific operation: The server analyzes the request and uses an NLP module to identify the request as "Find a cafe." This information is stored in the server's memory.

[1328] Step 6:

[1329] The server uses the Google Maps API based on the user's request to search for cafe information based on the user's current location.

[1330] Input: Analysis results (user request) and location information

[1331] Output: Cafe information list (cafe name, address, rating, etc.)

[1332] Specific operation: Call the Google Maps API to retrieve cafe information based on the current location. The retrieved information is stored in the server's memory.

[1333] Step 7:

[1334] The server generates a response text based on the cafe information it has acquired.

[1335] Input: Cafe information list

[1336] Output: Response text "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1337] Specific operation: Analyzes cafe information and generates response text in a user-friendly format.

[1338] Step 8:

[1339] The server translates the response text using a multilingual translation module as needed and edits the response text to suit the personality of the specific character.

[1340] Input: Response text

[1341] Output: Translated and edited response text

[1342] Specific actions: Use the Google Translate API to translate into multiple languages ​​and edit the text to suit the character's personality.

[1343] Step 9:

[1344] The server sends the edited and translated response text to the terminal.

[1345] Input: Response text (after translation and editing)

[1346] Output: Transmitted data

[1347] Specific operation: The edited and translated text data is sent to the terminal using the HTTPS protocol.

[1348] Step 10:

[1349] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[1350] Input: Response text

[1351] Output: Audio data file (MP3 format, etc.)

[1352] Specific operation: Text data is converted into speech data using a speech synthesis module such as Amazon Polly. The converted speech data is stored in the device's storage.

[1353] Step 11:

[1354] The device plays audio data through its speaker and responds to the user.

[1355] Input: Audio data file

[1356] Output: Physical audio output

[1357] Specific operation: The device's speaker is used to play the generated audio data, and the user is responded with, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1358] (Application Example 1)

[1359] 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."

[1360] In modern customer service, there is a demand for immediate answers to customer questions and provision of appropriate information within the store. However, conventional systems have difficulty providing information via voice input, and furthermore, multilingual support and responses using specific character voices are not adequately provided. As a result, the customer experience is limited, and improving service quality remains a challenge.

[1361] 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.

[1362] In this invention, the server includes means for capturing the user's voice, means for converting the captured voice into string data, means for analyzing the string data to identify the user's request, means for retrieving information based on the identified user's request, means for generating response text based on the retrieved information, means for converting the generated response text into audio data, means for playing the converted audio data back to the user, means for converting the generated response text into the voice of a specific virtual character, means for acquiring the user's current location information, and means for providing information about features within the store based on the acquired location information and the user's request. This enables immediate customer service within the store, and significantly improves the customer experience by enabling multilingual support and responses by specific characters.

[1363] "Means for capturing user voice" refers to a device for acquiring the voice emitted by a user as a digital signal using an audio input device such as a microphone.

[1364] "Methods for converting captured audio into text data" refers to the process of converting acquired audio data into corresponding text data using speech recognition technology.

[1365] "Means for analyzing the string data to identify the user's request" refers to a mechanism that uses natural language processing technology to understand and identify the user's intent and request from the string data.

[1366] "Means for retrieving information based on the requests of a specified user" refers to means for searching and extracting information from databases or the internet that meets the user's requirements.

[1367] "Means for generating response text based on searched information" refers to the process of creating an appropriate response to the user in text format based on search results.

[1368] "Means for converting generated response text into audio data" refers to a system that converts text into audio data using speech synthesis technology.

[1369] "Means for playing back converted audio data to the user" refers to a device for presenting the generated audio data to the user through a speaker or ear-hook type device.

[1370] "Means for converting generated response text into the voice of a specific virtual character" refers to the process of synthesizing generated text into speech using a pre-configured character and voice.

[1371] "Means for obtaining the user's current location information" refers to technologies that use GPS or other location information services to determine the user's current location.

[1372] "Means for providing information about in-store features based on acquired location information and user requests" refers to a system that provides specific product placement and guidance information within a store in accordance with user requests and location information.

[1373] System Overview

[1374] This invention is a system that retrieves information from a user's voice and returns a voice response. The system consists of various modules for realizing voice capture, speech recognition, natural language processing, information retrieval, response generation, speech synthesis, location information acquisition, and voice responses in the voice of a specific virtual character. The system uses a wearable device such as smart glasses to provide customer service in physical stores.

[1375] Hardware and software to be used

[1376] Hardware:

[1377] Smart glasses (voice input device, speaker)

[1378] GPS or location acquisition module

[1379] software:

[1380] Speech recognition modules (e.g., Google Speech-to-Text, Amazon Transcribe)

[1381] Natural Language Processing (NLP) modules (e.g., Google NLP API, OpenAI GPT-3)

[1382] Speech synthesis modules (e.g., Microsoft Azure TTS, Amazon Polly)

[1383] Data processing and data calculation

[1384] 1. Audio capture:

[1385] When a user speaks into the microphone attached to the smart glasses, audio data is captured. This data is stored digitally on the smart glasses.

[1386] 2. Speech recognition:

[1387] The captured audio data is sent to the speech recognition module in the smart glasses and converted into text data. This converts the audio into text format.

[1388] 3. Submit the request:

[1389] The converted text data is sent to the server along with the user's current location information. Here, the user's location information is obtained using a GPS module.

[1390] 4. Request Analysis:

[1391] The server analyzes the received request data and uses a natural language processing (NLP) module to identify the user's request.

[1392] 5. Information Retrieval:

[1393] Based on the specified request, the server searches for relevant information from databases and the internet.

[1394] 6. Response generation:

[1395] The server generates appropriate response text based on the search results. Furthermore, this response text can be translated into multiple languages ​​and edited to sound like a specific virtual character.

[1396] 7. Send a reply:

[1397] The response text is sent from the server to the smart glasses.

[1398] 8. Speech synthesis:

[1399] The speech synthesis module within the smart glasses converts the received response text into speech data. If the voice is converted to that of a specific character, a voice that reflects the character's characteristics will be generated.

[1400] 9. Audio Playback:

[1401] The generated audio data is played back to the user through the smart glasses' speaker.

[1402] Specific example

[1403] Situation: "Can you recommend a wine?"

[1404] 1. The user speaks to the smart glasses and says, "Please recommend a wine."

[1405] 2. The smart glasses capture the audio data and convert it into text data.

[1406] 3. The converted text data and current location information are sent to the server.

[1407] 4. The server analyzes the received data and identifies the request, "Please recommend a wine."

[1408] 5. The server retrieves recommended wine information from the database.

[1409] 6. Based on the information obtained by the server, it generates the response text: "The recommended wine is 'a specific wine'. It is available for purchase on the left side of the liquor section."

[1410] 7. Convert the generated text into the voice of a specific virtual character.

[1411] 8. The server sends the generated response text to the smart glasses.

[1412] 9. The smart glasses convert the speech into voice using a speech synthesis module and play it back to the user.

[1413] Example of a prompt:

[1414] When a user asks, "Can you recommend a wine?", the response text will be generated based on the following text:

[1415] Prompt: The customer asked, "Can you recommend a wine?" Please create a response text following the format below.

[1416] 1. Product Name

[1417] 2. A brief description of the sales location

[1418] Example response: "Our recommended wine is 'a specific wine.' It's located on the left side of the liquor section."

[1419] Text: "Please recommend a wine."

[1420] Thus, the invention provides a concrete means for implementing voice-based customer service in physical stores.

[1421] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1422] Step 1:

[1423] When a user speaks into the smart glasses, the smart glasses' microphone captures the voice and saves it as digital audio data. For example, the user might say, "Tell me your recommended wine." The input is an audio signal, and the output is digital audio data.

[1424] Step 2:

[1425] The speech recognition module on the device performs speech recognition using the captured audio data and converts it into text data. Specifically, the speech recognition module (e.g., Google Speech-to-Text) generates the text "Tell me your recommended wine" from the audio data. The input is audio data, and the output is text data.

[1426] Step 3:

[1427] The terminal obtains the converted text data and the user's current location information (such as GPS data), and sends a request containing this data to the server. Specifically, a location information acquisition module detects the current location and incorporates it into the transmission packet along with the text data. The input is text data and location information, and the output is request data.

[1428] Step 4:

[1429] The server analyzes the received request data and uses a natural language processing module (e.g., Google NLP API) to identify the user's request. Specifically, it analyzes the request "Tell me a wine recommendation" and identifies the request as "Wine Recommendations". The input is the request data, and the output is the analyzed request.

[1430] Step 5:

[1431] Based on the parsed request, the server searches for relevant information from the specified database and the internet. Specifically, it executes a database query to retrieve information about "recommended wines." The input is the parsed request, and the output is the search result.

[1432] Step 6:

[1433] The server generates response text based on the search results. Furthermore, it translates the generated text into multiple languages ​​as needed and edits it to the voice of a specific virtual character. For example, it might generate the text, "Our recommended wine is 'Specific Wine'. It's available on the left side of the liquor section," and edit it to the character's voice. The input is the search results, and the output is the response text.

[1434] Step 7:

[1435] The server sends the generated response text to the terminal. Specifically, it converts the response text into a packet and sends it to the terminal. The input is the response text, and the output is the transmitted packet.

[1436] Step 8:

[1437] The speech synthesis module on the terminal (e.g., Microsoft Azure TTS) converts the received response text into speech data. Specifically, it generates speech data from text data. The input is the response text, and the output is speech data.

[1438] Step 9:

[1439] The device plays audio data and provides a response to the user through the smart glasses' speaker. Specifically, it passes the generated audio data to the playback device and outputs the audio. The input is audio data, and the output is audio information provided to the user.

[1440] 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.

[1441] System Overview

[1442] This invention is a system that performs a series of processes, including capturing user voice, converting it into text data, analyzing it to identify requests, retrieving information, generating response text, and converting it back into audio data for playback to the user. Furthermore, this invention incorporates an emotion engine that recognizes user emotions, providing a function to improve the user experience. It also includes the ability to translate retrieved information into multiple languages ​​and edit it to suit the personality of a specific character.

[1443] Program processing flow

[1444] 1. Voice input (device)

[1445] The user says, "Find a nearby cafe."

[1446] The device uses the microphone to capture the user's voice and saves it as audio data.

[1447] 2. Voice recognition (device)

[1448] The device analyzes the voice data captured using its speech recognition module and converts it into text. For example, it might convert it to text like, "Find a nearby cafe."

[1449] 3. Emotion Recognition (Device)

[1450] The device uses an emotion engine to recognize the user's emotions (joy, anger, sadness, etc.) from voice and text data.

[1451] 4. Sending a request (from the terminal)

[1452] The device generates request data containing converted text data, user location information, and recognized sentiment data, and sends it to the server.

[1453] 5. Request parsing (server)

[1454] The server receives the request data. The request data includes the user's text, location information, and sentiment data.

[1455] 6. Information Retrieval (Server)

[1456] The server uses a natural language processing (NLP) module to parse the text portion of the request and identify the user's request. For example, it identifies the request "Find a cafe."

[1457] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[1458] 7. Response generation (server)

[1459] The server generates response text for the user based on the search results. For example, it might generate a sentence like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1460] The server adjusts the generated response text based on recognized emotion data. For example, if the user is angry, the tone of the response will be softened.

[1461] 8. Multilingual support and character editing (server)

[1462] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[1463] 9. Sending a response (server)

[1464] The server sends the generated and edited response text to the terminal.

[1465] 10. Speech synthesis (device)

[1466] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[1467] 11. Response playback (terminal)

[1468] The device plays audio data through its speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[1469] Specific example

[1470] Situation: "Look for a nearby cafe."

[1471] 1. The user says, "Find a nearby cafe."

[1472] 2. The device captures the audio and saves it as audio data.

[1473] 3. The device's voice recognition module converts the voice data into "Find a nearby cafe."

[1474] 4. The device uses an emotion engine to recognize the user's emotions and determines that they are "excited."

[1475] 5. The device generates a request containing the converted text, location information, and sentiment data, and sends it to the server.

[1476] 6. The server receives the request, parses it with the NLP module, and identifies the request as "Find a cafe".

[1477] 7. The server uses a map API to search for information about nearby cafes.

[1478] 8. The server generates the response text "There are 3 cafes nearby. We recommend 'Cafe Latte Street'." based on the search results.

[1479] 9. Based on the recognized emotion data (excited), the server adjusts the response text and generates it in a brighter tone.

[1480] 10. If necessary, the server will translate the response text into multiple languages ​​and edit it to suit the character's personality.

[1481] 11. The server sends a response text to the terminal.

[1482] 12. The terminal converts the response text into speech using a speech synthesis module.

[1483] 13. The device plays a voice message through its speaker, responding to the user in a cheerful tone, saying, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1484] In this way, the system allows users to enjoy conversations with their favorite characters while smoothly obtaining the necessary information. Furthermore, responses adapted to the user's emotions provide a more personalized experience.

[1485] The following describes the processing flow.

[1486] Step 1:

[1487] The user says, "Find a nearby cafe."

[1488] The device uses the microphone to capture audio and saves it as audio data.

[1489] Step 2:

[1490] The device uses a speech recognition module to analyze the captured audio data and convert it into text data such as "Find a nearby cafe."

[1491] Step 3:

[1492] The device uses an emotion engine to recognize the user's emotions from voice data and converted text data. For example, it might determine that the user is "excited" based on the tone and speed of their voice.

[1493] Step 4:

[1494] The device generates request data that includes converted text data, user location information, and recognized sentiment data. This request data contains the user's request, location information, and sentiment information.

[1495] Step 5:

[1496] The device sends the request data to the server via the internet.

[1497] Step 6:

[1498] The server receives the request data, which includes the user's text, location information, and sentiment data.

[1499] Step 7:

[1500] The server uses a natural language processing (NLP) module to analyze the text portion of the request and identify the user's request. For example, it might extract the request "Find a cafe."

[1501] Step 8:

[1502] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[1503] Step 9:

[1504] The server organizes the search results and generates response text for the user. For example, it might generate a response like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1505] Step 10:

[1506] The server adjusts the response text based on recognized sentiment data. For example, if the user is excited, the response tone becomes brighter and positive words are added.

[1507] Step 11:

[1508] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[1509] Step 12:

[1510] The server sends the generated and formatted response text to the terminal.

[1511] Step 13:

[1512] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[1513] Step 14:

[1514] The device plays audio data through its speaker and responds to the user in a cheerful tone, saying, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[1515] Through this series of processes, users can enjoy conversations with their favorite characters while quickly and effectively obtaining the information they need. Furthermore, emotion recognition provides more personalized responses, improving the user experience.

[1516] (Example 2)

[1517] 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."

[1518] Conventional voice dialogue systems struggled to generate flexible responses that reflected user emotions or to provide responses tailored to the individuality of specific characters. As a result, the user experience was limited, and it was difficult to address individual requests. Furthermore, multilingual support was insufficient, making it difficult to serve a global user base.

[1519] 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.

[1520] In this invention, the server includes means for acquiring user voice, means for converting the acquired voice into text data, means for analyzing the converted text data to identify the user's request, means for retrieving information based on the identified user's request, means for generating response data based on the retrieved information, means for converting the generated response data into voice data, means for providing the converted voice data to the user, means for recognizing the user's emotions, and means for adjusting the response data based on the recognized emotions. This enables flexible response generation based on the user's emotions, provision of responses tailored to specific characters, and multilingual support.

[1521] "Means of acquiring user voice" refers to devices and technologies that capture user voice and acquire it as data.

[1522] "Means of converting acquired audio into text data" refers to technologies and algorithms that convert audio data into string data.

[1523] "Means of analyzing converted text data to identify user requests" refers to technologies and methods that analyze text data to identify what the user wants.

[1524] "Means of retrieving information based on the requests of identified users" refers to technologies and systems that retrieve relevant information from databases and external services based on user requests.

[1525] "Means for generating response data based on retrieved information" refers to algorithms and technologies that generate appropriate responses for the user based on retrieved information.

[1526] "Means for converting generated response data into audio data" refers to technologies and devices that convert text-based response data into audio data.

[1527] "Means of providing the converted audio data to the user" refers to speakers or other audio output devices for playing the audio data to the user.

[1528] "Means of recognizing user emotions" refers to technologies and algorithms that identify emotions from user voice and text data.

[1529] "Means of adjusting response data based on recognized emotions" refers to technologies and methods that adjust the tone and content of responses according to the user's emotions.

[1530] "Means of translating into multiple languages" refers to technologies and systems that translate text data into different languages.

[1531] "Methods for editing to suit the characteristics of a specific character" refers to techniques and methods for modifying response data to match the personality and manner of speaking of a particular character.

[1532] This invention is a system that performs a series of processes: acquiring user voice, converting it into text data, analyzing it to identify requests, retrieving information, generating response data, converting it back into voice data, and providing it to the user. Furthermore, this invention provides a function to recognize the user's emotions and adjust the response data accordingly. It also includes a function to translate the retrieved information into multiple languages ​​and edit it to suit a specific character.

[1533] Hardware and software to be used

[1534] This system primarily uses the following hardware and software.

[1535] hardware

[1536] Device: A device equipped with a microphone and speaker for audio capture and playback (e.g., smartphone, tablet, smart speaker)

[1537] Server: A remote computer used for data processing and management.

[1538] software

[1539] Speech recognition module: An API for converting speech to text (e.g., Google Cloud Speech-to-Text API)

[1540] Emotion recognition module: A service for analyzing user emotions (e.g., Amazon Comprehend)

[1541] Natural Language Processing Module: A model for analyzing user requests (e.g., OpenAI GPT-3)

[1542] Map API: A service for searching for nearby information based on the user's location (e.g., Google Maps API).

[1543] Speech synthesis module: A technology for converting text into speech (e.g., Amazon Polly)

[1544] Translation Module: An API for translating response text into multiple languages ​​(e.g., Microsoft Translator).

[1545] Specific example

[1546] The following are examples of specific methods for implementing this invention.

[1547] Situation: "Looking for a nearby cafe"

[1548] 1. The user says, "Find a nearby cafe."

[1549] 2. The device uses the microphone to capture audio and saves it as audio data.

[1550] 3. The device uses a speech recognition module (Google Cloud Speech-to-Text API) to convert the captured audio into text: "Find a nearby cafe."

[1551] 4. The device uses an emotion recognition module (Amazon Comprehend) to analyze the voice and text data and determine that the user is "excited."

[1552] 5. The device generates a request containing text data, location information (GPS data), and sentiment data, and sends it to the server.

[1553] 6. The server receives the request, parses it using the NLP module (OpenAI GPT-3), and identifies the request as "Find a cafe."

[1554] 7. The server uses a map API (Google Maps API) to search for cafe information near the user's current location.

[1555] 8. Based on the search results, the server generates response data such as, "There are 3 cafes nearby. We recommend 'Cafe Latte Street'."

[1556] 9. Based on the recognized emotion data "excited," the server generates a response text in a cheerful tone.

[1557] 10. If necessary, the server will translate the response text into multiple languages ​​using a translation module (Microsoft Translator) and edit it to suit the characteristics of the specific character.

[1558] 11. The server sends the generated and edited response data to the terminal.

[1559] 12. The device uses a speech synthesis module (Amazon Polly) to convert the received response text into speech data.

[1560] 13. The device responds through the speaker, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1561] Example of a prompt

[1562] "Generate responses based on the user's emotions when they are looking for a nearby cafe."

[1563] "Generate a response for when the user is angry."

[1564] As described above, this invention is a system that efficiently executes a series of processes, starting with user voice input, going through text conversion and emotion recognition, searching for appropriate information, and generating and providing responses. As a result, users can obtain flexible responses that respond to their emotions, multilingual support, and personalized experiences tailored to their characters.

[1565] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1566] Program processing flow

[1567] Step 1: Voice input (on the device)

[1568] input:

[1569] The user says, "Find a nearby cafe."

[1570] Specific operation:

[1571] The device's microphone captures the audio and saves it as audio data.

[1572] output:

[1573] The acquired audio data is saved.

[1574] Step 2: Voice Recognition (Device)

[1575] input:

[1576] Captured audio data.

[1577] Specific operation:

[1578] The device calls the Google Cloud Speech-to-Text API and uploads the audio data. The API analyzes the audio and retrieves the text "Find a nearby cafe."

[1579] output:

[1580] Text data converted from audio data: "Find a nearby cafe."

[1581] Step 3: Emotion Recognition (Device)

[1582] input:

[1583] Audio and text data: "Find a nearby cafe."

[1584] Specific operation:

[1585] The device uses Amazon Comprehend to analyze voice and text data. From the analysis results, it extracts emotions and determines that the user is "excited."

[1586] output:

[1587] The recognized emotion data is "excited".

[1588] Step 4: Sending the request (on the device)

[1589] input:

[1590] Text data "Looking for a nearby cafe," sentiment data "Excited," and current location obtained from GPS.

[1591] Specific operation:

[1592] The terminal generates request data containing this information and sends it to the server.

[1593] output:

[1594] Request data sent to the server.

[1595] Step 5: Request parsing (server)

[1596] input:

[1597] Request data received by the server.

[1598] Specific operation:

[1599] The server logs the request data to a log file. A natural language processing (NLP) module (OpenAI GPT-3) is used to analyze the text data "Find a nearby cafe". The user's request "Looking for a cafe" is identified from the analysis results.

[1600] output:

[1601] User request: "I'm looking for a cafe."

[1602] Step 6: Information Retrieval (Server)

[1603] input:

[1604] The user's request is "I'm looking for a cafe" and their location information is provided.

[1605] Specific operation:

[1606] The server accesses the Google Maps API to search for nearby cafes based on location information. It then parses the response from the API to retrieve relevant cafe information.

[1607] output:

[1608] The retrieved cafe information (e.g., 3 nearby cafes including "Cafe Latte Street").

[1609] Step 7: Response generation (server)

[1610] input:

[1611] The acquired cafe information and emotion data indicated "excited."

[1612] Specific operation:

[1613] The server generates a text response based on cafe information: "There are three cafes nearby. We recommend 'Cafe Latte Street'." Based on recognized sentiment data, the tone of the response text is adjusted to be brighter.

[1614] output:

[1615] Adjusted response text: "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1616] Step 8: Multilingual support and character editing (server)

[1617] input:

[1618] The generated response text reads, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1619] Specific operation:

[1620] If necessary, the server translates the response text into multiple languages ​​using a translation module. It then edits it to suit the characteristics of the specific character.

[1621] output:

[1622] Translated into multiple languages, response text tailored to the character's characteristics.

[1623] Step 9: Send response (server)

[1624] input:

[1625] Edited response text.

[1626] Specific operation:

[1627] The server sends the edited response text to the terminal.

[1628] output:

[1629] The response text sent to the terminal.

[1630] Step 10: Speech synthesis (device)

[1631] input:

[1632] The received response text.

[1633] Specific operation:

[1634] The device uses a speech synthesis module (Amazon Polly) to convert the response text into speech data.

[1635] output:

[1636] Audio data.

[1637] Step 11: Response playback (device)

[1638] input:

[1639] Generated audio data.

[1640] Specific operation:

[1641] The device plays audio data through its built-in speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[1642] output:

[1643] Voice response to the user.

[1644] (Application Example 2)

[1645] 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."

[1646] Conventional voice assistant systems are limited to simple information retrieval and responses to user requests, making it difficult to provide personalized responses that take into account the user's emotions and circumstances. Furthermore, they lack multilingual support and the ability to edit responses to match the character's personality, posing a challenge in providing an optimal user experience, particularly in virtual stores. Additionally, the system's ability to provide information considering the user's location is underdeveloped, making it difficult to deliver appropriate information to users in real time.

[1647] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1648] In this invention, the server includes means for recognizing the user's emotions, means for acquiring the user's location information, and means for searching relevant data based on the acquired user location information and generating response text. This makes it possible to provide personalized responses in real time that are tailored to the user's emotions and current situation. In addition, by incorporating multilingual support and response editing functions that match character personalities, a more natural and engaging user experience can be achieved.

[1649] A "user" refers to a person who uses the system.

[1650] "Means for capturing speech" refers to a device or technology for acquiring a user's speech as audio data.

[1651] "Means of converting to string data" refers to technology or equipment for converting audio data into text data.

[1652] "Means of analysis to identify user requests" refers to technology or devices that process converted text data to identify what the user is asking for.

[1653] "Means of recognizing user emotions" refers to technology or devices that determine a user's emotions (e.g., joy, anger, sadness, etc.) from their voice data or speech content.

[1654] "Means of retrieving information" refers to technologies or devices for finding appropriate information based on the user's needs and feelings.

[1655] "Means for generating response text" refers to technology or equipment for creating text that serves as a response to a user based on retrieved information.

[1656] "Means of converting to audio data" refers to technology or equipment for converting generated response text into audio output.

[1657] "Means for playing audio data" refers to the technology or device used to allow the user to listen to the converted audio data.

[1658] "Means of translation into multiple languages" refers to technologies or devices that translate generated response text into different languages.

[1659] "Means of editing to match a character's personality" refers to techniques or devices that adjust generated response text to suit a specific character.

[1660] "Means for acquiring location information" refers to technology or devices used to determine the user's current location.

[1661] "Means for searching relevant data and generating response text" refers to a technology or device that searches for appropriate information based on acquired location information and user requests and creates response text.

[1662] This invention relates to a system that captures user voice, converts it into text data, analyzes it to identify requests, retrieves information, generates response text, converts it back into audio data, and plays it back to the user. Furthermore, this system has the capability to recognize user emotions and provides personalized responses according to the user's emotions. The detailed configuration and operation of the system are described below.

[1663] Hardware and software configuration

[1664] This system primarily consists of user terminals and servers. User terminals include devices such as smartphones, smart glasses, and head-mounted displays (HMDs), and these devices include the following software modules.

[1665] 1. Voice Capture Module: Captures the user's voice using the device's microphone.

[1666] 2. Speech Recognition Module: Converts speech data into text data. Specifically, it uses the Google Cloud Speech-to-Text API.

[1667] 3. Emotion Recognition Module: Recognizes emotions from the user's voice and text data. Specifically, it uses the Microsoft Azure Emotion API.

[1668] 4. Location Information Acquisition Module: Acquires the current location information using the GPS function of the user's device.

[1669] The server side includes the following software modules:

[1670] 1. Natural Language Processing (NLP) Module: Uses the Google Cloud Natural Language API to analyze and identify user requests.

[1671] 2. Information Retrieval Module: Searches for appropriate information based on user requests and location information.

[1672] 3. Response Generation Module: Generates response text based on the searched information and the user's sentiment.

[1673] 4. Multilingual Translation Module: Translates response text into multiple languages ​​as needed.

[1674] 5. Character Editing Module: Edit response text to match the personality of a specific character.

[1675] System Operation Description

[1676] When a user says, "Tell me your recommended shampoo," the smart device's microphone captures the voice and saves it as audio data. Then, a speech recognition module (Google Cloud Speech-to-Text API) is used to convert the captured audio data into text, "Tell me your recommended shampoo." At the same time, an emotion recognition module (Microsoft Azure Emotion API) recognizes the user's emotion and determines that they are "interested."

[1677] Request data is generated, containing the converted text data, user location information, and recognized sentiment data, and sent to the server. The server uses a natural language processing (NLP) module to parse the request and identify the request for "shampoo recommendations."

[1678] The server uses an information retrieval module to search the database for relevant data and generates the response text "We recommend 'Shampoo Excellence'." Based on the recognized sentiment data, the response text is edited and presented in a user-friendly tone. Additionally, a multilingual translation module is used to translate the text as needed, and a character editing module is used to edit it to match the character's personality.

[1679] The device receives the response text and converts it into speech data using a text-to-speech (TTS) module (Amazon Polly). It then plays the audio through the smart device's speaker, telling the user, "We recommend 'Shampoo Excellence'."

[1680] Specific examples and prompt statements

[1681] A concrete example of this scenario would be a series of actions taken by a smart device in response to a user's request, such as "Tell me your recommended shampoo." The following is an example of a prompt to a generative AI model.

[1682] Example of a prompt:

[1683] You are a customer service assistant in a virtual store. Generate a response when a user asks for a "recommended shampoo." Respond in a friendly and positive tone, especially if the user shows interest. Remember to mention specific product names.

[1684] In this way, the present invention can provide personalized information in response to user requests and improve the user experience.

[1685] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1686] Step 1:

[1687] Voice input

[1688] User: "Can you recommend a shampoo?"

[1689] Device: Captures the user's voice using the microphone and saves it as audio data.

[1690] Input: User's speech (audio data)

[1691] Output: Audio data

[1692] Step 2:

[1693] Speech recognition

[1694] Terminal: Analyzes captured audio data using a speech recognition module (Google Cloud Speech-to-Text API) and converts it to text.

[1695] Input: Audio data

[1696] Output: Text data "Please recommend a shampoo"

[1697] Step 3:

[1698] emotion recognition

[1699] Terminal: Uses an emotion recognition module (Microsoft Azure Emotion API) to recognize the user's emotions from voice and text data and determine that they are "interested."

[1700] Input: Audio data, text data

[1701] Output: Sentiment data "Interested"

[1702] Step 4:

[1703] Send Request

[1704] Terminal: Generates request data including converted text data, user location information, and recognized sentiment data, and sends it to the server.

[1705] Input: Text data "Please recommend a shampoo," location information, sentiment data "Interested"

[1706] Output: Request data

[1707] Step 5:

[1708] Request Analysis

[1709] Server: Receives request data, uses a natural language processing (NLP) module to parse the text portion and identify the user's request. Identifies the request as "shampoo recommendation".

[1710] Input: Request data

[1711] Output: User request "Shampoo recommendation"

[1712] Step 6:

[1713] Information Retrieval

[1714] Server: Retrieves information on relevant products from the database based on the identified request and location information.

[1715] Input: User request "Shampoo recommendation", location information

[1716] Output: Search result "Recommended shampoo: 'Shampoo Excellence'"

[1717] Step 7:

[1718] Response generation

[1719] Server: Based on the obtained search results and sentiment data, it generates response text for the user. It generates "We recommend 'Shampoo Excellence'."

[1720] Input: Search results, sentiment data "Interested"

[1721] Output: Response text: "We recommend 'Shampoo Excellence'."

[1722] Step 8:

[1723] Multilingual support and character editing

[1724] Server: Translates response text into multiple languages ​​as needed and edits it to suit the character's personality.

[1725] Input: Response text: "My recommendation is 'Shampoo Excellence'."

[1726] Output: Edited response text

[1727] Step 9:

[1728] Send response

[1729] Server: Sends the edited response text to the terminal.

[1730] Input: Edited response text

[1731] Output: Response text received by the terminal

[1732] Step 10:

[1733] Speech synthesis

[1734] Terminal: Converts the received response text into speech data using a text-to-speech (TTS) module (Amazon Polly).

[1735] Input: Response text: "My recommendation is 'Shampoo Excellence'."

[1736] Output: Audio data

[1737] Step 11:

[1738] Response playback

[1739] Device: Plays audio data through the speaker and tells the user, "We recommend 'Shampoo Excellence'."

[1740] Input: Audio data

[1741] Output: Voice response to the user: "We recommend 'Shampoo Excellence'."

[1742] 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.

[1743] 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.

[1744] 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.

[1745] [Fourth Embodiment]

[1746] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1747] 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.

[1748] 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).

[1749] 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.

[1750] 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.

[1751] 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).

[1752] 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.

[1753] 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.

[1754] 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.

[1755] 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.

[1756] 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.

[1757] 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.

[1758] 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".

[1759] This invention relates to a system that accurately captures a user's voice and uses that voice for navigation and information retrieval. The following describes a detailed configuration for implementing this system.

[1760] System Overview

[1761] This system performs a series of processes: capturing the user's voice, converting it into text data, analyzing it to identify the request, searching for information to generate response text, converting it back into audio data, and playing it back to the user. Furthermore, it includes functions to translate information into multiple languages ​​and to edit response text to match the personality of a specific character.

[1762] Program processing flow

[1763] 1. Voice input (device)

[1764] The user says, "Find a nearby cafe."

[1765] The device uses the microphone to capture the user's voice and saves it as audio data.

[1766] 2. Voice recognition (device)

[1767] The device uses a speech recognition module to analyze the voice data and convert it into text, "Find a nearby cafe."

[1768] 3. Sending a request (from the terminal)

[1769] The terminal generates a request containing the converted text data and the user's current location information, and sends it to the server.

[1770] 4. Request parsing (server)

[1771] The server receives the request data, uses a natural language processing (NLP) module to analyze the text, and identifies the request as "Find a cafe."

[1772] 5. Information Retrieval (Server)

[1773] The server searches for relevant cafe information from the internet and databases based on the specified request. For example, it might use a map API to identify nearby cafes.

[1774] 6. Response generation (server)

[1775] The server generates response text for the user based on the search results. For example, it might generate, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1776] 7. Multilingual support and character editing (server)

[1777] The server translates response text into multiple languages ​​as needed. It also edits response text to suit the personality of specific characters.

[1778] 8. Sending a response (server)

[1779] The server sends the generated response text to the terminal.

[1780] 9. Speech synthesis (device)

[1781] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[1782] 10. Response playback (terminal)

[1783] The device plays audio data through its speaker and responds to the user, saying, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[1784] Specific example

[1785] Situation: "Look for a nearby cafe."

[1786] 1. The user says, "Find a nearby cafe."

[1787] 2. The device captures the audio and saves it as audio data.

[1788] 3. The device's voice recognition module converts the voice data into "Find a nearby cafe."

[1789] 4. The device generates a request containing the converted text and current location information and sends it to the server.

[1790] 5. The server receives the request, parses it with the NLP module, and identifies the request as "Find a cafe".

[1791] 6. The server searches for information about cafes related to it using a map API.

[1792] 7. The server generates the response text "There are 3 cafes nearby. We recommend 'Cafe Latte Street'." based on the search results.

[1793] 8. If necessary, the server will translate the response text into multiple languages ​​and edit it to suit the character's personality.

[1794] 9. The server sends a response text to the terminal.

[1795] 10. The terminal converts the response text into speech using a speech synthesis module.

[1796] 11. The device plays a voice message through its speaker, responding to the user with, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1797] In this way, the system allows users to enjoy conversations with their favorite characters while obtaining the information they need.

[1798] The following describes the processing flow.

[1799] Step 1:

[1800] The user says, "Find a nearby cafe."

[1801] The device uses the microphone to capture this audio and saves it as audio data.

[1802] Step 2:

[1803] The device analyzes the voice data captured using its speech recognition module and converts it into text. For example, it might convert it to text like, "Find a nearby cafe."

[1804] Step 3:

[1805] The terminal generates request data that includes generated text data and user location information. This request data contains the user's request and location information.

[1806] Step 4:

[1807] The device sends the request data to the server via the internet.

[1808] Step 5:

[1809] The server receives the request data. The request data includes text and location information from the user.

[1810] Step 6:

[1811] The server uses a natural language processing (NLP) module to analyze the text portion of the request and identify the user's request. For example, it might extract the phrase "find a cafe."

[1812] Step 7:

[1813] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[1814] Step 8:

[1815] The server organizes the search results and generates response text for the user. For example, it might generate a sentence like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1816] Step 9:

[1817] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[1818] Step 10:

[1819] The server sends the generated and edited response text to the terminal.

[1820] Step 11:

[1821] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[1822] Step 12:

[1823] The device plays audio data through its speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[1824] In this way, the system allows users to enjoy conversations with their favorite characters while smoothly obtaining the information they need.

[1825] (Example 1)

[1826] 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".

[1827] Conventional speech recognition systems have had the problem of difficulty in accurately capturing user voices and providing information that meets the user's requests. Furthermore, they are limited to single-language support and cannot provide responses tailored to the personality of a character, resulting in a limited user experience. This invention aims to solve these problems and provide a more interactive and multi-functional information provision system.

[1828] 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.

[1829] In this invention, the server includes means for analyzing a request to identify the user's request, means for retrieving information based on the identified user's request, and means for generating response text based on the retrieved information. This makes it possible to analyze the user's request with high accuracy and provide appropriate information quickly.

[1830] "User" refers to a person who uses a system or an end-user.

[1831] "Audio capture" is the process of recording audio signals using a microphone.

[1832] "Converting to string data" refers to converting audio data into text format using speech recognition technology.

[1833] A "request" refers to a data packet containing a user's request or command, which is sent to the server.

[1834] "Location information" refers to geographical coordinate information obtained using GPS or other positioning technologies.

[1835] A "server" refers to a computer system that processes data and provides services to clients via a network.

[1836] "Natural language processing" refers to the technology that enables computers to understand, generate, and analyze human language.

[1837] "Information retrieval" is the process of obtaining relevant information from databases or the internet based on user requests.

[1838] "Response text" refers to a text-based representation of a user's request or answer to a user's request.

[1839] "Translating into multiple languages" is the process of converting text written in one language into another language.

[1840] "Editing to match the character's personality" refers to modifying text to reflect the characteristics and speaking style of a specific character.

[1841] "Converting to audio data" refers to the process of changing text information into computer-generated speech.

[1842] "Audio playback" refers to the physical output of audio data using sound devices such as speakers.

[1843] This invention is a system that captures user voice and uses it for navigation and information retrieval. The following describes a detailed configuration for implementing this system.

[1844] General overview

[1845] This system includes a voice input device, a speech recognition module, a data communication module, an information retrieval module, a natural language processing module, a response generation module, and a speech synthesis module. Through a series of processes, it is possible to analyze and respond to user voice requests. In addition, it also provides multilingual translation and character-based response generation capabilities.

[1846] Hardware and software to be used

[1847] Device: Mobile information terminals such as smartphones and tablets

[1848] Microphone: Built-in microphone or external microphone device

[1849] Servers: Cloud-based servers, such as AWS or Google Cloud

[1850] Speech recognition module: Google Speech-to-Text API

[1851] Natural language processing modules: SpaCy and Google NLP API

[1852] Information retrieval module: Google Maps API or any database

[1853] Multilingual translation module: Google Translate API

[1854] Speech synthesis module: Amazon Polly

[1855] Processing flow

[1856] 1. The user speaks through a voice input device, for example, saying, "Find a nearby cafe."

[1857] 2. The device's built-in microphone captures this audio and saves it as audio data. This data is temporarily stored in the device's storage.

[1858] 3. The device uses a speech recognition module to convert this speech data into text data. The converted text will read, "Find a nearby cafe."

[1859] 4. The terminal generates the converted text and location information as request data and sends it to the server. This includes precise location information using GPS and other positioning technologies.

[1860] 5. The server receives the request data and parses the text using a natural language processing module. This analysis identifies the user's request (e.g., "find a cafe").

[1861] 6. The server uses the Google Maps API to search for nearby cafes based on the user's request. The retrieved information includes details such as the cafe's name, address, and rating.

[1862] 7. The server generates response text based on the search results. For example, it might generate something like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1863] 8. If necessary, the server will use a multilingual translation module to translate response text. It may also edit the text to suit the character's personality.

[1864] 9. The server sends the edited and translated text to the terminal.

[1865] 10. The terminal converts the received text into speech data using a speech synthesis module.

[1866] 11. Finally, the device plays this audio data to the user, providing the desired information verbally.

[1867] Specific example

[1868] Let's take the situation "finding a nearby cafe" as an example.

[1869] The user says, "Find a nearby cafe."

[1870] The device captures the audio and converts it to text using a speech recognition module.

[1871] The device sends the converted text and location information to the server.

[1872] The server uses an NLP module to analyze the request and identify the request to search for a cafe.

[1873] The server uses the Google Maps API to search for cafe information and generates a response text.

[1874] The server performs multilingual translation and character editing, and then sends the text to the terminal.

[1875] The device converts text into speech using its speech synthesis module.

[1876] The device plays the generated audio and provides the user with cafe information.

[1877] This process allows users to quickly and accurately obtain information through voice input. Furthermore, multilingual support and character-based responses provide a more user-friendly experience.

[1878] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1879] Step 1:

[1880] The user says, "Find a nearby cafe."

[1881] Input: User's spoken voice

[1882] Specific action: The user speaks the desired information (in this case, information about a cafe) into the voice input device.

[1883] Step 2:

[1884] The device uses its built-in microphone to capture the user's voice and saves it as audio data.

[1885] Input: User's spoken voice

[1886] Output: Audio data file (WAV or MP3 format)

[1887] Specific operation: The device's microphone acquires audio as a digital signal, which is then converted and saved as an audio data file.

[1888] Step 3:

[1889] The device uses a speech recognition module (e.g., Google Speech-to-Text API) to convert the audio data into text data.

[1890] Input: Audio data file

[1891] Output: Text data "Find a nearby cafe"

[1892] Specific operation: The speech recognition module receives an audio data file as input, performs speech analysis, and converts it into text. The converted text data is saved to the device's memory.

[1893] Step 4:

[1894] The device generates request data based on the converted text data and location information obtained from its GPS function, and sends it to the server.

[1895] Input: Text data "Find a nearby cafe" and location information

[1896] Output: Request data (API request format)

[1897] Specific operation: Generate a request combining the converted text and location information, and send this request to the server using the HTTPS protocol.

[1898] Step 5:

[1899] The server receives the request data, uses a natural language processing (NLP) module to parse the text data, and identifies the user's request.

[1900] Input: Request data (text and location information)

[1901] Output: Analysis results (User request: Find a cafe)

[1902] Specific operation: The server analyzes the request and uses an NLP module to identify the request as "Find a cafe." This information is stored in the server's memory.

[1903] Step 6:

[1904] The server uses the Google Maps API based on the user's request to search for cafe information based on the user's current location.

[1905] Input: Analysis results (user request) and location information

[1906] Output: Cafe information list (cafe name, address, rating, etc.)

[1907] Specific operation: Call the Google Maps API to retrieve cafe information based on the current location. The retrieved information is stored in the server's memory.

[1908] Step 7:

[1909] The server generates a response text based on the cafe information it has acquired.

[1910] Input: Cafe information list

[1911] Output: Response text "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1912] Specific operation: Analyzes cafe information and generates response text in a user-friendly format.

[1913] Step 8:

[1914] The server translates the response text using a multilingual translation module as needed and edits the response text to suit the personality of the specific character.

[1915] Input: Response text

[1916] Output: Translated and edited response text

[1917] Specific actions: Use the Google Translate API to translate into multiple languages ​​and edit the text to suit the character's personality.

[1918] Step 9:

[1919] The server sends the edited and translated response text to the terminal.

[1920] Input: Response text (after translation and editing)

[1921] Output: Transmitted data

[1922] Specific operation: The edited and translated text data is sent to the terminal using the HTTPS protocol.

[1923] Step 10:

[1924] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[1925] Input: Response text

[1926] Output: Audio data file (MP3 format, etc.)

[1927] Specific operation: Text data is converted into speech data using a speech synthesis module such as Amazon Polly. The converted speech data is stored in the device's storage.

[1928] Step 11:

[1929] The device plays audio data through its speaker and responds to the user.

[1930] Input: Audio data file

[1931] Output: Physical audio output

[1932] Specific operation: The device's speaker is used to play the generated audio data, and the user is responded with, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[1933] (Application Example 1)

[1934] 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".

[1935] In modern customer service, there is a demand for immediate answers to customer questions and provision of appropriate information within the store. However, conventional systems have difficulty providing information via voice input, and furthermore, multilingual support and responses using specific character voices are not adequately provided. As a result, the customer experience is limited, and improving service quality remains a challenge.

[1936] 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.

[1937] In this invention, the server includes means for capturing the user's voice, means for converting the captured voice into string data, means for analyzing the string data to identify the user's request, means for retrieving information based on the identified user's request, means for generating response text based on the retrieved information, means for converting the generated response text into audio data, means for playing the converted audio data back to the user, means for converting the generated response text into the voice of a specific virtual character, means for acquiring the user's current location information, and means for providing information about features within the store based on the acquired location information and the user's request. This enables immediate customer service within the store, and significantly improves the customer experience by enabling multilingual support and responses by specific characters.

[1938] "Means for capturing user voice" refers to a device for acquiring the voice emitted by a user as a digital signal using an audio input device such as a microphone.

[1939] "Methods for converting captured audio into text data" refers to the process of converting acquired audio data into corresponding text data using speech recognition technology.

[1940] "Means for analyzing the string data to identify the user's request" refers to a mechanism that uses natural language processing technology to understand and identify the user's intent and request from the string data.

[1941] "Means for retrieving information based on the requests of a specified user" refers to means for searching and extracting information from databases or the internet that meets the user's requirements.

[1942] "Means for generating response text based on searched information" refers to the process of creating an appropriate response to the user in text format based on search results.

[1943] "Means for converting generated response text into audio data" refers to a system that converts text into audio data using speech synthesis technology.

[1944] "Means for playing back converted audio data to the user" refers to a device for presenting the generated audio data to the user through a speaker or ear-hook type device.

[1945] "Means for converting generated response text into the voice of a specific virtual character" refers to the process of synthesizing generated text into speech using a pre-configured character and voice.

[1946] "Means for obtaining the user's current location information" refers to technologies that use GPS or other location information services to determine the user's current location.

[1947] "Means for providing information about in-store features based on acquired location information and user requests" refers to a system that provides specific product placement and guidance information within a store in accordance with user requests and location information.

[1948] System Overview

[1949] This invention is a system that retrieves information from a user's voice and returns a voice response. The system consists of various modules for realizing voice capture, speech recognition, natural language processing, information retrieval, response generation, speech synthesis, location information acquisition, and voice responses in the voice of a specific virtual character. The system uses a wearable device such as smart glasses to provide customer service in physical stores.

[1950] Hardware and software to be used

[1951] Hardware:

[1952] Smart glasses (voice input device, speaker)

[1953] GPS or location acquisition module

[1954] software:

[1955] Speech recognition modules (e.g., Google Speech-to-Text, Amazon Transcribe)

[1956] Natural Language Processing (NLP) modules (e.g., Google NLP API, OpenAI GPT-3)

[1957] Speech synthesis modules (e.g., Microsoft Azure TTS, Amazon Polly)

[1958] Data processing and data calculation

[1959] 1. Audio capture:

[1960] When a user speaks into the microphone attached to the smart glasses, audio data is captured. This data is stored digitally on the smart glasses.

[1961] 2. Speech recognition:

[1962] The captured audio data is sent to the speech recognition module in the smart glasses and converted into text data. This converts the audio into text format.

[1963] 3. Submit the request:

[1964] The converted text data is sent to the server along with the user's current location information. Here, the user's location information is obtained using a GPS module.

[1965] 4. Request Analysis:

[1966] The server analyzes the received request data and uses a natural language processing (NLP) module to identify the user's request.

[1967] 5. Information Retrieval:

[1968] Based on the specified request, the server searches for relevant information from databases and the internet.

[1969] 6. Response generation:

[1970] The server generates appropriate response text based on the search results. Furthermore, this response text can be translated into multiple languages ​​and edited to sound like a specific virtual character.

[1971] 7. Send a reply:

[1972] The response text is sent from the server to the smart glasses.

[1973] 8. Speech synthesis:

[1974] The speech synthesis module within the smart glasses converts the received response text into speech data. If the voice is converted to that of a specific character, a voice that reflects the character's characteristics will be generated.

[1975] 9. Audio Playback:

[1976] The generated audio data is played back to the user through the smart glasses' speaker.

[1977] Specific example

[1978] Situation: "Can you recommend a wine?"

[1979] 1. The user speaks to the smart glasses and says, "Please recommend a wine."

[1980] 2. The smart glasses capture the audio data and convert it into text data.

[1981] 3. The converted text data and current location information are sent to the server.

[1982] 4. The server analyzes the received data and identifies the request, "Please recommend a wine."

[1983] 5. The server retrieves recommended wine information from the database.

[1984] 6. Based on the information obtained by the server, it generates the response text: "The recommended wine is 'a specific wine'. It is available for purchase on the left side of the liquor section."

[1985] 7. Convert the generated text into the voice of a specific virtual character.

[1986] 8. The server sends the generated response text to the smart glasses.

[1987] 9. The smart glasses convert the speech into voice using a speech synthesis module and play it back to the user.

[1988] Example of a prompt:

[1989] When a user asks, "Can you recommend a wine?", the response text will be generated based on the following text:

[1990] Prompt: The customer asked, "Can you recommend a wine?" Please create a response text following the format below.

[1991] 1. Product Name

[1992] 2. A brief description of the sales location

[1993] Example response: "Our recommended wine is 'a specific wine.' It's located on the left side of the liquor section."

[1994] Text: "Please recommend a wine."

[1995] Thus, the invention provides a concrete means for implementing voice-based customer service in physical stores.

[1996] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1997] Step 1:

[1998] When a user speaks into the smart glasses, the smart glasses' microphone captures the voice and saves it as digital audio data. For example, the user might say, "Tell me your recommended wine." The input is an audio signal, and the output is digital audio data.

[1999] Step 2:

[2000] The speech recognition module on the device performs speech recognition using the captured audio data and converts it into text data. Specifically, the speech recognition module (e.g., Google Speech-to-Text) generates the text "Tell me your recommended wine" from the audio data. The input is audio data, and the output is text data.

[2001] Step 3:

[2002] The terminal obtains the converted text data and the user's current location information (such as GPS data), and sends a request containing this data to the server. Specifically, a location information acquisition module detects the current location and incorporates it into the transmission packet along with the text data. The input is text data and location information, and the output is request data.

[2003] Step 4:

[2004] The server analyzes the received request data and uses a natural language processing module (e.g., Google NLP API) to identify the user's request. Specifically, it analyzes the request "Tell me a wine recommendation" and identifies the request as "Wine Recommendations". The input is the request data, and the output is the analyzed request.

[2005] Step 5:

[2006] Based on the parsed request, the server searches for relevant information from the specified database and the internet. Specifically, it executes a database query to retrieve information about "recommended wines." The input is the parsed request, and the output is the search result.

[2007] Step 6:

[2008] The server generates response text based on the search results. Furthermore, it translates the generated text into multiple languages ​​as needed and edits it to the voice of a specific virtual character. For example, it might generate the text, "Our recommended wine is 'Specific Wine'. It's available on the left side of the liquor section," and edit it to the character's voice. The input is the search results, and the output is the response text.

[2009] Step 7:

[2010] The server sends the generated response text to the terminal. Specifically, it converts the response text into a packet and sends it to the terminal. The input is the response text, and the output is the transmitted packet.

[2011] Step 8:

[2012] The speech synthesis module on the terminal (e.g., Microsoft Azure TTS) converts the received response text into speech data. Specifically, it generates speech data from text data. The input is the response text, and the output is speech data.

[2013] Step 9:

[2014] The device plays audio data and provides a response to the user through the smart glasses' speaker. Specifically, it passes the generated audio data to the playback device and outputs the audio. The input is audio data, and the output is audio information provided to the user.

[2015] 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.

[2016] System Overview

[2017] This invention is a system that performs a series of processes, including capturing user voice, converting it into text data, analyzing it to identify requests, retrieving information, generating response text, and converting it back into audio data for playback to the user. Furthermore, this invention incorporates an emotion engine that recognizes user emotions, providing a function to improve the user experience. It also includes the ability to translate retrieved information into multiple languages ​​and edit it to suit the personality of a specific character.

[2018] Program processing flow

[2019] 1. Voice input (device)

[2020] The user says, "Find a nearby cafe."

[2021] The device uses the microphone to capture the user's voice and saves it as audio data.

[2022] 2. Voice recognition (device)

[2023] The device analyzes the voice data captured using its speech recognition module and converts it into text. For example, it might convert it to text like, "Find a nearby cafe."

[2024] 3. Emotion Recognition (Device)

[2025] The device uses an emotion engine to recognize the user's emotions (joy, anger, sadness, etc.) from voice and text data.

[2026] 4. Sending a request (from the terminal)

[2027] The device generates request data containing converted text data, user location information, and recognized sentiment data, and sends it to the server.

[2028] 5. Request parsing (server)

[2029] The server receives the request data. The request data includes the user's text, location information, and sentiment data.

[2030] 6. Information Retrieval (Server)

[2031] The server uses a natural language processing (NLP) module to parse the text portion of the request and identify the user's request. For example, it identifies the request "Find a cafe."

[2032] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[2033] 7. Response generation (server)

[2034] The server generates response text for the user based on the search results. For example, it might generate a sentence like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[2035] The server adjusts the generated response text based on recognized emotion data. For example, if the user is angry, the tone of the response will be softened.

[2036] 8. Multilingual support and character editing (server)

[2037] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[2038] 9. Sending a response (server)

[2039] The server sends the generated and edited response text to the terminal.

[2040] 10. Speech synthesis (device)

[2041] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[2042] 11. Response playback (terminal)

[2043] The device plays audio data through its speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[2044] Specific example

[2045] Situation: "Look for a nearby cafe."

[2046] 1. The user says, "Find a nearby cafe."

[2047] 2. The device captures the audio and saves it as audio data.

[2048] 3. The device's voice recognition module converts the voice data into "Find a nearby cafe."

[2049] 4. The device uses an emotion engine to recognize the user's emotions and determines that they are "excited."

[2050] 5. The device generates a request containing the converted text, location information, and sentiment data, and sends it to the server.

[2051] 6. The server receives the request, parses it with the NLP module, and identifies the request as "Find a cafe".

[2052] 7. The server uses a map API to search for information about nearby cafes.

[2053] 8. The server generates the response text "There are 3 cafes nearby. We recommend 'Cafe Latte Street'." based on the search results.

[2054] 9. Based on the recognized emotion data (excited), the server adjusts the response text and generates it in a brighter tone.

[2055] 10. If necessary, the server will translate the response text into multiple languages ​​and edit it to suit the character's personality.

[2056] 11. The server sends a response text to the terminal.

[2057] 12. The terminal converts the response text into speech using a speech synthesis module.

[2058] 13. The device plays a voice message through its speaker, responding to the user in a cheerful tone, saying, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[2059] In this way, the system allows users to enjoy conversations with their favorite characters while smoothly obtaining the necessary information. Furthermore, responses adapted to the user's emotions provide a more personalized experience.

[2060] The following describes the processing flow.

[2061] Step 1:

[2062] The user says, "Find a nearby cafe."

[2063] The device uses the microphone to capture audio and saves it as audio data.

[2064] Step 2:

[2065] The device uses a speech recognition module to analyze the captured audio data and convert it into text data such as "Find a nearby cafe."

[2066] Step 3:

[2067] The device uses an emotion engine to recognize the user's emotions from voice data and converted text data. For example, it might determine that the user is "excited" based on the tone and speed of their voice.

[2068] Step 4:

[2069] The device generates request data that includes converted text data, user location information, and recognized sentiment data. This request data contains the user's request, location information, and sentiment information.

[2070] Step 5:

[2071] The device sends the request data to the server via the internet.

[2072] Step 6:

[2073] The server receives the request data, which includes the user's text, location information, and sentiment data.

[2074] Step 7:

[2075] The server uses a natural language processing (NLP) module to analyze the text portion of the request and identify the user's request. For example, it might extract the request "Find a cafe."

[2076] Step 8:

[2077] The server searches for relevant information based on the identified request and location information. Specifically, it uses a map API to retrieve information about cafes near the user's current location.

[2078] Step 9:

[2079] The server organizes the search results and generates response text for the user. For example, it might generate a response like, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[2080] Step 10:

[2081] The server adjusts the response text based on recognized sentiment data. For example, if the user is excited, the response tone becomes brighter and positive words are added.

[2082] Step 11:

[2083] If necessary, the server translates the generated response text into multiple languages ​​and edits it to suit the personality of the specific character.

[2084] Step 12:

[2085] The server sends the generated and formatted response text to the terminal.

[2086] Step 13:

[2087] The terminal converts the received response text into speech data using a text-to-speech (TTS) module.

[2088] Step 14:

[2089] The device plays audio data through its speaker and responds to the user in a cheerful tone, saying, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[2090] Through this series of processes, users can enjoy conversations with their favorite characters while quickly and effectively obtaining the information they need. Furthermore, emotion recognition provides more personalized responses, improving the user experience.

[2091] (Example 2)

[2092] 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".

[2093] Conventional voice dialogue systems struggled to generate flexible responses that reflected user emotions or to provide responses tailored to the individuality of specific characters. As a result, the user experience was limited, and it was difficult to address individual requests. Furthermore, multilingual support was insufficient, making it difficult to serve a global user base.

[2094] 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.

[2095] In this invention, the server includes means for acquiring user voice, means for converting the acquired voice into text data, means for analyzing the converted text data to identify the user's request, means for retrieving information based on the identified user's request, means for generating response data based on the retrieved information, means for converting the generated response data into voice data, means for providing the converted voice data to the user, means for recognizing the user's emotions, and means for adjusting the response data based on the recognized emotions. This enables flexible response generation based on the user's emotions, provision of responses tailored to specific characters, and multilingual support.

[2096] "Means of acquiring user voice" refers to devices and technologies that capture user voice and acquire it as data.

[2097] "Means of converting acquired audio into text data" refers to technologies and algorithms that convert audio data into string data.

[2098] "Means of analyzing converted text data to identify user requests" refers to technologies and methods that analyze text data to identify what the user wants.

[2099] "Means of retrieving information based on the requests of identified users" refers to technologies and systems that retrieve relevant information from databases and external services based on user requests.

[2100] "Means for generating response data based on retrieved information" refers to algorithms and technologies that generate appropriate responses for the user based on retrieved information.

[2101] "Means for converting generated response data into audio data" refers to technologies and devices that convert text-based response data into audio data.

[2102] "Means of providing the converted audio data to the user" refers to speakers or other audio output devices for playing the audio data to the user.

[2103] "Means of recognizing user emotions" refers to technologies and algorithms that identify emotions from user voice and text data.

[2104] "Means of adjusting response data based on recognized emotions" refers to technologies and methods that adjust the tone and content of responses according to the user's emotions.

[2105] "Means of translating into multiple languages" refers to technologies and systems that translate text data into different languages.

[2106] "Methods for editing to suit the characteristics of a specific character" refers to techniques and methods for modifying response data to match the personality and manner of speaking of a particular character.

[2107] This invention is a system that performs a series of processes: acquiring user voice, converting it into text data, analyzing it to identify requests, retrieving information, generating response data, converting it back into voice data, and providing it to the user. Furthermore, this invention provides a function to recognize the user's emotions and adjust the response data accordingly. It also includes a function to translate the retrieved information into multiple languages ​​and edit it to suit a specific character.

[2108] Hardware and software to be used

[2109] This system primarily uses the following hardware and software.

[2110] hardware

[2111] Device: A device equipped with a microphone and speaker for audio capture and playback (e.g., smartphone, tablet, smart speaker)

[2112] Server: A remote computer used for data processing and management.

[2113] software

[2114] Speech recognition module: An API for converting speech to text (e.g., Google Cloud Speech-to-Text API)

[2115] Emotion recognition module: A service for analyzing user emotions (e.g., Amazon Comprehend)

[2116] Natural Language Processing Module: A model for analyzing user requests (e.g., OpenAI GPT-3)

[2117] Map API: A service for searching for nearby information based on the user's location (e.g., Google Maps API).

[2118] Speech synthesis module: A technology for converting text into speech (e.g., Amazon Polly)

[2119] Translation Module: An API for translating response text into multiple languages ​​(e.g., Microsoft Translator).

[2120] Specific example

[2121] The following are examples of specific methods for implementing this invention.

[2122] Situation: "Looking for a nearby cafe"

[2123] 1. The user says, "Find a nearby cafe."

[2124] 2. The device uses the microphone to capture audio and saves it as audio data.

[2125] 3. The device uses a speech recognition module (Google Cloud Speech-to-Text API) to convert the captured audio into text: "Find a nearby cafe."

[2126] 4. The device uses an emotion recognition module (Amazon Comprehend) to analyze the voice and text data and determine that the user is "excited."

[2127] 5. The device generates a request containing text data, location information (GPS data), and sentiment data, and sends it to the server.

[2128] 6. The server receives the request, parses it using the NLP module (OpenAI GPT-3), and identifies the request as "Find a cafe."

[2129] 7. The server uses a map API (Google Maps API) to search for cafe information near the user's current location.

[2130] 8. Based on the search results, the server generates response data such as, "There are 3 cafes nearby. We recommend 'Cafe Latte Street'."

[2131] 9. Based on the recognized emotion data "excited," the server generates a response text in a cheerful tone.

[2132] 10. If necessary, the server will translate the response text into multiple languages ​​using a translation module (Microsoft Translator) and edit it to suit the characteristics of the specific character.

[2133] 11. The server sends the generated and edited response data to the terminal.

[2134] 12. The device uses a speech synthesis module (Amazon Polly) to convert the received response text into speech data.

[2135] 13. The device responds through the speaker, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[2136] Example of a prompt

[2137] "Generate responses based on the user's emotions when they are looking for a nearby cafe."

[2138] "Generate a response for when the user is angry."

[2139] As described above, this invention is a system that efficiently executes a series of processes, starting with user voice input, going through text conversion and emotion recognition, searching for appropriate information, and generating and providing responses. As a result, users can obtain flexible responses that respond to their emotions, multilingual support, and personalized experiences tailored to their characters.

[2140] The flow of the specific processing in Example 2 will be explained using Figure 13.

[2141] Program processing flow

[2142] Step 1: Voice input (on the device)

[2143] input:

[2144] The user says, "Find a nearby cafe."

[2145] Specific operation:

[2146] The device's microphone captures the audio and saves it as audio data.

[2147] output:

[2148] The acquired audio data is saved.

[2149] Step 2: Voice Recognition (Device)

[2150] input:

[2151] Captured audio data.

[2152] Specific operation:

[2153] The device calls the Google Cloud Speech-to-Text API and uploads the audio data. The API analyzes the audio and retrieves the text "Find a nearby cafe."

[2154] output:

[2155] Text data converted from audio data: "Find a nearby cafe."

[2156] Step 3: Emotion Recognition (Device)

[2157] input:

[2158] Audio and text data: "Find a nearby cafe."

[2159] Specific operation:

[2160] The device uses Amazon Comprehend to analyze voice and text data. From the analysis results, it extracts emotions and determines that the user is "excited."

[2161] output:

[2162] The recognized emotion data is "excited".

[2163] Step 4: Sending the request (on the device)

[2164] input:

[2165] Text data "Looking for a nearby cafe," sentiment data "Excited," and current location obtained from GPS.

[2166] Specific operation:

[2167] The terminal generates request data containing this information and sends it to the server.

[2168] output:

[2169] Request data sent to the server.

[2170] Step 5: Request parsing (server)

[2171] input:

[2172] Request data received by the server.

[2173] Specific operation:

[2174] The server logs the request data to a log file. A natural language processing (NLP) module (OpenAI GPT-3) is used to analyze the text data "Find a nearby cafe". The user's request "Looking for a cafe" is identified from the analysis results.

[2175] output:

[2176] User request: "I'm looking for a cafe."

[2177] Step 6: Information Retrieval (Server)

[2178] input:

[2179] The user's request is "I'm looking for a cafe" and their location information is provided.

[2180] Specific operation:

[2181] The server accesses the Google Maps API to search for nearby cafes based on location information. It then parses the response from the API to retrieve relevant cafe information.

[2182] output:

[2183] The retrieved cafe information (e.g., 3 nearby cafes including "Cafe Latte Street").

[2184] Step 7: Response generation (server)

[2185] input:

[2186] The acquired cafe information and emotion data indicated "excited."

[2187] Specific operation:

[2188] The server generates a text response based on cafe information: "There are three cafes nearby. We recommend 'Cafe Latte Street'." Based on recognized sentiment data, the tone of the response text is adjusted to be brighter.

[2189] output:

[2190] Adjusted response text: "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[2191] Step 8: Multilingual support and character editing (server)

[2192] input:

[2193] The generated response text reads, "There are three cafes nearby. We recommend 'Cafe Latte Street'."

[2194] Specific operation:

[2195] If necessary, the server translates the response text into multiple languages ​​using a translation module. It then edits it to suit the characteristics of the specific character.

[2196] output:

[2197] Translated into multiple languages, response text tailored to the character's characteristics.

[2198] Step 9: Send response (server)

[2199] input:

[2200] Edited response text.

[2201] Specific operation:

[2202] The server sends the edited response text to the terminal.

[2203] output:

[2204] The response text sent to the terminal.

[2205] Step 10: Speech synthesis (device)

[2206] input:

[2207] The received response text.

[2208] Specific operation:

[2209] The device uses a speech synthesis module (Amazon Polly) to convert the response text into speech data.

[2210] output:

[2211] Audio data.

[2212] Step 11: Response playback (device)

[2213] input:

[2214] Generated audio data.

[2215] Specific operation:

[2216] The device plays audio data through its built-in speaker and responds to the user with, "There are three cafes nearby. I recommend 'Cafe Latte Street'."

[2217] output:

[2218] Voice response to the user.

[2219] (Application Example 2)

[2220] 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".

[2221] Conventional voice assistant systems are limited to simple information retrieval and responses to user requests, making it difficult to provide personalized responses that take into account the user's emotions and circumstances. Furthermore, they lack multilingual support and the ability to edit responses to match the character's personality, posing a challenge in providing an optimal user experience, particularly in virtual stores. Additionally, the system's ability to provide information considering the user's location is underdeveloped, making it difficult to deliver appropriate information to users in real time.

[2222] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[2223] In this invention, the server includes means for recognizing the user's emotions, means for acquiring the user's location information, and means for searching relevant data based on the acquired user location information and generating response text. This makes it possible to provide personalized responses in real time that are tailored to the user's emotions and current situation. In addition, by incorporating multilingual support and response editing functions that match character personalities, a more natural and engaging user experience can be achieved.

[2224] A "user" refers to a person who uses the system.

[2225] "Means for capturing speech" refers to a device or technology for acquiring a user's speech as audio data.

[2226] "Means of converting to string data" refers to technology or equipment for converting audio data into text data.

[2227] "Means of analysis to identify user requests" refers to technology or devices that process converted text data to identify what the user is asking for.

[2228] "Means of recognizing user emotions" refers to technology or devices that determine a user's emotions (e.g., joy, anger, sadness, etc.) from their voice data or speech content.

[2229] "Means of retrieving information" refers to technologies or devices for finding appropriate information based on the user's needs and feelings.

[2230] "Means for generating response text" refers to technology or equipment for creating text that serves as a response to a user based on retrieved information.

[2231] "Means of converting to audio data" refers to technology or equipment for converting generated response text into audio output.

[2232] "Means for playing audio data" refers to the technology or device used to allow the user to listen to the converted audio data.

[2233] "Means of translation into multiple languages" refers to technologies or devices that translate generated response text into different languages.

[2234] "Means of editing to match a character's personality" refers to techniques or devices that adjust generated response text to suit a specific character.

[2235] "Means for acquiring location information" refers to technology or devices used to determine the user's current location.

[2236] "Means for searching relevant data and generating response text" refers to a technology or device that searches for appropriate information based on acquired location information and user requests and creates response text.

[2237] This invention relates to a system that captures user voice, converts it into text data, analyzes it to identify requests, retrieves information, generates response text, converts it back into audio data, and plays it back to the user. Furthermore, this system has the capability to recognize user emotions and provides personalized responses according to the user's emotions. The detailed configuration and operation of the system are described below.

[2238] Hardware and software configuration

[2239] This system primarily consists of user terminals and servers. User terminals include devices such as smartphones, smart glasses, and head-mounted displays (HMDs), and these devices include the following software modules.

[2240] 1. Voice Capture Module: Captures the user's voice using the device's microphone.

[2241] 2. Speech Recognition Module: Converts speech data into text data. Specifically, it uses the Google Cloud Speech-to-Text API.

[2242] 3. Emotion Recognition Module: Recognizes emotions from the user's voice and text data. Specifically, it uses the Microsoft Azure Emotion API.

[2243] 4. Location Information Acquisition Module: Acquires the current location information using the GPS function of the user's device.

[2244] The server side includes the following software modules:

[2245] 1. Natural Language Processing (NLP) Module: Uses the Google Cloud Natural Language API to analyze and identify user requests.

[2246] 2. Information Retrieval Module: Searches for appropriate information based on user requests and location information.

[2247] 3. Response Generation Module: Generates response text based on the searched information and the user's sentiment.

[2248] 4. Multilingual Translation Module: Translates response text into multiple languages ​​as needed.

[2249] 5. Character Editing Module: Edit response text to match the personality of a specific character.

[2250] System Operation Description

[2251] When a user says, "Tell me your recommended shampoo," the smart device's microphone captures the voice and saves it as audio data. Then, a speech recognition module (Google Cloud Speech-to-Text API) is used to convert the captured audio data into text, "Tell me your recommended shampoo." At the same time, an emotion recognition module (Microsoft Azure Emotion API) recognizes the user's emotion and determines that they are "interested."

[2252] Request data is generated, containing the converted text data, user location information, and recognized sentiment data, and sent to the server. The server uses a natural language processing (NLP) module to parse the request and identify the request for "shampoo recommendations."

[2253] The server uses an information retrieval module to search the database for relevant data and generates the response text "We recommend 'Shampoo Excellence'." Based on the recognized sentiment data, the response text is edited and presented in a user-friendly tone. Additionally, a multilingual translation module is used to translate the text as needed, and a character editing module is used to edit it to match the character's personality.

[2254] The device receives the response text and converts it into speech data using a text-to-speech (TTS) module (Amazon Polly). It then plays the audio through the smart device's speaker, telling the user, "We recommend 'Shampoo Excellence'."

[2255] Specific examples and prompt statements

[2256] A concrete example of this scenario would be a series of actions taken by a smart device in response to a user's request, such as "Tell me your recommended shampoo." The following is an example of a prompt to a generative AI model.

[2257] Example of a prompt:

[2258] You are a customer service assistant in a virtual store. Generate a response when a user asks for a "recommended shampoo." Respond in a friendly and positive tone, especially if the user shows interest. Remember to mention specific product names.

[2259] In this way, the present invention can provide personalized information in response to user requests and improve the user experience.

[2260] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[2261] Step 1:

[2262] Voice input

[2263] User: "Can you recommend a shampoo?"

[2264] Device: Captures the user's voice using the microphone and saves it as audio data.

[2265] Input: User's speech (audio data)

[2266] Output: Audio data

[2267] Step 2:

[2268] Speech recognition

[2269] Terminal: Analyzes captured audio data using a speech recognition module (Google Cloud Speech-to-Text API) and converts it to text.

[2270] Input: Audio data

[2271] Output: Text data "Please recommend a shampoo"

[2272] Step 3:

[2273] emotion recognition

[2274] Terminal: Uses an emotion recognition module (Microsoft Azure Emotion API) to recognize the user's emotions from voice and text data and determine that they are "interested."

[2275] Input: Audio data, text data

[2276] Output: Sentiment data "Interested"

[2277] Step 4:

[2278] Send Request

[2279] Terminal: Generates request data including converted text data, user location information, and recognized sentiment data, and sends it to the server.

[2280] Input: Text data "Please recommend a shampoo," location information, sentiment data "Interested"

[2281] Output: Request data

[2282] Step 5:

[2283] Request Analysis

[2284] Server: Receives request data, uses a natural language processing (NLP) module to parse the text portion and identify the user's request. Identifies the request as "shampoo recommendation".

[2285] Input: Request data

[2286] Output: User request "Shampoo recommendation"

[2287] Step 6:

[2288] Information Retrieval

[2289] Server: Retrieves information on relevant products from the database based on the identified request and location information.

[2290] Input: User request "Shampoo recommendation", location information

[2291] Output: Search result "Recommended shampoo: 'Shampoo Excellence'"

[2292] Step 7:

[2293] Response generation

[2294] Server: Based on the obtained search results and sentiment data, it generates response text for the user. It generates "We recommend 'Shampoo Excellence'."

[2295] Input: Search results, sentiment data "Interested"

[2296] Output: Response text: "We recommend 'Shampoo Excellence'."

[2297] Step 8:

[2298] Multilingual support and character editing

[2299] Server: Translates response text into multiple languages ​​as needed and edits it to suit the character's personality.

[2300] Input: Response text: "My recommendation is 'Shampoo Excellence'."

[2301] Output: Edited response text

[2302] Step 9:

[2303] Send response

[2304] Server: Sends the edited response text to the terminal.

[2305] Input: Edited response text

[2306] Output: Response text received by the terminal

[2307] Step 10:

[2308] Speech synthesis

[2309] Terminal: Converts the received response text into speech data using a text-to-speech (TTS) module (Amazon Polly).

[2310] Input: Response text: "My recommendation is 'Shampoo Excellence'."

[2311] Output: Audio data

[2312] Step 11:

[2313] Response playback

[2314] Device: Plays audio data through the speaker and tells the user, "We recommend 'Shampoo Excellence'."

[2315] Input: Audio data

[2316] Output: Voice response to the user: "We recommend 'Shampoo Excellence'."

[2317] 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.

[2318] 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.

[2319] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[2320] 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.

[2321] 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.

[2322] 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.

[2323] 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.

[2324] 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.

[2325] 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."

[2326] 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.

[2327] 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.

[2328] 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.

[2329] 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.

[2330] 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.

[2331] 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.

[2332] 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.

[2333] 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.

[2334] 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.

[2335] 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.

[2336] 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.

[2337] 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 to be incorporated by reference.

[2338] The following is further disclosed regarding the embodiments described above.

[2339] (Claim 1)

[2340] A means of capturing the user's voice,

[2341] A means of converting captured audio into text data,

[2342] A means for analyzing the string data to identify the user's request,

[2343] A means of retrieving information based on the requests of an identified user,

[2344] Means for generating response text based on retrieved information,

[2345] A means for converting the generated response text into audio data,

[2346] A means of playing the converted audio data to the user,

[2347] A system that includes this.

[2348] (Claim 2)

[2349] The system according to claim 1, further comprising means for translating retrieved information into multiple languages.

[2350] (Claim 3)

[2351] The system according to claim 1, further comprising means for editing the generated response text to suit the personality of a particular character.

[2352] "Example 1"

[2353] (Claim 1)

[2354] A means of capturing the user's voice,

[2355] A means of converting captured audio into text data,

[2356] A means for generating a request including the string data and location information and sending it to the server,

[2357] A means of analyzing requests to identify user requests,

[2358] A means of retrieving information based on the requests of an identified user,

[2359] Means for generating response text based on retrieved information,

[2360] A means of translating the generated response text into multiple languages,

[2361] A means of editing the generated response text to suit the personality of a specific character,

[2362] A means for converting the generated response text into audio data,

[2363] A means of playing the converted audio data to the user,

[2364] A system that includes this.

[2365] (Claim 2)

[2366] The system according to claim 1, further comprising means for providing the user with translated and edited text.

[2367] (Claim 3)

[2368] The system according to claim 1, further comprising means for using a natural language processing module in analyzing user requests.

[2369] "Application Example 1"

[2370] (Claim 1)

[2371] A means of capturing the user's voice,

[2372] A means of converting captured audio into text data,

[2373] A means for analyzing the string data to identify the user's request,

[2374] A means of retrieving information based on the requests of an identified user,

[2375] Means for generating response text based on retrieved information,

[2376] A means for converting the generated response text into audio data,

[2377] A means of playing the converted audio data to the user,

[2378] A means for converting the generated response text into the voice of a specific virtual character,

[2379] A means of obtaining the user's current location information,

[2380] A means for providing information about features within a store based on acquired location information and user requests,

[2381] A system that includes this.

[2382] (Claim 2)

[2383] The system according to claim 1, further comprising means for translating retrieved information into multiple languages.

[2384] (Claim 3)

[2385] The system according to claim 1, further comprising means for editing the generated response text to suit the personality of a particular character.

[2386] "Example 2 of combining an emotion engine"

[2387] (Claim 1)

[2388] A means of acquiring the user's voice,

[2389] A means of converting acquired audio into text data,

[2390] A means of analyzing the converted text data to identify the user's request,

[2391] A means of retrieving information based on the requests of an identified user,

[2392] Means for generating response data based on retrieved information,

[2393] A means for converting the generated response data into audio data,

[2394] A means of providing the converted audio data to the user,

[2395] Means of recognizing user emotions,

[2396] Means for adjusting response data based on recognized emotions,

[2397] A system that includes this.

[2398] (Claim 2)

[2399] The system according to claim 1, further comprising means for translating retrieved information into multiple languages.

[2400] (Claim 3)

[2401] The system according to claim 1, further comprising means for editing the generated response data to suit the characteristics of a specific character.

[2402] "Application example 2 of combining emotional engines"

[2403] (Claim 1)

[2404] A means of capturing the user's voice,

[2405] A means of converting captured audio into text data,

[2406] A means for analyzing the string data to identify the user's request,

[2407] Means of recognizing user emotions,

[2408] A means of retrieving information based on the requests and sentiments of identified users,

[2409] Means for generating response text based on retrieved information,

[2410] A means for converting the generated response text into audio data,

[2411] A means of playing the converted audio data to the user,

[2412] A system that includes this.

[2413] (Claim 2)

[2414] The system according to claim 1, further comprising means for translating retrieved information into multiple languages.

[2415] (Claim 3)

[2416] The system according to claim 1, further comprising means for editing the generated response text to suit the personality of a particular character.

[2417] (Claim 4)

[2418] The system according to claim 1, further comprising means for obtaining the user's location information.

[2419] (Claim 5)

[2420] The system according to claim 4, further comprising means for searching for relevant data based on acquired user location information and generating response text. [Explanation of Symbols]

[2421] 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 means of capturing the user's voice, A means of converting captured audio into text data, A means for analyzing the string data to identify the user's request, A means of retrieving information based on the requests of an identified user, Means for generating response text based on retrieved information, A means for converting the generated response text into audio data, A means of playing the converted audio data to the user, A system that includes this.

2. The system according to claim 1, further comprising means for translating retrieved information into multiple languages.

3. The system according to claim 1, further comprising means for editing the generated response text to suit the personality of a specific character.

Images