system

Abstract

We provide the system. [Solution] A means for acquiring voice data from the user, Means for converting the aforementioned audio data into text data, A means for transmitting the aforementioned text data to a natural language generation AI and receiving the generated response text data, A means for converting the aforementioned response text data into audio data, Means for providing the aforementioned audio data to the user, A system that includes this.

Description

Technical Field

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

Background Art

[0002] Patent Document 1 discloses a method for controlling a persona chatbot, which is performed by at least one processor, including steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In recent years, the number of overseas travelers has increased, and language barriers have become a major obstacle in communication. There are many cases where travelers cannot understand the local language and feel difficulties when seeking information or help. In particular, the language barrier in situations where accurate transmission of information is required during emergencies or important times is serious. To solve such problems, a system that can provide accurate real-time translation is needed, but current translation tools often have delays, misrecognitions, and mistranslations in voice-to-voice conversion, failing to meet user expectations.

Means for Solving the Problems

[0005] This invention proposes a system that acquires voice data from a user, converts it into text data using a speech recognition API, sends that text data to a natural language generation AI to receive the generated response text data, and then converts that response text data back into voice data to provide to the user. This enables users to have natural interactions in real time and effectively overcome language barriers. This system provides rapid and accurate translation by combining speech recognition technology that converts voice data into text data, a natural language generation AI with advanced response generation capabilities, and speech synthesis technology that converts text data into high-quality voice data. Furthermore, by efficiently implementing the entire process through an API, practicality and reliability are also improved.

[0006] "Audio data" refers to a digital recording of a user's voice.

[0007] "Text data" refers to string information converted from audio data, and is data in a format that allows for natural language processing.

[0008] "Natural language generation AI" is an artificial intelligence technology that generates natural-sounding sentences that are easy for humans to understand from input text data.

[0009] A "speech recognition API" is an application programming interface that provides the functionality to convert speech data into text data.

[0010] A "speech synthesis API" is an application programming interface that provides the functionality to convert text data into speech data.

[0011] An "API request" is a protocol used to request a specific service to perform a function.

[0012] A "user" is a person who uses the system, and in particular, a person who uses the translation function to communicate.

[0013] A "terminal" is a device used by a user to input voice, send the voice data to a server, and ultimately play back the voice data.

[0014] A "server" is a central computing system that performs tasks such as processing audio data, running natural language generation AI, and converting audio data.

[0015] "Real-time" refers to processing user actions immediately and providing results without delay. [Brief explanation of the drawing]

[0016] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] 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 combined with an emotion engine. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when combined with an emotion engine.

Mode for Carrying Out the Invention

[0017] Hereinafter, an example of an embodiment of the system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0018] First, the language used in the following description will be explained.

[0019] In the following embodiments, the numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0020] In the following embodiments, the numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0021] 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.

[0022] 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).

[0023] 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."

[0024] [First Embodiment]

[0025] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0026] 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.

[0027] 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).

[0028] 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.

[0029] 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.

[0030] 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.

[0031] 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.

[0032] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0033] 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.

[0034] 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.

[0035] 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.

[0036] 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".

[0037] This invention is a system that integrates speech recognition, natural language generation, and speech synthesis technologies to enable travelers and others to communicate smoothly across language barriers. This system is implemented as follows.

[0038] User

[0039] The user inputs their questions and requests into the device using voice. For example, they might voice-input the question, "Where is the nearest pharmacy?"

[0040] terminal

[0041] The user's voice is recorded and saved as audio data.

[0042] The recorded audio data is sent to the server. Here, the audio data is sent, for example, using an HTTP request.

[0043] server

[0044] Receive audio data.

[0045] Use a speech recognition API to convert received audio data into text data. For example, use Google's (registered trademark) speech recognition API.

[0046] The converted text data is sent to a natural language generation AI to generate appropriate response text data. Here, for example, OpenAI's GPT-3® is used as the generation AI.

[0047] The generated response text data is sent to a speech synthesis API and converted into speech data. Examples of speech synthesis APIs used include Google Text-to-Speech and gTTS.

[0048] The final audio data is sent to the device as the response. This audio data is sent in the form of an HTTP response or similar.

[0049] terminal

[0050] Receive audio data sent from the server.

[0051] The system plays back the received audio data and provides the user with an answer.

[0052] Specific example

[0053] Let's take a scenario where a user asks, "Where is the nearest pharmacy?" as an example. When the user inputs the question by voice, the device records the voice and sends it to the server. The server uses speech recognition to convert it into text, "Where is the nearest pharmacy?", and sends that text to a natural language generation AI to obtain an answer. For example, if the answer is, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk," this is converted into audio data and sent to the device. The device then plays the received audio data and provides it to the user.

[0054] This allows users to receive real-time audio answers to their questions, enabling smoother communication across language barriers.

[0055] The following describes the processing flow.

[0056] Step 1:

[0057] The user makes a voice input to the device. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as audio data.

[0058] Step 2:

[0059] The device sends the recorded audio data to the server. Specifically, it uploads the audio file to the server using an HTTP POST request.

[0060] Step 3:

[0061] The server receives the audio data. After receiving it, it temporarily stores the audio data and converts it into text data using a speech recognition API. For example, it uses Google's speech recognition API to convert the audio into the text "Where is the nearest pharmacy?".

[0062] Step 4:

[0063] The server sends the converted text data to a natural language generation AI. For example, it calls OpenAI's GPT-3 API to generate an appropriate answer. In this case, it sends the text data "Where is the nearest pharmacy?" and receives the generated answer text data.

[0064] Step 5:

[0065] The server sends the received response text data to a speech synthesis API and converts it into speech data. Specifically, it uses the Google Text-to-Speech API or gTTS to generate speech data such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0066] Step 6:

[0067] The server sends the generated audio data to the terminal as an HTTP response. Here, it is sent as an audio file and converted to a format that the terminal can receive.

[0068] Step 7:

[0069] The device saves the received audio data and plays it back to the user. This allows the user to hear the response, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0070] These steps enable users to interact naturally in real time and effectively overcome language barriers.

[0071] (Example 1)

[0072] 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."

[0073] Conventional speech recognition technologies are highly dependent on specific languages ​​and environments, making it difficult to achieve smooth communication for travelers or in situations requiring cross-language communication. In particular, existing systems have limitations in response speed and accuracy when real-time, accurate translation and responses are required. Furthermore, the process of converting speech data to text data and then converting the response back to speech data is complex and time-consuming, posing a significant challenge.

[0074] 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.

[0075] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to a natural language generation algorithm and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. This enables users to communicate smoothly and accurately in different languages ​​in real time, and facilitates smooth conversations even when used by travelers or in multilingual environments.

[0076] "User" refers to an individual or group that uses this system.

[0077] "Audio data" refers to digital data that records the content of a user's speech.

[0078] "Text data" refers to data obtained by analyzing audio data and representing it as textual information.

[0079] A "natural language generation algorithm" refers to a computer algorithm used to generate natural-sounding language text based on text data.

[0080] A "speech recognition interface" refers to a program or service used to convert speech data into text data.

[0081] An "Application Programming Interface Request" refers to a request used for different software systems to communicate with each other.

[0082] A "communication system" refers to a collection of devices and programs that enable users and systems to interact through voice.

[0083] This invention is a system that integrates speech recognition, natural language generation, and speech synthesis technologies to enable travelers and others to communicate smoothly across language barriers. This system is implemented based on the roles of server, terminal, and user.

[0084] 1. User

[0085] The user inputs their questions and requests into the device using voice. For example, consider a scenario where the user asks, "Where is the nearest pharmacy?"

[0086] 2. Terminal

[0087] The device records the user's voice using its built-in microphone. After recording, the audio data is sent to the server using an HTTP request. At this time, the audio data is saved in a format such as WAV or MP3, and the HTTP POST protocol is used for transmission.

[0088] 3. Server

[0089] The server receives audio data sent from the terminal. Using a speech recognition API (e.g., a speech recognition interface), it converts this audio data into text data. The converted text data is then sent to a natural language generation AI (e.g., a natural language generation algorithm) to generate an appropriate response. Examples of natural language generation algorithms include widely used models.

[0090] The generated response text is sent to a speech synthesis API (e.g., a speech synthesis interface) and converted into audio data. The final generated audio data is then sent to the terminal as an HTTP response.

[0091] 4. Terminal

[0092] The terminal receives audio data sent from the server and plays it for the user. For example, the terminal uses its built-in speaker to convey the message, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0093] Specific example

[0094] As a concrete example, let's explain a scenario where a user asks, "Where is the nearest pharmacy?" When the user speaks the question into the terminal, the terminal records the voice and sends it to the server. The server uses speech recognition to convert this voice into text data: "Where is the nearest pharmacy?"

[0095] Next, the server sends this text data to a natural language generation AI to request an appropriate response. For example, it might generate a response like, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The generated text data is converted into audio data using a speech synthesis API and sent to the device. The device then plays this audio data to provide the user with the answer.

[0096] Examples of prompt statements to input into a generative AI model include the following:

[0097] "The user wants to know the nearest pharmacy. Question: Where is the nearest pharmacy? Please provide the appropriate address and travel time as your answer."

[0098] This allows users to receive real-time audio answers to their questions, enabling smooth communication across language barriers.

[0099] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0100] Processing steps

[0101] Step 1:

[0102] The user performs voice input. The user speaks their questions or requests into the terminal. For example, they might say, "Where is the nearest pharmacy?" The input data is the voice data of the user's question or request. The output data is the voice data recorded on the terminal.

[0103] Step 2:

[0104] The device records audio and sends it to the server. The device uses its built-in microphone to record audio and saves the audio data in a format such as WAV or MP3. Then, it sends the audio data to the server using an HTTP POST request. The input data is the user's audio data, and the output data is the audio data sent to the server.

[0105] Step 3:

[0106] The server receives the audio data and converts it to text. The server receives an HTTP request and retrieves the audio data. Next, it uses a speech recognition interface (e.g., a speech recognition API) to convert the audio data into text data. The input data is the audio data sent from the terminal, and the output data is the text "Where is the nearest pharmacy?".

[0107] Step 4:

[0108] The server sends the converted text to a natural language generation algorithm to generate a response. The server analyzes the generated text data and sends it as a prompt to the natural language generation algorithm (e.g., a generative AI model) to generate an appropriate response. The input data is the text "Where is the nearest pharmacy?", and the output data is the response text "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0109] Step 5:

[0110] The server converts the generated response text into audio data. The server sends the generated response text data to a speech synthesis interface (e.g., a speech synthesis API) and converts it into audio data. The input data is the generated response text data, and the output data is audio data.

[0111] Step 6:

[0112] The server sends the final audio data to the terminal. The server sends the generated audio data to the terminal as an HTTP response. The input data is the audio data obtained from the speech synthesis API, and the output data is the audio data sent to the terminal.

[0113] Step 7:

[0114] The terminal plays audio data and provides answers to the user. The terminal plays audio data received from the server through its built-in speaker and provides answers to the user. For example, it might tell the user, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The input data is audio data received from the server, and the output data is the audio information the user hears.

[0115] (Application Example 1)

[0116] 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."

[0117] There is a problem in that users of shopping facilities lack the means to quickly and accurately obtain information within the facility. In particular, users who speak different languages ​​often have difficulty obtaining the necessary information, so effective communication methods are needed.

[0118] 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.

[0119] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to a natural language generation AI and receiving the generated response text data, means for converting the response text data into voice data, and means for guiding the user to information within the shopping facility in order to provide appropriate answers based on the user's questions. This enables the user to obtain answers to their questions in real time and to smoothly obtain user-oriented information, overcoming language barriers.

[0120] A "user" is someone who operates the system and uses voice input.

[0121] "Audio data" refers to data that records user-generated audio in digital format.

[0122] "Text data" refers to the transcribed text data of speech converted by speech recognition.

[0123] "Natural language generation AI" refers to artificial intelligence that generates appropriate responses based on text data entered by the user.

[0124] A "speech recognition API" is an application programming interface for converting speech data into text data.

[0125] "Answer text data" refers to the text data of the answer generated by the natural language generation AI.

[0126] A "speech synthesis API" is an application programming interface for converting text data into speech data.

[0127] A "shopping facility" refers to a commercial area where many shops and facilities are gathered together.

[0128] An "information guidance means" is a means of providing relevant information within a shopping facility based on a user's question.

[0129] A "server" is a central computer system that processes, converts, generates, and redistributes audio data.

[0130] A specific system for implementing this invention provides voice-based information guidance to users within a shopping facility. The embodiments of this system are described in detail below.

[0131] Hardware and software configuration

[0132] This system uses terminals, servers, and multiple APIs. A terminal refers to a device where the user performs voice input and receives the results as voice. The server is a central computer system that processes voice data, converts it to text data, generates natural language, and synthesizes speech.

[0133] Hardware: A device requiring a microphone, speaker, and internet connection (e.g., smartphone, dedicated information terminal).

[0134] software:

[0135] Speech recognition API (e.g., Google Speech Recognition API)

[0136] Natural language generation AI (e.g. OpenAI GPT-3)

[0137] Text-to-speech APIs (e.g., Google Text-to-Speech, gTTS)

[0138] Data processing and data calculation

[0139] 1. Acquisition of audio data:

[0140] The device acquires voice data from the user and stores it in digital format.

[0141] 2. Speech recognition:

[0142] The device uses a speech recognition API to convert the acquired audio data into text data. For example, it uses the Google Speech Recognition API.

[0143] 3. Natural Language Processing:

[0144] The server receives text data and sends it to a natural language generation AI (e.g., GPT-3) to generate appropriate response text data.

[0145] 4. Speech synthesis:

[0146] The generated response text data is converted into speech data using a speech synthesis API, such as Google Text-to-Speech or gTTS.

[0147] 5. Audio Playback:

[0148] The terminal plays audio data transmitted from the server and provides it to the user.

[0149] Specific example

[0150] When a user asks the terminal, "Where is the nearest cafe?", the terminal will act as follows:

[0151] 1. The device receives voice input and saves it as voice data.

[0152] 2. Use a speech recognition API to convert the text into "Where is the nearest cafe?".

[0153] 3. This text data is sent to the server, which then passes it to a natural language generation AI to generate a response. For example, the response might be, "There is a cafe on the east side of the second floor."

[0154] 4. The generated text data is converted into speech data using a speech synthesis API.

[0155] 5. Finally, the device plays this audio data and provides it to the user.

[0156] Example of a prompt

[0157] The prompt message when a user asks "Where is the restroom?" is as follows:

[0158] A user is asking about the location of restrooms in a shopping mall. Please generate appropriate answers to the following questions.

[0159] User question: Where is the restroom?

[0160] In this way, a system is realized that allows users to receive necessary information via voice within shopping facilities without experiencing language barriers. This invention significantly improves the user experience within shopping facilities.

[0161] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0162] Step 1:

[0163] The device acquires voice data from the user.

[0164] When a user speaks into the device, the voice data is saved digitally via the microphone. The input is the user's voice question, and the output is digital voice data.

[0165] Step 2:

[0166] The device uses a speech recognition API to convert the audio data into text data.

[0167] The device sends the stored audio data to a speech recognition API, which, for example, uses the Google Speech Recognition API to convert the audio to text. The input is audio data in digital format, and the output is question text as character data.

[0168] Step 3:

[0169] The terminal sends text data to the server.

[0170] The converted text data is sent to the server via an HTTP request. The input is text data, and the output is the text data sent to the server.

[0171] Step 4:

[0172] The server sends text data to the natural language generation AI and receives the generated response text data.

[0173] The server sends the received text data to a natural language generation AI (e.g., GPT-3) to generate an appropriate response. The input is the text data of the user's question, and the output is the generated response text data.

[0174] Step 5:

[0175] The server sends the response text data to a speech synthesis API, where it is converted into speech data.

[0176] The generated response text data is sent to a speech synthesis API (e.g., Google Text-to-Speech or gTTS) and converted into speech data. The input is the response text data, and the output is the synthesized speech data.

[0177] Step 6:

[0178] The server sends the audio data to the terminal.

[0179] The generated audio data is sent from the server to the terminal in the form of an HTTP response or similar. The input is synthesized speech data, and the output is the transmission of audio data to the terminal.

[0180] Step 7:

[0181] The device plays audio data and provides it to the user.

[0182] The device plays the received audio data through its speaker and provides the user with an answer. The input is the received audio data, and the output is the audio answer that the user can hear.

[0183] 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.

[0184] This invention is a system that integrates speech recognition, natural language generation, speech synthesis, and emotion recognition technologies to enable users to communicate smoothly across language barriers. In particular, by adding a function that recognizes the user's emotions and generates the optimal response based on them, it achieves more natural and effective communication.

[0185] User

[0186] The user inputs their questions or requests into the device using voice. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as audio data.

[0187] terminal

[0188] The user's voice data is sent to the server. Specifically, the voice file is uploaded to the server using an HTTP POST request.

[0189] server

[0190] Receive and temporarily store audio data.

[0191] This process uses a speech recognition API to convert received audio data into text data. For example, it uses Google's speech recognition API to convert the audio into the text "Where is the nearest pharmacy?".

[0192] The voice data is passed to an emotion engine, which recognizes the user's emotions. For example, the emotion engine can detect emotions such as "excited" or "tired."

[0193] The recognized text data and user sentiment information are sent to a natural language generation AI to generate appropriate response text data. For example, OpenAI's GPT-3 is used. In this process, the content and tone of the response are adjusted according to the user's sentiment.

[0194] The generated response text data is sent to a speech synthesis API and converted into speech data. For example, using the Google Text-to-Speech API or gTTS, the speech data "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated.

[0195] The final audio data is sent to the terminal. This audio data is sent in the form of an HTTP response or similar.

[0196] terminal

[0197] Receive and save audio data sent from the server.

[0198] The system plays back the received audio data and provides the user with an answer. This allows the user to hear the answer, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0199] Specific example

[0200] Let's take a scenario where a user asks, "Where is the nearest pharmacy?" The user inputs the question by voice, and the device records the voice and sends it to the server. The server uses speech recognition to convert it into text, "Where is the nearest pharmacy?" Next, the emotion engine detects the user's level of excitement. Based on this emotional information, the natural language generation AI generates a calming response in a soothing tone: "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The generated response text is converted into audio data and sent to the device. The device plays the received audio data and provides it to the user.

[0201] This system allows users to receive real-time voice responses to their questions, and these responses are appropriately adjusted according to the user's emotions. This enables more natural and effective communication.

[0202] The following describes the processing flow.

[0203] Step 1:

[0204] The user makes a voice input to the device. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as an audio file.

[0205] Step 2:

[0206] The device sends the recorded audio file to the server. Specifically, it uploads the audio data to the server using an HTTP POST request.

[0207] Step 3:

[0208] The server receives the audio file and uses a speech recognition API to convert the audio data into text data. The received audio data is passed to Google's speech recognition API to obtain the text data "Where is the nearest pharmacy?".

[0209] Step 4:

[0210] The server passes the voice data to the emotion engine, which then recognizes the user's emotions. The emotion engine analyzes the voice to extract emotional information, such as "excited" or "tired."

[0211] Step 5:

[0212] The server sends the converted text data and sentiment information obtained by the sentiment engine to the natural language generation AI to generate response text data. Using OpenAI's GPT-3 API, the response is generated based on the text data "Where is the nearest pharmacy?" and the sentiment information. Based on the sentiment information, for example, a response such as "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated in a calming tone.

[0213] Step 6:

[0214] The server sends the generated response text data to a text-to-speech API to convert it into speech data. For example, it uses the Google Text-to-Speech API to generate speech data such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0215] Step 7:

[0216] The server sends the generated audio data to the terminal as an HTTP response. The audio data is converted to the appropriate format so that the terminal can receive it.

[0217] Step 8:

[0218] The device saves the received audio data and plays it back to the user. The device plays the audio data, and the user can hear the response, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0219] This allows users to engage in natural, real-time interactions and receive voice responses, as well as emotionally sensitive responses.

[0220] (Example 2)

[0221] 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".

[0222] Conventional speech recognition systems have a problem in that they cannot achieve effective and natural communication with users because they convert speech data into text data and provide generated responses without considering the user's emotions. In particular, it has been difficult to provide responses with appropriate tone and content that correspond to the user's emotional state.

[0223] 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.

[0224] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to an emotion recognition engine and acquiring the user's emotion information, means for transmitting the text data to a natural language generation device based on the emotion information and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. This enables natural and effective communication that takes the user's emotions into consideration.

[0225] A "user" refers to a person who uses the system to input information using voice commands.

[0226] "Audio data" refers to data that records the voice spoken by a user in digital format.

[0227] "Text data" refers to data obtained by converting audio data into a string format.

[0228] An "emotion recognition engine" refers to a device or program that analyzes and determines a user's emotional state from text data or audio data.

[0229] A "natural language generation device" refers to a device or program that generates responses in natural language that is easy for humans to understand, based on input text data and emotional information.

[0230] A "data transmission device" refers to a device or program for transmitting text data or audio data to other devices or systems over a network.

[0231] A "speech recognition device" refers to a device or program that analyzes speech data and converts it into text data.

[0232] "Answer text data" refers to data that represents, in string format, the answers to user questions and requests generated by a natural language generation device.

[0233] This invention is a system that integrates speech recognition, natural language generation, speech synthesis, and emotion recognition technologies to enable users to communicate smoothly across language barriers. This system acquires voice data from the user, converts that voice data into text data, further recognizes the user's emotions to generate appropriate responses, and finally provides them to the user as voice data.

[0234] The user inputs questions and requests by voice into the terminal. For example, they might say, "Where is the nearest pharmacy?" The terminal uses a microphone and a voice recording API to record the voice and save it as audio data. The recorded audio data is sent to the server using an HTTP POST request.

[0235] The server temporarily stores the received audio data and converts it into text data using a speech recognition API (for example, a commonly used speech recognition service). The converted text data is then transformed into a string of text, such as "Where is the nearest pharmacy?".

[0236] Next, the server passes text and audio data to the emotion recognition engine to recognize the user's emotions. The emotion recognition engine analyzes the tone of voice, speaking speed, volume, etc., to determine the user's emotional state. For example, the emotion recognition engine might detect that the user is "excited."

[0237] Next, the server sends the recognized text data and user sentiment information to a natural language generator. Here, the natural language generator uses, for example, a commonly available natural language generation technology (AI model). The generated response is adjusted in content and tone according to the user's sentiment. For example, a prompt might be given stating, "The user is agitated, please generate a calming response."

[0238] The generated response text data is converted into speech data using a speech synthesis API (e.g., a commonly available speech synthesis service). For example, the speech data "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated.

[0239] Finally, the server sends the generated audio data to the terminal. The terminal plays the received audio data and provides it to the user. This allows the user to receive the response in audio format.

[0240] This system allows users to receive real-time audio answers to their questions, with the responses appropriately adjusted according to the user's emotions. This enables more natural and effective communication.

[0241] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0242] Step 1:

[0243] The user performs voice input.

[0244] Users input questions and requests into the terminal using voice. For example, they might say, "Where is the nearest pharmacy?"

[0245] Input: User's spoken voice

[0246] The device uses a microphone and an audio recording API to record this audio and save it as audio data.

[0247] Output: Audio data stored on the device

[0248] Step 2:

[0249] The device sends voice data to the server.

[0250] The device sends the recorded audio data to the server using an HTTP POST request.

[0251] Input: Audio data stored on the device

[0252] Specifically, the request contains binary data of the audio.

[0253] Output: Audio data embedded in the HTTP request

[0254] Step 3:

[0255] The server receives the audio data.

[0256] The server receives audio data via an API that has an endpoint for receiving audio data, and stores it temporarily.

[0257] Input: Audio data embedded in an HTTP request

[0258] Output: Audio data temporarily stored on the server

[0259] Step 4:

[0260] The server converts the audio data into text data.

[0261] The server uses a speech recognition API (assuming a typical speech recognition service) to convert the audio data into text data.

[0262] Input: Audio data stored on the server

[0263] A request is sent to a speech recognition API, and text data is received as a response from the API. For example, the string data "Where is the nearest pharmacy?" is obtained.

[0264] Output: Converted text data

[0265] Step 5:

[0266] The server performs emotion recognition.

[0267] The server passes text and audio data to the emotion recognition engine to recognize the user's emotions.

[0268] Input: Converted text data and audio data

[0269] The emotion recognition engine analyzes the tone, speed, and volume of the voice to determine the user's emotional state. For example, it might detect the emotion of "excitement."

[0270] Output: User sentiment information

[0271] Step 6:

[0272] The server generates the appropriate answer.

[0273] The server sends emotion information and text data to the natural language generation device to generate an appropriate response.

[0274] Input: Transformed text data and the user's emotion information

[0275] Based on the emotion information, the natural language generation device generates a response according to, for example, the prompt sentence "Since the user is excited, please generate an answer to calm them down."

[0276] Output: Generated response text data

[0277] Step 7:

[0278] The server converts the response text data into voice data

[0279] The server sends the generated response text data to a voice synthesis API (using a general voice synthesis service) to convert it into voice data.

[0280] Input: Generated response text data

[0281] Send the text data to the voice synthesis API and receive the voice data as the response from the API. For example, voice data such as "The nearest pharmacy is in front of JR Koutouji Station. It takes about 5 minutes on foot." is generated.

[0282] Output: Generated response voice data

[0283] [[ID=A1]]Step 8:

[0284] The server sends the final voice data to the terminal

[0285] The server sends the generated voice data to the terminal as an HTTP response.

[0286] Input: Generated response voice data

[0287] The audio data is sent to the terminal as an HTTP response.

[0288] Output: Audio data sent to the terminal

[0289] Step 9:

[0290] The device receives and plays the audio data.

[0291] The terminal receives and stores the audio data sent from the server.

[0292] Input: Sent audio data

[0293] Using the built-in voice playback function, the system plays audio data to the user and provides an answer. This allows the user to hear an answer such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0294] Output: Played response audio

[0295] The above describes the specific processing steps of the program.

[0296] (Application Example 2)

[0297] 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".

[0298] Conventional speech recognition systems have struggled to generate natural responses that take into account the user's emotions. Furthermore, there has been a lack of technology to automatically provide optimal responses based on the user's emotions. This has hindered improvements in the user experience, and flexible responses to diverse situations are particularly needed in services where emotion recognition is crucial, such as food delivery.

[0299] The specific processing by the specific processing unit 290 of the data processing apparatus 12 in Application Example 2 is realized by the following means.

[0300] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to means for recognizing emotions to recognize the user's emotions, means for transmitting the text data to natural language generation AI based on the emotion information and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. Thereby, it becomes possible to provide a natural and appropriate response according to the user's emotions.

[0301] The "user" refers to a person who makes voice inquiries or instructions using the system.

[0302] The "voice data" refers to sound information recording the user's speech.

[0303] The "text data" refers to data that is the result of converting voice data into character information.

[0304] The "means for recognizing emotions" refers to an algorithm or system for discriminating the user's emotions from text data.

[0305] The "natural language generation AI" refers to a type of artificial intelligence that automatically generates documents or sentences based on given input data.

[0306] The "response text data" refers to character data as a response generated by the natural language generation AI.

[0307] The "voice recognition API" refers to an application programming interface for converting voice data into text data.

[0308] The "means for converting into voice data" refers to a process or system for converting text data into voice sounds.

[0309] An "API request" refers to the act of software sending a request to access an API (Application Programming Interface) and use its services.

[0310] A "prompt sentence" refers to a predefined input sentence used to give instructions to a natural language generation AI.

[0311] This invention provides a system for food delivery services that integrates speech recognition, emotion recognition, natural language generation, and speech synthesis technologies to provide an optimal response tailored to the user's emotions when the user places an order or makes an inquiry using their voice.

[0312] Users place orders or make inquiries by voice into their smartphone's microphone. This voice data is acquired by the device and sent to the server. The server is configured with various APIs and algorithms, and performs the following processing.

[0313] 1. Speech recognition:

[0314] The server converts the acquired audio data into text data using Google's speech recognition API. This converted text data represents the user's spoken content.

[0315] 2. Emotion recognition:

[0316] The converted text data is sent to an emotion recognition system to determine the user's emotional state. For example, if a user says, "I'm really hungry right now, and I want to order a pizza," the emotion recognition engine will determine that the user's emotion is "excited."

[0317] 3. Natural language generation:

[0318] The emotional information and text data obtained through emotion recognition are sent to a natural language generation AI such as OpenAI. In this process, appropriate response text data that takes the user's emotions into account is generated. For example, if the prompt is given as "The customer is feeling excited. Respond to their query: 'I'm really hungry right now, but I'd like to order a pizza'", a response will be generated.

[0319] 4. Speech synthesis:

[0320] The generated response text data is converted into audio data using a speech synthesis API such as Google's Text-to-Speech API. This results in the response being produced as spoken audio.

[0321] 5. Providing a response:

[0322] The audio data generated in the process described above is sent to the device, which then plays the audio data. The user receives an audio response such as, "You ordered pizza, right? We recommend the Margherita. We'll deliver it within 30 minutes."

[0323] The key to this system is its ability to respond flexibly while taking into account the user's emotional state. A concrete example of the invention is a user utterance such as, "I'm really hungry right now, and I'd like to order a pizza." In this case, the emotion of "being hungry" is recognized, and a response in a friendly tone appropriate to that emotion is generated.

[0324] This will improve the user experience in food delivery services and enable more natural and effective communication.

[0325] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0326] Step 1:

[0327] The user speaks their order or inquiry into the microphone of their smartphone. For example, they might say, "I'd like to order a pizza." This voice data is then captured by the device.

[0328] Input: User's voice data

[0329] Output: Audio data stored on the device

[0330] Step 2:

[0331] The device sends the user's voice data to the server. Specifically, it uploads the voice file using an HTTP POST request.

[0332] Input: Audio data stored on the device

[0333] Output: Audio data sent to the server

[0334] Step 3:

[0335] The server converts the received audio data into text data using a speech recognition API (for example, Google's speech recognition API).

[0336] Input: Audio data stored on the server

[0337] Output: Converted text data (e.g., "I would like to order a pizza")

[0338] Step 4:

[0339] The server sends the text data to an emotion recognition engine (for example, the emotion recognition model in Hugging Face) to recognize the user's emotions.

[0340] Input: Converted text data

[0341] Output: Recognized emotion information (e.g., "excited")

[0342] Step 5:

[0343] The server sends text data and recognized sentiment information to a natural language generation AI (e.g., OpenAI's GPT-3) to generate appropriate response text data. The prompt is "The customer is feeling excited. Respond to their query: 'I would like to order a pizza.'"

[0344] Input: Text data and sentiment information

[0345] Output: Generated response text data (Example: "Your order is pizza. We recommend the Margherita. We will deliver it within 30 minutes.")

[0346] Step 6:

[0347] The server passes the generated response text data to a speech synthesis API (for example, Google Text-to-Speech API) to generate audio data.

[0348] Input: Generated response text data

[0349] Output: Generated audio data (Example: "You ordered a pizza, correct? We recommend the Margherita. We'll deliver it within 30 minutes." - audio recording)

[0350] Step 7:

[0351] The server sends the generated audio data to the terminal. The terminal receives this audio data and plays it.

[0352] Input: Generated audio data

[0353] Output: Voice response provided to the user

[0354] The above outlines the process for generating voice responses that respond to user emotions in a food delivery service. Through this process, users can receive natural and appropriate voice responses to their utterances.

[0355] 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.

[0356] Data generation model 58 is a type of 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 those described above. 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 shown 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.

[0357] 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.

[0358] [Second Embodiment]

[0359] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0360] 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.

[0361] 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).

[0362] 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.

[0363] 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.

[0364] 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).

[0365] 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.

[0366] 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.

[0367] 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.

[0368] 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.

[0369] 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.

[0370] 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".

[0371] This invention is a system that integrates speech recognition, natural language generation, and speech synthesis technologies to enable travelers and others to communicate smoothly across language barriers. This system is implemented as follows.

[0372] User

[0373] The user inputs their questions and requests into the device using voice. For example, they might voice-input the question, "Where is the nearest pharmacy?"

[0374] terminal

[0375] The user's voice is recorded and saved as audio data.

[0376] The recorded audio data is sent to the server. Here, the audio data is sent, for example, using an HTTP request.

[0377] server

[0378] Receive audio data.

[0379] Use a speech recognition API to convert received audio data into text data. For example, use Google's speech recognition API.

[0380] The converted text data is sent to a natural language generation AI to generate appropriate response text data. Here, for example, OpenAI's GPT-3 is used as the generation AI.

[0381] The generated response text data is sent to a speech synthesis API and converted into speech data. Examples of speech synthesis APIs used include Google Text-to-Speech and gTTS.

[0382] The final audio data is sent to the device as the response. This audio data is sent in the form of an HTTP response or similar.

[0383] terminal

[0384] Receive audio data sent from the server.

[0385] The system plays back the received audio data and provides the user with an answer.

[0386] Specific example

[0387] Let's take a scenario where a user asks, "Where is the nearest pharmacy?" as an example. When the user inputs the question by voice, the device records the voice and sends it to the server. The server uses speech recognition to convert it into text, "Where is the nearest pharmacy?", and sends that text to a natural language generation AI to obtain an answer. For example, if the answer is, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk," this is converted into audio data and sent to the device. The device then plays the received audio data and provides it to the user.

[0388] This allows users to receive real-time audio answers to their questions, enabling smoother communication across language barriers.

[0389] The following describes the processing flow.

[0390] Step 1:

[0391] The user makes a voice input to the device. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as audio data.

[0392] Step 2:

[0393] The device sends the recorded audio data to the server. Specifically, it uploads the audio file to the server using an HTTP POST request.

[0394] Step 3:

[0395] The server receives the audio data. After receiving it, it temporarily stores the audio data and converts it into text data using a speech recognition API. For example, it uses Google's speech recognition API to convert the audio into the text "Where is the nearest pharmacy?".

[0396] Step 4:

[0397] The server sends the converted text data to a natural language generation AI. For example, it calls OpenAI's GPT-3 API to generate an appropriate answer. In this case, it sends the text data "Where is the nearest pharmacy?" and receives the generated answer text data.

[0398] Step 5:

[0399] The server sends the received response text data to a speech synthesis API and converts it into speech data. Specifically, it uses the Google Text-to-Speech API or gTTS to generate speech data such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0400] Step 6:

[0401] The server sends the generated audio data to the terminal as an HTTP response. Here, it is sent as an audio file and converted to a format that the terminal can receive.

[0402] Step 7:

[0403] The device saves the received audio data and plays it back to the user. This allows the user to hear the response, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0404] These steps enable users to interact naturally in real time and effectively overcome language barriers.

[0405] (Example 1)

[0406] 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."

[0407] Conventional speech recognition technologies are highly dependent on specific languages ​​and environments, making it difficult to achieve smooth communication for travelers or in situations requiring cross-language communication. In particular, existing systems have limitations in response speed and accuracy when real-time, accurate translation and responses are required. Furthermore, the process of converting speech data to text data and then converting the response back to speech data is complex and time-consuming, posing a significant challenge.

[0408] 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.

[0409] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to a natural language generation algorithm and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. This enables users to communicate smoothly and accurately in different languages ​​in real time, and facilitates smooth conversations even when used by travelers or in multilingual environments.

[0410] "User" refers to an individual or group that uses this system.

[0411] "Audio data" refers to digital data that records the content of a user's speech.

[0412] "Text data" refers to data obtained by analyzing audio data and representing it as textual information.

[0413] A "natural language generation algorithm" refers to a computer algorithm used to generate natural-sounding language text based on text data.

[0414] A "speech recognition interface" refers to a program or service used to convert speech data into text data.

[0415] An "Application Programming Interface Request" refers to a request used for different software systems to communicate with each other.

[0416] A "communication system" refers to a collection of devices and programs that enable users and systems to interact through voice.

[0417] This invention is a system that integrates speech recognition, natural language generation, and speech synthesis technologies to enable travelers and others to communicate smoothly across language barriers. This system is implemented based on the roles of server, terminal, and user.

[0418] 1. User

[0419] The user inputs their questions and requests into the device using voice. For example, consider a scenario where the user asks, "Where is the nearest pharmacy?"

[0420] 2. Terminal

[0421] The device records the user's voice using its built-in microphone. After recording, the audio data is sent to the server using an HTTP request. At this time, the audio data is saved in a format such as WAV or MP3, and the HTTP POST protocol is used for transmission.

[0422] 3. Server

[0423] The server receives audio data sent from the terminal. Using a speech recognition API (e.g., a speech recognition interface), it converts this audio data into text data. The converted text data is then sent to a natural language generation AI (e.g., a natural language generation algorithm) to generate an appropriate response. Examples of natural language generation algorithms include widely used models.

[0424] The generated response text is sent to a speech synthesis API (e.g., a speech synthesis interface) and converted into audio data. The final generated audio data is then sent to the terminal as an HTTP response.

[0425] 4. Terminal

[0426] The terminal receives audio data sent from the server and plays it for the user. For example, the terminal uses its built-in speaker to convey the message, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0427] Specific example

[0428] As a concrete example, let's explain a scenario where a user asks, "Where is the nearest pharmacy?" When the user speaks the question into the terminal, the terminal records the voice and sends it to the server. The server uses speech recognition to convert this voice into text data: "Where is the nearest pharmacy?"

[0429] Next, the server sends this text data to a natural language generation AI to request an appropriate response. For example, it might generate a response like, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The generated text data is converted into audio data using a speech synthesis API and sent to the device. The device then plays this audio data to provide the user with the answer.

[0430] Examples of prompt statements to input into a generative AI model include the following:

[0431] "The user wants to know the nearest pharmacy. Question: Where is the nearest pharmacy? Please provide the appropriate address and travel time as your answer."

[0432] This allows users to receive real-time audio answers to their questions, enabling smooth communication across language barriers.

[0433] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0434] Processing steps

[0435] Step 1:

[0436] The user performs voice input. The user speaks their questions or requests into the terminal. For example, they might say, "Where is the nearest pharmacy?" The input data is the voice data of the user's question or request. The output data is the voice data recorded on the terminal.

[0437] Step 2:

[0438] The device records audio and sends it to the server. The device uses its built-in microphone to record audio and saves the audio data in a format such as WAV or MP3. Then, it sends the audio data to the server using an HTTP POST request. The input data is the user's audio data, and the output data is the audio data sent to the server.

[0439] Step 3:

[0440] The server receives the audio data and converts it to text. The server receives an HTTP request and retrieves the audio data. Next, it uses a speech recognition interface (e.g., a speech recognition API) to convert the audio data into text data. The input data is the audio data sent from the terminal, and the output data is the text "Where is the nearest pharmacy?".

[0441] Step 4:

[0442] The server sends the converted text to a natural language generation algorithm to generate a response. The server analyzes the generated text data and sends it as a prompt to the natural language generation algorithm (e.g., a generative AI model) to generate an appropriate response. The input data is the text "Where is the nearest pharmacy?", and the output data is the response text "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0443] Step 5:

[0444] The server converts the generated response text into audio data. The server sends the generated response text data to a speech synthesis interface (e.g., a speech synthesis API) and converts it into audio data. The input data is the generated response text data, and the output data is audio data.

[0445] Step 6:

[0446] The server sends the final audio data to the terminal. The server sends the generated audio data to the terminal as an HTTP response. The input data is the audio data obtained from the speech synthesis API, and the output data is the audio data sent to the terminal.

[0447] Step 7:

[0448] The terminal plays audio data and provides answers to the user. The terminal plays audio data received from the server through its built-in speaker and provides answers to the user. For example, it might tell the user, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The input data is audio data received from the server, and the output data is the audio information the user hears.

[0449] (Application Example 1)

[0450] 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."

[0451] There is a problem in that users of shopping facilities lack the means to quickly and accurately obtain information within the facility. In particular, users who speak different languages ​​often have difficulty obtaining the necessary information, so effective communication methods are needed.

[0452] 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.

[0453] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to a natural language generation AI and receiving the generated response text data, means for converting the response text data into voice data, and means for guiding the user to information within the shopping facility in order to provide appropriate answers based on the user's questions. This enables the user to obtain answers to their questions in real time and to smoothly obtain user-oriented information, overcoming language barriers.

[0454] A "user" is someone who operates the system and uses voice input.

[0455] "Audio data" refers to data that records user-generated audio in digital format.

[0456] "Text data" refers to the transcribed text data of speech converted by speech recognition.

[0457] "Natural language generation AI" refers to artificial intelligence that generates appropriate responses based on text data entered by the user.

[0458] A "speech recognition API" is an application programming interface for converting speech data into text data.

[0459] "Answer text data" refers to the text data of the answer generated by the natural language generation AI.

[0460] A "speech synthesis API" is an application programming interface for converting text data into speech data.

[0461] A "shopping facility" refers to a commercial area where many shops and facilities are gathered together.

[0462] An "information guidance means" is a means of providing relevant information within a shopping facility based on a user's question.

[0463] A "server" is a central computer system that processes, converts, generates, and redistributes audio data.

[0464] A specific system for implementing this invention provides voice-based information guidance to users within a shopping facility. The embodiments of this system are described in detail below.

[0465] Hardware and software configuration

[0466] This system uses terminals, servers, and multiple APIs. A terminal refers to a device where the user performs voice input and receives the results as voice. The server is a central computer system that processes voice data, converts it to text data, generates natural language, and synthesizes speech.

[0467] Hardware: A device requiring a microphone, speaker, and internet connection (e.g., smartphone, dedicated information terminal).

[0468] software:

[0469] Speech recognition API (e.g., Google Speech Recognition API)

[0470] Natural language generation AI (e.g. OpenAI GPT-3)

[0471] Text-to-speech APIs (e.g., Google Text-to-Speech, gTTS)

[0472] Data processing and data calculation

[0473] 1. Acquisition of audio data:

[0474] The device acquires voice data from the user and stores it in digital format.

[0475] 2. Speech recognition:

[0476] The device uses a speech recognition API to convert the acquired audio data into text data. For example, it uses the Google Speech Recognition API.

[0477] 3. Natural Language Processing:

[0478] The server receives text data and sends it to a natural language generation AI (e.g., GPT-3) to generate appropriate response text data.

[0479] 4. Speech synthesis:

[0480] The generated response text data is converted into speech data using a speech synthesis API, such as Google Text-to-Speech or gTTS.

[0481] 5. Audio Playback:

[0482] The terminal plays audio data transmitted from the server and provides it to the user.

[0483] Specific example

[0484] When a user asks the terminal, "Where is the nearest cafe?", the terminal will act as follows:

[0485] 1. The device receives voice input and saves it as voice data.

[0486] 2. Use a speech recognition API to convert the text into "Where is the nearest cafe?".

[0487] 3. This text data is sent to the server, which then passes it to a natural language generation AI to generate a response. For example, the response might be, "There is a cafe on the east side of the second floor."

[0488] 4. The generated text data is converted into speech data using a speech synthesis API.

[0489] 5. Finally, the device plays this audio data and provides it to the user.

[0490] Example of a prompt

[0491] The prompt message when a user asks "Where is the restroom?" is as follows:

[0492] A user is asking about the location of restrooms in a shopping mall. Please generate appropriate answers to the following questions.

[0493] User question: Where is the restroom?

[0494] In this way, a system is realized that allows users to receive necessary information via voice within shopping facilities without experiencing language barriers. This invention significantly improves the user experience within shopping facilities.

[0495] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0496] Step 1:

[0497] The device acquires voice data from the user.

[0498] When a user speaks into the device, the voice data is saved digitally via the microphone. The input is the user's voice question, and the output is digital voice data.

[0499] Step 2:

[0500] The device uses a speech recognition API to convert the audio data into text data.

[0501] The device sends the stored audio data to a speech recognition API, which, for example, uses the Google Speech Recognition API to convert the audio to text. The input is audio data in digital format, and the output is question text as character data.

[0502] Step 3:

[0503] The terminal sends text data to the server.

[0504] The converted text data is sent to the server via an HTTP request. The input is text data, and the output is the text data sent to the server.

[0505] Step 4:

[0506] The server sends text data to the natural language generation AI and receives the generated response text data.

[0507] The server sends the received text data to a natural language generation AI (e.g., GPT-3) to generate an appropriate response. The input is the text data of the user's question, and the output is the generated response text data.

[0508] Step 5:

[0509] The server sends the response text data to a speech synthesis API, where it is converted into speech data.

[0510] The generated response text data is sent to a speech synthesis API (e.g., Google Text-to-Speech or gTTS) and converted into speech data. The input is the response text data, and the output is the synthesized speech data.

[0511] Step 6:

[0512] The server sends the audio data to the terminal.

[0513] The generated audio data is sent from the server to the terminal in the form of an HTTP response or similar. The input is synthesized speech data, and the output is the transmission of audio data to the terminal.

[0514] Step 7:

[0515] The device plays audio data and provides it to the user.

[0516] The device plays the received audio data through its speaker and provides the user with an answer. The input is the received audio data, and the output is the audio answer that the user can hear.

[0517] 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.

[0518] This invention is a system that integrates speech recognition, natural language generation, speech synthesis, and emotion recognition technologies to enable users to communicate smoothly across language barriers. In particular, by adding a function that recognizes the user's emotions and generates the optimal response based on them, it achieves more natural and effective communication.

[0519] User

[0520] The user inputs their questions or requests into the device using voice. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as audio data.

[0521] terminal

[0522] The user's voice data is sent to the server. Specifically, the voice file is uploaded to the server using an HTTP POST request.

[0523] server

[0524] Receive and temporarily store audio data.

[0525] This process uses a speech recognition API to convert received audio data into text data. For example, it uses Google's speech recognition API to convert the audio into the text "Where is the nearest pharmacy?".

[0526] The voice data is passed to an emotion engine, which recognizes the user's emotions. For example, the emotion engine can detect emotions such as "excited" or "tired."

[0527] The recognized text data and user sentiment information are sent to a natural language generation AI to generate appropriate response text data. For example, OpenAI's GPT-3 is used. In this process, the content and tone of the response are adjusted according to the user's sentiment.

[0528] The generated response text data is sent to a speech synthesis API and converted into speech data. For example, using the Google Text-to-Speech API or gTTS, the speech data "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated.

[0529] The final audio data is sent to the terminal. This audio data is sent in the form of an HTTP response or similar.

[0530] terminal

[0531] Receive and save audio data sent from the server.

[0532] The system plays back the received audio data and provides the user with an answer. This allows the user to hear the answer, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0533] Specific example

[0534] Let's take a scenario where a user asks, "Where is the nearest pharmacy?" The user inputs the question by voice, and the device records the voice and sends it to the server. The server uses speech recognition to convert it into text, "Where is the nearest pharmacy?" Next, the emotion engine detects the user's level of excitement. Based on this emotional information, the natural language generation AI generates a calming response in a soothing tone: "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The generated response text is converted into audio data and sent to the device. The device plays the received audio data and provides it to the user.

[0535] This system allows users to receive real-time voice responses to their questions, and these responses are appropriately adjusted according to the user's emotions. This enables more natural and effective communication.

[0536] The following describes the processing flow.

[0537] Step 1:

[0538] The user makes a voice input to the device. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as an audio file.

[0539] Step 2:

[0540] The device sends the recorded audio file to the server. Specifically, it uploads the audio data to the server using an HTTP POST request.

[0541] Step 3:

[0542] The server receives the audio file and uses a speech recognition API to convert the audio data into text data. The received audio data is passed to Google's speech recognition API to obtain the text data "Where is the nearest pharmacy?".

[0543] Step 4:

[0544] The server passes the voice data to the emotion engine, which then recognizes the user's emotions. The emotion engine analyzes the voice to extract emotional information, such as "excited" or "tired."

[0545] Step 5:

[0546] The server sends the converted text data and sentiment information obtained by the sentiment engine to the natural language generation AI to generate response text data. Using OpenAI's GPT-3 API, the response is generated based on the text data "Where is the nearest pharmacy?" and the sentiment information. Based on the sentiment information, for example, a response such as "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated in a calming tone.

[0547] Step 6:

[0548] The server sends the generated response text data to a text-to-speech API to convert it into speech data. For example, it uses the Google Text-to-Speech API to generate speech data such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0549] Step 7:

[0550] The server sends the generated audio data to the terminal as an HTTP response. The audio data is converted to the appropriate format so that the terminal can receive it.

[0551] Step 8:

[0552] The device saves the received audio data and plays it back to the user. The device plays the audio data, and the user can hear the response, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0553] This allows users to engage in natural, real-time interactions and receive voice responses, as well as emotionally sensitive responses.

[0554] (Example 2)

[0555] 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".

[0556] Conventional speech recognition systems have a problem in that they cannot achieve effective and natural communication with users because they convert speech data into text data and provide generated responses without considering the user's emotions. In particular, it has been difficult to provide responses with appropriate tone and content that correspond to the user's emotional state.

[0557] 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.

[0558] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to an emotion recognition engine and acquiring the user's emotion information, means for transmitting the text data to a natural language generation device based on the emotion information and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. This enables natural and effective communication that takes the user's emotions into consideration.

[0559] A "user" refers to a person who uses the system to input information using voice commands.

[0560] "Audio data" refers to data that records the voice spoken by a user in digital format.

[0561] "Text data" refers to data obtained by converting audio data into a string format.

[0562] An "emotion recognition engine" refers to a device or program that analyzes and determines a user's emotional state from text data or audio data.

[0563] A "natural language generation device" refers to a device or program that generates responses in natural language that is easy for humans to understand, based on input text data and emotional information.

[0564] A "data transmission device" refers to a device or program for transmitting text data or audio data to other devices or systems over a network.

[0565] A "speech recognition device" refers to a device or program that analyzes speech data and converts it into text data.

[0566] "Answer text data" refers to data that represents, in string format, the answers to user questions and requests generated by a natural language generation device.

[0567] This invention is a system that integrates speech recognition, natural language generation, speech synthesis, and emotion recognition technologies to enable users to communicate smoothly across language barriers. This system acquires voice data from the user, converts that voice data into text data, further recognizes the user's emotions to generate appropriate responses, and finally provides them to the user as voice data.

[0568] The user inputs questions and requests by voice into the terminal. For example, they might say, "Where is the nearest pharmacy?" The terminal uses a microphone and a voice recording API to record the voice and save it as audio data. The recorded audio data is sent to the server using an HTTP POST request.

[0569] The server temporarily stores the received audio data and converts it into text data using a speech recognition API (for example, a commonly used speech recognition service). The converted text data is then transformed into a string of text, such as "Where is the nearest pharmacy?".

[0570] Next, the server passes text and audio data to the emotion recognition engine to recognize the user's emotions. The emotion recognition engine analyzes the tone of voice, speaking speed, volume, etc., to determine the user's emotional state. For example, the emotion recognition engine might detect that the user is "excited."

[0571] Next, the server sends the recognized text data and user sentiment information to a natural language generator. Here, the natural language generator uses, for example, a commonly available natural language generation technology (AI model). The generated response is adjusted in content and tone according to the user's sentiment. For example, a prompt might be given stating, "The user is agitated, please generate a calming response."

[0572] The generated response text data is converted into speech data using a speech synthesis API (e.g., a commonly available speech synthesis service). For example, the speech data "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated.

[0573] Finally, the server sends the generated audio data to the terminal. The terminal plays the received audio data and provides it to the user. This allows the user to receive the response in audio format.

[0574] This system allows users to receive real-time audio answers to their questions, with the responses appropriately adjusted according to the user's emotions. This enables more natural and effective communication.

[0575] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0576] Step 1:

[0577] The user performs voice input.

[0578] Users input questions and requests into the terminal using voice. For example, they might say, "Where is the nearest pharmacy?"

[0579] Input: User's spoken voice

[0580] The device uses a microphone and an audio recording API to record this audio and save it as audio data.

[0581] Output: Audio data stored on the device

[0582] Step 2:

[0583] The device sends voice data to the server.

[0584] The device sends the recorded audio data to the server using an HTTP POST request.

[0585] Input: Audio data stored on the device

[0586] Specifically, the request contains binary data of the audio.

[0587] Output: Audio data embedded in the HTTP request

[0588] Step 3:

[0589] The server receives the audio data.

[0590] The server receives audio data via an API that has an endpoint for receiving audio data, and stores it temporarily.

[0591] Input: Audio data embedded in an HTTP request

[0592] Output: Audio data temporarily stored on the server

[0593] Step 4:

[0594] The server converts the audio data into text data.

[0595] The server uses a speech recognition API (assuming a typical speech recognition service) to convert the audio data into text data.

[0596] Input: Audio data stored on the server

[0597] A request is sent to a speech recognition API, and text data is received as a response from the API. For example, the string data "Where is the nearest pharmacy?" is obtained.

[0598] Output: Converted text data

[0599] Step 5:

[0600] The server performs emotion recognition.

[0601] The server passes text and audio data to the emotion recognition engine to recognize the user's emotions.

[0602] Input: Converted text data and audio data

[0603] The emotion recognition engine analyzes the tone, speed, and volume of the voice to determine the user's emotional state. For example, it might detect the emotion of "excitement."

[0604] Output: User sentiment information

[0605] Step 6:

[0606] The server generates the appropriate answer.

[0607] The server sends emotional information and text data to a natural language generator to produce an appropriate response.

[0608] Input: Converted text data and user sentiment information

[0609] The natural language generator uses emotional information to generate responses, for example, according to a prompt such as "The user is agitated, please generate a calming response."

[0610] Output: Generated response text data

[0611] Step 7:

[0612] The server converts the response text data into audio data.

[0613] The server sends the generated response text data to a speech synthesis API (using a common speech synthesis service) and converts it into speech data.

[0614] Input: Generated response text data

[0615] Text data is sent to a speech synthesis API, and speech data is received as a response from the API. For example, the speech data "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated.

[0616] Output: Generated response audio data

[0617] Step 8:

[0618] The server sends the final audio data to the terminal.

[0619] The server sends the generated audio data to the terminal as an HTTP response.

[0620] Input: Generated response audio data

[0621] The audio data is sent to the terminal as an HTTP response.

[0622] Output: Audio data sent to the terminal

[0623] Step 9:

[0624] The device receives and plays the audio data.

[0625] The terminal receives and stores the audio data sent from the server.

[0626] Input: Sent audio data

[0627] Using the built-in voice playback function, the system plays audio data to the user and provides an answer. This allows the user to hear an answer such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0628] Output: Played response audio

[0629] The above describes the specific processing steps of the program.

[0630] (Application Example 2)

[0631] 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."

[0632] Conventional speech recognition systems have struggled to generate natural responses that take into account the user's emotions. Furthermore, there has been a lack of technology to automatically provide optimal responses based on the user's emotions. This has hindered improvements in the user experience, and flexible responses to diverse situations are particularly needed in services where emotion recognition is crucial, such as food delivery.

[0633] 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.

[0634] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to means for recognizing emotions and recognizing the user's emotions, means for transmitting the text data to a natural language generation AI based on the emotion information and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. This makes it possible to provide natural and appropriate responses that correspond to the user's emotions.

[0635] A "user" refers to anyone who uses the system to make inquiries or give instructions via voice.

[0636] "Audio data" refers to sound information recorded from the user's speech.

[0637] "Text data" refers to data that results from converting audio data into text information.

[0638] "Means of recognizing emotions" refers to algorithms and systems that determine a user's emotions from text data.

[0639] "Natural language generation AI" refers to a type of artificial intelligence that automatically generates documents and texts based on given input data.

[0640] "Response text data" refers to character data generated as a response by a natural language generation AI.

[0641] A "speech recognition API" refers to an application programming interface for converting speech data into text data.

[0642] "Means of converting to audio data" refers to the process or system of converting text data into audio.

[0643] An "API request" refers to the act of software sending a request to access an API (Application Programming Interface) and use its services.

[0644] A "prompt sentence" refers to a predefined input sentence used to give instructions to a natural language generation AI.

[0645] This invention provides a system for food delivery services that integrates speech recognition, emotion recognition, natural language generation, and speech synthesis technologies to provide an optimal response tailored to the user's emotions when the user places an order or makes an inquiry using their voice.

[0646] Users place orders or make inquiries by voice into their smartphone's microphone. This voice data is acquired by the device and sent to the server. The server is configured with various APIs and algorithms, and performs the following processing.

[0647] 1. Speech recognition:

[0648] The server converts the acquired audio data into text data using Google's speech recognition API. This converted text data represents the user's spoken content.

[0649] 2. Emotion recognition:

[0650] The converted text data is sent to an emotion recognition system to determine the user's emotional state. For example, if a user says, "I'm really hungry right now, and I want to order a pizza," the emotion recognition engine will determine that the user's emotion is "excited."

[0651] 3. Natural language generation:

[0652] The emotional information and text data obtained through emotion recognition are sent to a natural language generation AI such as OpenAI. In this process, appropriate response text data that takes the user's emotions into account is generated. For example, if the prompt is given as "The customer is feeling excited. Respond to their query: 'I'm really hungry right now, but I'd like to order a pizza'", a response will be generated.

[0653] 4. Speech synthesis:

[0654] The generated response text data is converted into audio data using a speech synthesis API such as Google's Text-to-Speech API. This results in the response being produced as spoken audio.

[0655] 5. Providing a response:

[0656] The audio data generated in the process described above is sent to the device, which then plays this audio data. The user receives an audio response such as, "You ordered pizza, right? We recommend the Margherita. We'll deliver it within 30 minutes."

[0657] The key to this system is its ability to respond flexibly while taking into account the user's emotional state. A concrete example of the invention is a user utterance such as, "I'm really hungry right now, and I'd like to order a pizza." In this case, the emotion of "being hungry" is recognized, and a response in a friendly tone appropriate to that emotion is generated.

[0658] This will improve the user experience in food delivery services and enable more natural and effective communication.

[0659] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0660] Step 1:

[0661] The user speaks their order or inquiry into the microphone of their smartphone. For example, they might say, "I'd like to order a pizza." This voice data is then captured by the device.

[0662] Input: User's voice data

[0663] Output: Audio data stored on the device

[0664] Step 2:

[0665] The device sends the user's voice data to the server. Specifically, it uploads the voice file using an HTTP POST request.

[0666] Input: Audio data stored on the device

[0667] Output: Audio data sent to the server

[0668] Step 3:

[0669] The server converts the received audio data into text data using a speech recognition API (for example, Google's speech recognition API).

[0670] Input: Audio data stored on the server

[0671] Output: Converted text data (e.g., "I would like to order a pizza")

[0672] Step 4:

[0673] The server sends the text data to an emotion recognition engine (for example, the emotion recognition model in Hugging Face) to recognize the user's emotions.

[0674] Input: Converted text data

[0675] Output: Recognized emotion information (e.g., "excited")

[0676] Step 5:

[0677] The server sends text data and recognized sentiment information to a natural language generation AI (e.g., OpenAI's GPT-3) to generate appropriate response text data. The prompt is "The customer is feeling excited. Respond to their query: 'I would like to order a pizza.'"

[0678] Input: Text data and sentiment information

[0679] Output: Generated response text data (Example: "Your order is pizza. We recommend the Margherita. We will deliver it within 30 minutes.")

[0680] Step 6:

[0681] The server passes the generated response text data to a speech synthesis API (for example, Google Text-to-Speech API) to generate the audio data.

[0682] Input: Generated response text data

[0683] Output: Generated audio data (Example: "You ordered a pizza, correct? We recommend the Margherita. We'll deliver it within 30 minutes." - audio recording)

[0684] Step 7:

[0685] The server sends the generated audio data to the terminal. The terminal receives this audio data and plays it.

[0686] Input: Generated audio data

[0687] Output: Voice response provided to the user

[0688] The above outlines the process for generating voice responses that respond to user emotions in a food delivery service. Through this process, users can receive natural and appropriate voice responses to their utterances.

[0689] 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.

[0690] 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 those described above. 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 shown 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.

[0691] 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.

[0692] [Third Embodiment]

[0693] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0694] 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.

[0695] 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).

[0696] 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.

[0697] 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.

[0698] 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).

[0699] 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.

[0700] 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.

[0701] 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.

[0702] 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.

[0703] 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.

[0704] 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".

[0705] This invention is a system that integrates speech recognition, natural language generation, and speech synthesis technologies to enable travelers and others to communicate smoothly across language barriers. This system is implemented as follows.

[0706] User

[0707] The user inputs their questions and requests into the device using voice. For example, they might voice-input the question, "Where is the nearest pharmacy?"

[0708] terminal

[0709] The user's voice is recorded and saved as audio data.

[0710] The recorded audio data is sent to the server. Here, the audio data is sent, for example, using an HTTP request.

[0711] server

[0712] Receive audio data.

[0713] Use a speech recognition API to convert received audio data into text data. For example, use Google's speech recognition API.

[0714] The converted text data is sent to a natural language generation AI to generate appropriate response text data. Here, for example, OpenAI's GPT-3 is used as the generation AI.

[0715] The generated response text data is sent to a speech synthesis API and converted into speech data. Examples of speech synthesis APIs used include Google Text-to-Speech and gTTS.

[0716] The final audio data is sent to the device as the response. This audio data is sent in the form of an HTTP response or similar.

[0717] terminal

[0718] Receive audio data sent from the server.

[0719] The system plays back the received audio data and provides the user with an answer.

[0720] Specific example

[0721] Let's take a scenario where a user asks, "Where is the nearest pharmacy?" as an example. When the user inputs the question by voice, the device records the voice and sends it to the server. The server uses speech recognition to convert it into text, "Where is the nearest pharmacy?", and sends that text to a natural language generation AI to obtain an answer. For example, if the answer is, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk," this is converted into audio data and sent to the device. The device then plays the received audio data and provides it to the user.

[0722] This allows users to receive real-time audio answers to their questions, enabling smoother communication across language barriers.

[0723] The following describes the processing flow.

[0724] Step 1:

[0725] The user makes a voice input to the device. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as audio data.

[0726] Step 2:

[0727] The device sends the recorded audio data to the server. Specifically, it uploads the audio file to the server using an HTTP POST request.

[0728] Step 3:

[0729] The server receives the audio data. After receiving it, it temporarily stores the audio data and converts it into text data using a speech recognition API. For example, it uses Google's speech recognition API to convert the audio into the text "Where is the nearest pharmacy?".

[0730] Step 4:

[0731] The server sends the converted text data to a natural language generation AI. For example, it calls OpenAI's GPT-3 API to generate an appropriate answer. In this case, it sends the text data "Where is the nearest pharmacy?" and receives the generated answer text data.

[0732] Step 5:

[0733] The server sends the received response text data to a speech synthesis API and converts it into speech data. Specifically, it uses the Google Text-to-Speech API or gTTS to generate speech data such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0734] Step 6:

[0735] The server sends the generated audio data to the terminal as an HTTP response. Here, it is sent as an audio file and converted to a format that the terminal can receive.

[0736] Step 7:

[0737] The device saves the received audio data and plays it back to the user. This allows the user to hear the response, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0738] These steps enable users to interact naturally in real time and effectively overcome language barriers.

[0739] (Example 1)

[0740] 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."

[0741] Conventional speech recognition technologies are highly dependent on specific languages ​​and environments, making it difficult to achieve smooth communication for travelers or in situations requiring cross-language communication. In particular, existing systems have limitations in response speed and accuracy when real-time, accurate translation and responses are required. Furthermore, the process of converting speech data to text data and then converting the response back to speech data is complex and time-consuming, posing a significant challenge.

[0742] 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.

[0743] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to a natural language generation algorithm and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. This enables users to communicate smoothly and accurately in different languages ​​in real time, and facilitates smooth conversations even when used by travelers or in multilingual environments.

[0744] "User" refers to an individual or group that uses this system.

[0745] "Audio data" refers to digital data that records the content of a user's speech.

[0746] "Text data" refers to data obtained by analyzing audio data and representing it as textual information.

[0747] A "natural language generation algorithm" refers to a computer algorithm used to generate natural-sounding language text based on text data.

[0748] A "speech recognition interface" refers to a program or service used to convert speech data into text data.

[0749] An "Application Programming Interface Request" refers to a request used for different software systems to communicate with each other.

[0750] A "communication system" refers to a collection of devices and programs that enable users and systems to interact through voice.

[0751] This invention is a system that integrates speech recognition, natural language generation, and speech synthesis technologies to enable travelers and others to communicate smoothly across language barriers. This system is implemented based on the roles of server, terminal, and user.

[0752] 1. User

[0753] The user inputs their questions and requests into the device using voice. For example, consider a scenario where the user asks, "Where is the nearest pharmacy?"

[0754] 2. Terminal

[0755] The device records the user's voice using its built-in microphone. After recording, the audio data is sent to the server using an HTTP request. At this time, the audio data is saved in a format such as WAV or MP3, and the HTTP POST protocol is used for transmission.

[0756] 3. Server

[0757] The server receives audio data sent from the terminal. Using a speech recognition API (e.g., a speech recognition interface), it converts this audio data into text data. The converted text data is then sent to a natural language generation AI (e.g., a natural language generation algorithm) to generate an appropriate response. Examples of natural language generation algorithms include widely used models.

[0758] The generated response text is sent to a speech synthesis API (e.g., a speech synthesis interface) and converted into audio data. The final generated audio data is then sent to the terminal as an HTTP response.

[0759] 4. Terminal

[0760] The terminal receives audio data sent from the server and plays it for the user. For example, the terminal uses its built-in speaker to convey the message, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0761] Specific example

[0762] As a concrete example, let's explain a scenario where a user asks, "Where is the nearest pharmacy?" When the user speaks the question into the terminal, the terminal records the voice and sends it to the server. The server uses speech recognition to convert this voice into text data: "Where is the nearest pharmacy?"

[0763] Next, the server sends this text data to a natural language generation AI to request an appropriate response. For example, it might generate a response like, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The generated text data is converted into audio data using a speech synthesis API and sent to the device. The device then plays this audio data to provide the user with the answer.

[0764] Examples of prompt statements to input into a generative AI model include the following:

[0765] "The user wants to know the nearest pharmacy. Question: Where is the nearest pharmacy? Please provide the appropriate address and travel time as your answer."

[0766] This allows users to receive real-time audio answers to their questions, enabling smooth communication across language barriers.

[0767] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0768] Processing steps

[0769] Step 1:

[0770] The user performs voice input. The user speaks their questions or requests into the terminal. For example, they might say, "Where is the nearest pharmacy?" The input data is the voice data of the user's question or request. The output data is the voice data recorded on the terminal.

[0771] Step 2:

[0772] The device records audio and sends it to the server. The device uses its built-in microphone to record audio and saves the audio data in a format such as WAV or MP3. Then, it sends the audio data to the server using an HTTP POST request. The input data is the user's audio data, and the output data is the audio data sent to the server.

[0773] Step 3:

[0774] The server receives the audio data and converts it to text. The server receives an HTTP request and retrieves the audio data. Next, it uses a speech recognition interface (e.g., a speech recognition API) to convert the audio data into text data. The input data is the audio data sent from the terminal, and the output data is the text "Where is the nearest pharmacy?".

[0775] Step 4:

[0776] The server sends the converted text to a natural language generation algorithm to generate a response. The server analyzes the generated text data and sends it as a prompt to the natural language generation algorithm (e.g., a generative AI model) to generate an appropriate response. The input data is the text "Where is the nearest pharmacy?", and the output data is the response text "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0777] Step 5:

[0778] The server converts the generated response text into audio data. The server sends the generated response text data to a speech synthesis interface (e.g., a speech synthesis API) and converts it into audio data. The input data is the generated response text data, and the output data is audio data.

[0779] Step 6:

[0780] The server sends the final audio data to the terminal. The server sends the generated audio data to the terminal as an HTTP response. The input data is the audio data obtained from the speech synthesis API, and the output data is the audio data sent to the terminal.

[0781] Step 7:

[0782] The terminal plays audio data and provides answers to the user. The terminal plays audio data received from the server through its built-in speaker and provides answers to the user. For example, it might tell the user, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The input data is audio data received from the server, and the output data is the audio information the user hears.

[0783] (Application Example 1)

[0784] 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."

[0785] There is a problem in that users of shopping facilities lack the means to quickly and accurately obtain information within the facility. In particular, users who speak different languages ​​often have difficulty obtaining the necessary information, so effective communication methods are needed.

[0786] 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.

[0787] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to a natural language generation AI and receiving the generated response text data, means for converting the response text data into voice data, and means for guiding the user to information within the shopping facility in order to provide appropriate answers based on the user's questions. This enables the user to obtain answers to their questions in real time and to smoothly obtain user-oriented information, overcoming language barriers.

[0788] A "user" is someone who operates the system and uses voice input.

[0789] "Audio data" refers to data that records user-generated audio in digital format.

[0790] "Text data" refers to the transcribed text data of speech converted by speech recognition.

[0791] "Natural language generation AI" refers to artificial intelligence that generates appropriate responses based on text data entered by the user.

[0792] A "speech recognition API" is an application programming interface for converting speech data into text data.

[0793] "Answer text data" refers to the text data of the answer generated by the natural language generation AI.

[0794] A "speech synthesis API" is an application programming interface for converting text data into speech data.

[0795] A "shopping facility" refers to a commercial area where many shops and facilities are gathered together.

[0796] An "information guidance means" is a means of providing relevant information within a shopping facility based on a user's question.

[0797] A "server" is a central computer system that processes, converts, generates, and redistributes audio data.

[0798] A specific system for implementing this invention provides voice-based information guidance to users within a shopping facility. The embodiments of this system are described in detail below.

[0799] Hardware and software configuration

[0800] This system uses terminals, servers, and multiple APIs. A terminal refers to a device where the user performs voice input and receives the results as voice. The server is a central computer system that processes voice data, converts it to text data, generates natural language, and synthesizes speech.

[0801] Hardware: A device requiring a microphone, speaker, and internet connection (e.g., smartphone, dedicated information terminal).

[0802] software:

[0803] Speech recognition API (e.g., Google Speech Recognition API)

[0804] Natural language generation AI (e.g. OpenAI GPT-3)

[0805] Text-to-speech APIs (e.g., Google Text-to-Speech, gTTS)

[0806] Data processing and data calculation

[0807] 1. Acquisition of audio data:

[0808] The device acquires voice data from the user and stores it in digital format.

[0809] 2. Speech recognition:

[0810] The device uses a speech recognition API to convert the acquired audio data into text data. For example, it uses the Google Speech Recognition API.

[0811] 3. Natural Language Processing:

[0812] The server receives text data and sends it to a natural language generation AI (e.g., GPT-3) to generate appropriate response text data.

[0813] 4. Speech synthesis:

[0814] The generated response text data is converted into speech data using a speech synthesis API, such as Google Text-to-Speech or gTTS.

[0815] 5. Audio Playback:

[0816] The terminal plays audio data transmitted from the server and provides it to the user.

[0817] Specific example

[0818] When a user asks the terminal, "Where is the nearest cafe?", the terminal will act as follows:

[0819] 1. The device receives voice input and saves it as voice data.

[0820] 2. Use a speech recognition API to convert the text into "Where is the nearest cafe?".

[0821] 3. This text data is sent to the server, which then passes it to a natural language generation AI to generate a response. For example, the response might be, "There is a cafe on the east side of the second floor."

[0822] 4. The generated text data is converted into speech data using a speech synthesis API.

[0823] 5. Finally, the device plays this audio data and provides it to the user.

[0824] Example of a prompt

[0825] The prompt message when a user asks "Where is the restroom?" is as follows:

[0826] A user is asking about the location of restrooms in a shopping mall. Please generate appropriate answers to the following questions.

[0827] User question: Where is the restroom?

[0828] In this way, a system is realized that allows users to receive necessary information via voice within shopping facilities without experiencing language barriers. This invention significantly improves the user experience within shopping facilities.

[0829] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0830] Step 1:

[0831] The device acquires voice data from the user.

[0832] When a user speaks into the device, the voice data is saved digitally via the microphone. The input is the user's voice question, and the output is digital voice data.

[0833] Step 2:

[0834] The device uses a speech recognition API to convert the audio data into text data.

[0835] The device sends the stored audio data to a speech recognition API, which, for example, uses the Google Speech Recognition API to convert the audio to text. The input is audio data in digital format, and the output is question text as character data.

[0836] Step 3:

[0837] The terminal sends text data to the server.

[0838] The converted text data is sent to the server via an HTTP request. The input is text data, and the output is the text data sent to the server.

[0839] Step 4:

[0840] The server sends text data to the natural language generation AI and receives the generated response text data.

[0841] The server sends the received text data to a natural language generation AI (e.g., GPT-3) to generate an appropriate response. The input is the text data of the user's question, and the output is the generated response text data.

[0842] Step 5:

[0843] The server sends the response text data to a speech synthesis API, where it is converted into speech data.

[0844] The generated response text data is sent to a speech synthesis API (e.g., Google Text-to-Speech or gTTS) and converted into speech data. The input is the response text data, and the output is the synthesized speech data.

[0845] Step 6:

[0846] The server sends the audio data to the terminal.

[0847] The generated audio data is sent from the server to the terminal in the form of an HTTP response or similar. The input is synthesized speech data, and the output is the transmission of audio data to the terminal.

[0848] Step 7:

[0849] The device plays audio data and provides it to the user.

[0850] The device plays the received audio data through its speaker and provides the user with an answer. The input is the received audio data, and the output is the audio answer that the user can hear.

[0851] 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.

[0852] This invention is a system that integrates speech recognition, natural language generation, speech synthesis, and emotion recognition technologies to enable users to communicate smoothly across language barriers. In particular, by adding a function that recognizes the user's emotions and generates the optimal response based on them, it achieves more natural and effective communication.

[0853] User

[0854] The user inputs their questions or requests into the device using voice. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as audio data.

[0855] terminal

[0856] The user's voice data is sent to the server. Specifically, the voice file is uploaded to the server using an HTTP POST request.

[0857] server

[0858] Receive and temporarily store audio data.

[0859] This process uses a speech recognition API to convert received audio data into text data. For example, it uses Google's speech recognition API to convert the audio into the text "Where is the nearest pharmacy?".

[0860] The voice data is passed to an emotion engine, which recognizes the user's emotions. For example, the emotion engine can detect emotions such as "excited" or "tired."

[0861] The recognized text data and user sentiment information are sent to a natural language generation AI to generate appropriate response text data. For example, OpenAI's GPT-3 is used. In this process, the content and tone of the response are adjusted according to the user's sentiment.

[0862] The generated response text data is sent to a speech synthesis API and converted into speech data. For example, using the Google Text-to-Speech API or gTTS, the speech data "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated.

[0863] The final audio data is sent to the terminal. This audio data is sent in the form of an HTTP response or similar.

[0864] terminal

[0865] Receive and save audio data sent from the server.

[0866] The system plays back the received audio data and provides the user with an answer. This allows the user to hear the answer, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0867] Specific example

[0868] Let's take a scenario where a user asks, "Where is the nearest pharmacy?" The user inputs the question by voice, and the device records the voice and sends it to the server. The server uses speech recognition to convert it into text, "Where is the nearest pharmacy?" Next, the emotion engine detects the user's level of excitement. Based on this emotional information, the natural language generation AI generates a calming response in a soothing tone: "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The generated response text is converted into audio data and sent to the device. The device plays the received audio data and provides it to the user.

[0869] This system allows users to receive real-time voice responses to their questions, and these responses are appropriately adjusted according to the user's emotions. This enables more natural and effective communication.

[0870] The following describes the processing flow.

[0871] Step 1:

[0872] The user makes a voice input to the device. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as an audio file.

[0873] Step 2:

[0874] The device sends the recorded audio file to the server. Specifically, it uploads the audio data to the server using an HTTP POST request.

[0875] Step 3:

[0876] The server receives the audio file and uses a speech recognition API to convert the audio data into text data. The received audio data is passed to Google's speech recognition API to obtain the text data "Where is the nearest pharmacy?".

[0877] Step 4:

[0878] The server passes the voice data to the emotion engine, which then recognizes the user's emotions. The emotion engine analyzes the voice to extract emotional information, such as "excited" or "tired."

[0879] Step 5:

[0880] The server sends the converted text data and sentiment information obtained by the sentiment engine to the natural language generation AI to generate response text data. Using OpenAI's GPT-3 API, the response is generated based on the text data "Where is the nearest pharmacy?" and the sentiment information. Based on the sentiment information, for example, a response such as "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated in a calming tone.

[0881] Step 6:

[0882] The server sends the generated response text data to a text-to-speech API to convert it into speech data. For example, it uses the Google Text-to-Speech API to generate speech data such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0883] Step 7:

[0884] The server sends the generated audio data to the terminal as an HTTP response. The audio data is converted to the appropriate format so that the terminal can receive it.

[0885] Step 8:

[0886] The device saves the received audio data and plays it back to the user. The device plays the audio data, and the user can hear the response, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0887] This allows users to engage in natural, real-time interactions and receive voice responses, as well as emotionally sensitive responses.

[0888] (Example 2)

[0889] 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."

[0890] Conventional speech recognition systems have a problem in that they cannot achieve effective and natural communication with users because they convert speech data into text data and provide generated responses without considering the user's emotions. In particular, it has been difficult to provide responses with appropriate tone and content that correspond to the user's emotional state.

[0891] 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.

[0892] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to an emotion recognition engine and acquiring the user's emotion information, means for transmitting the text data to a natural language generation device based on the emotion information and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. This enables natural and effective communication that takes the user's emotions into consideration.

[0893] A "user" refers to a person who uses the system to input information using voice commands.

[0894] "Audio data" refers to data that records the voice spoken by a user in digital format.

[0895] "Text data" refers to data obtained by converting audio data into a string format.

[0896] An "emotion recognition engine" refers to a device or program that analyzes and determines a user's emotional state from text data or audio data.

[0897] A "natural language generation device" refers to a device or program that generates responses in natural language that is easy for humans to understand, based on input text data and emotional information.

[0898] A "data transmission device" refers to a device or program for transmitting text data or audio data to other devices or systems over a network.

[0899] A "speech recognition device" refers to a device or program that analyzes speech data and converts it into text data.

[0900] "Answer text data" refers to data that represents, in string format, the answers to user questions and requests generated by a natural language generation device.

[0901] This invention is a system that integrates speech recognition, natural language generation, speech synthesis, and emotion recognition technologies to enable users to communicate smoothly across language barriers. This system acquires voice data from the user, converts that voice data into text data, further recognizes the user's emotions to generate appropriate responses, and finally provides them to the user as voice data.

[0902] The user inputs questions and requests by voice into the terminal. For example, they might say, "Where is the nearest pharmacy?" The terminal uses a microphone and a voice recording API to record the voice and save it as audio data. The recorded audio data is sent to the server using an HTTP POST request.

[0903] The server temporarily stores the received audio data and converts it into text data using a speech recognition API (for example, a commonly used speech recognition service). The converted text data is then transformed into a string of text, such as "Where is the nearest pharmacy?".

[0904] Next, the server passes text and audio data to the emotion recognition engine to recognize the user's emotions. The emotion recognition engine analyzes the tone of voice, speaking speed, volume, etc., to determine the user's emotional state. For example, the emotion recognition engine might detect that the user is "excited."

[0905] Next, the server sends the recognized text data and user sentiment information to a natural language generator. Here, the natural language generator uses, for example, a commonly available natural language generation technology (AI model). The generated response is adjusted in content and tone according to the user's sentiment. For example, a prompt might be given stating, "The user is agitated, please generate a calming response."

[0906] The generated response text data is converted into speech data using a speech synthesis API (e.g., a commonly available speech synthesis service). For example, the speech data "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated.

[0907] Finally, the server sends the generated audio data to the terminal. The terminal plays the received audio data and provides it to the user. This allows the user to receive the response in audio format.

[0908] This system allows users to receive real-time audio answers to their questions, with the responses appropriately adjusted according to the user's emotions. This enables more natural and effective communication.

[0909] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0910] Step 1:

[0911] The user performs voice input.

[0912] Users input questions and requests into the terminal using voice. For example, they might say, "Where is the nearest pharmacy?"

[0913] Input: User's spoken voice

[0914] The device uses a microphone and an audio recording API to record this audio and save it as audio data.

[0915] Output: Audio data stored on the device

[0916] Step 2:

[0917] The device sends voice data to the server.

[0918] The device sends the recorded audio data to the server using an HTTP POST request.

[0919] Input: Audio data stored on the device

[0920] Specifically, the request contains binary data of the audio.

[0921] Output: Audio data embedded in the HTTP request

[0922] Step 3:

[0923] The server receives the audio data.

[0924] The server receives audio data via an API that has an endpoint for receiving audio data, and stores it temporarily.

[0925] Input: Audio data embedded in an HTTP request

[0926] Output: Audio data temporarily stored on the server

[0927] Step 4:

[0928] The server converts the audio data into text data.

[0929] The server uses a speech recognition API (assuming a typical speech recognition service) to convert the audio data into text data.

[0930] Input: Audio data stored on the server

[0931] A request is sent to a speech recognition API, and text data is received as a response from the API. For example, the string data "Where is the nearest pharmacy?" is obtained.

[0932] Output: Converted text data

[0933] Step 5:

[0934] The server performs emotion recognition.

[0935] The server passes text and audio data to the emotion recognition engine to recognize the user's emotions.

[0936] Input: Converted text data and audio data

[0937] The emotion recognition engine analyzes the tone, speed, and volume of the voice to determine the user's emotional state. For example, it might detect the emotion of "excitement."

[0938] Output: User sentiment information

[0939] Step 6:

[0940] The server generates the appropriate answer.

[0941] The server sends emotional information and text data to a natural language generator to produce an appropriate response.

[0942] Input: Converted text data and user sentiment information

[0943] The natural language generator uses emotional information to generate responses, for example, according to a prompt such as "The user is agitated, please generate a calming response."

[0944] Output: Generated response text data

[0945] Step 7:

[0946] The server converts the response text data into audio data.

[0947] The server sends the generated response text data to a speech synthesis API (using a common speech synthesis service) and converts it into speech data.

[0948] Input: Generated response text data

[0949] Text data is sent to a speech synthesis API, and speech data is received as a response from the API. For example, the speech data "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated.

[0950] Output: Generated response audio data

[0951] Step 8:

[0952] The server sends the final audio data to the terminal.

[0953] The server sends the generated audio data to the terminal as an HTTP response.

[0954] Input: Generated response audio data

[0955] The audio data is sent to the terminal as an HTTP response.

[0956] Output: Audio data sent to the terminal

[0957] Step 9:

[0958] The device receives and plays the audio data.

[0959] The terminal receives and stores the audio data sent from the server.

[0960] Input: Sent audio data

[0961] Using the built-in voice playback function, the system plays audio data to the user and provides an answer. This allows the user to hear an answer such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[0962] Output: Played response audio

[0963] The above describes the specific processing steps of the program.

[0964] (Application Example 2)

[0965] 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."

[0966] Conventional speech recognition systems have struggled to generate natural responses that take into account the user's emotions. Furthermore, there has been a lack of technology to automatically provide optimal responses based on the user's emotions. This has hindered improvements in the user experience, and flexible responses to diverse situations are particularly needed in services where emotion recognition is crucial, such as food delivery.

[0967] 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.

[0968] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to means for recognizing emotions and recognizing the user's emotions, means for transmitting the text data to a natural language generation AI based on the emotion information and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. This makes it possible to provide natural and appropriate responses that correspond to the user's emotions.

[0969] A "user" refers to anyone who uses the system to make inquiries or give instructions via voice.

[0970] "Audio data" refers to sound information recorded from the user's speech.

[0971] "Text data" refers to data that results from converting audio data into text information.

[0972] "Means of recognizing emotions" refers to algorithms and systems that determine a user's emotions from text data.

[0973] "Natural language generation AI" refers to a type of artificial intelligence that automatically generates documents and texts based on given input data.

[0974] "Response text data" refers to character data generated as a response by a natural language generation AI.

[0975] A "speech recognition API" refers to an application programming interface for converting speech data into text data.

[0976] "Means of converting to audio data" refers to the process or system of converting text data into audio.

[0977] An "API request" refers to the act of software sending a request to access an API (Application Programming Interface) and use its services.

[0978] A "prompt sentence" refers to a predefined input sentence used to give instructions to a natural language generation AI.

[0979] This invention provides a system for food delivery services that integrates speech recognition, emotion recognition, natural language generation, and speech synthesis technologies to provide an optimal response tailored to the user's emotions when the user places an order or makes an inquiry using their voice.

[0980] Users place orders or make inquiries by voice into their smartphone's microphone. This voice data is acquired by the device and sent to the server. The server is configured with various APIs and algorithms, and performs the following processing.

[0981] 1. Speech recognition:

[0982] The server converts the acquired audio data into text data using Google's speech recognition API. This converted text data represents the user's spoken content.

[0983] 2. Emotion recognition:

[0984] The converted text data is sent to an emotion recognition system to determine the user's emotional state. For example, if a user says, "I'm really hungry right now, and I want to order a pizza," the emotion recognition engine will determine that the user's emotion is "excited."

[0985] 3. Natural language generation:

[0986] The emotional information and text data obtained through emotion recognition are sent to a natural language generation AI such as OpenAI. In this process, appropriate response text data that takes the user's emotions into account is generated. For example, if the prompt is given as "The customer is feeling excited. Respond to their query: 'I'm really hungry right now, but I'd like to order a pizza'", a response will be generated.

[0987] 4. Speech synthesis:

[0988] The generated response text data is converted into audio data using a speech synthesis API such as Google's Text-to-Speech API. This results in the response being produced as spoken audio.

[0989] 5. Providing a response:

[0990] The audio data generated in the process described above is sent to the device, which then plays this audio data. The user receives an audio response such as, "You ordered pizza, right? We recommend the Margherita. We'll deliver it within 30 minutes."

[0991] The key to this system is its ability to respond flexibly while taking into account the user's emotional state. A concrete example of the invention is a user utterance such as, "I'm really hungry right now, and I'd like to order a pizza." In this case, the emotion of "being hungry" is recognized, and a response in a friendly tone appropriate to that emotion is generated.

[0992] This will improve the user experience in food delivery services and enable more natural and effective communication.

[0993] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0994] Step 1:

[0995] The user speaks their order or inquiry into the microphone of their smartphone. For example, they might say, "I'd like to order a pizza." This voice data is then captured by the device.

[0996] Input: User's voice data

[0997] Output: Audio data stored on the device

[0998] Step 2:

[0999] The device sends the user's voice data to the server. Specifically, it uploads the voice file using an HTTP POST request.

[1000] Input: Audio data stored on the device

[1001] Output: Audio data sent to the server

[1002] Step 3:

[1003] The server converts the received audio data into text data using a speech recognition API (for example, Google's speech recognition API).

[1004] Input: Audio data stored on the server

[1005] Output: Converted text data (e.g., "I would like to order a pizza")

[1006] Step 4:

[1007] The server sends the text data to an emotion recognition engine (for example, the emotion recognition model in Hugging Face) to recognize the user's emotions.

[1008] Input: Converted text data

[1009] Output: Recognized emotion information (e.g., "excited")

[1010] Step 5:

[1011] The server sends text data and recognized sentiment information to a natural language generation AI (e.g., OpenAI's GPT-3) to generate appropriate response text data. The prompt is "The customer is feeling excited. Respond to their query: 'I would like to order a pizza.'"

[1012] Input: Text data and sentiment information

[1013] Output: Generated response text data (Example: "Your order is pizza. We recommend the Margherita. We will deliver it within 30 minutes.")

[1014] Step 6:

[1015] The server passes the generated response text data to a speech synthesis API (for example, Google Text-to-Speech API) to generate the audio data.

[1016] Input: Generated response text data

[1017] Output: Generated audio data (Example: "You ordered a pizza, correct? We recommend the Margherita. We'll deliver it within 30 minutes." - audio recording)

[1018] Step 7:

[1019] The server sends the generated audio data to the terminal. The terminal receives this audio data and plays it.

[1020] Input: Generated audio data

[1021] Output: Voice response provided to the user

[1022] The above outlines the process for generating voice responses that respond to user emotions in a food delivery service. Through this process, users can receive natural and appropriate voice responses to their utterances.

[1023] 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.

[1024] 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 those described above. 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 shown 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.

[1025] 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.

[1026] [Fourth Embodiment]

[1027] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1028] 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.

[1029] 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).

[1030] 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.

[1031] 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.

[1032] 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).

[1033] 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.

[1034] 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.

[1035] 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.

[1036] 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.

[1037] 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.

[1038] 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.

[1039] 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".

[1040] This invention is a system that integrates speech recognition, natural language generation, and speech synthesis technologies to enable travelers and others to communicate smoothly across language barriers. This system is implemented as follows.

[1041] User

[1042] The user inputs their questions and requests into the device using voice. For example, they might voice-input the question, "Where is the nearest pharmacy?"

[1043] terminal

[1044] The user's voice is recorded and saved as audio data.

[1045] The recorded audio data is sent to the server. Here, the audio data is sent, for example, using an HTTP request.

[1046] server

[1047] Receive audio data.

[1048] Use a speech recognition API to convert received audio data into text data. For example, use Google's speech recognition API.

[1049] The converted text data is sent to a natural language generation AI to generate appropriate response text data. Here, for example, OpenAI's GPT-3 is used as the generation AI.

[1050] The generated response text data is sent to a speech synthesis API and converted into speech data. Examples of speech synthesis APIs used include Google Text-to-Speech and gTTS.

[1051] The final audio data is sent to the device as the response. This audio data is sent in the form of an HTTP response or similar.

[1052] terminal

[1053] Receive audio data sent from the server.

[1054] The system plays back the received audio data and provides the user with an answer.

[1055] Specific example

[1056] Let's take a scenario where a user asks, "Where is the nearest pharmacy?" as an example. When the user inputs the question by voice, the device records the voice and sends it to the server. The server uses speech recognition to convert it into text, "Where is the nearest pharmacy?", and sends that text to a natural language generation AI to obtain an answer. For example, if the answer is, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk," this is converted into audio data and sent to the device. The device then plays the received audio data and provides it to the user.

[1057] This allows users to receive real-time audio answers to their questions, enabling smoother communication across language barriers.

[1058] The following describes the processing flow.

[1059] Step 1:

[1060] The user makes a voice input to the device. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as audio data.

[1061] Step 2:

[1062] The device sends the recorded audio data to the server. Specifically, it uploads the audio file to the server using an HTTP POST request.

[1063] Step 3:

[1064] The server receives the audio data. After receiving it, it temporarily stores the audio data and converts it into text data using a speech recognition API. For example, it uses Google's speech recognition API to convert the audio into the text "Where is the nearest pharmacy?".

[1065] Step 4:

[1066] The server sends the converted text data to a natural language generation AI. For example, it calls OpenAI's GPT-3 API to generate an appropriate answer. In this case, it sends the text data "Where is the nearest pharmacy?" and receives the generated answer text data.

[1067] Step 5:

[1068] The server sends the received response text data to a speech synthesis API and converts it into speech data. Specifically, it uses the Google Text-to-Speech API or gTTS to generate speech data such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[1069] Step 6:

[1070] The server sends the generated audio data to the terminal as an HTTP response. Here, it is sent as an audio file and converted to a format that the terminal can receive.

[1071] Step 7:

[1072] The device saves the received audio data and plays it back to the user. This allows the user to hear the response, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[1073] These steps enable users to interact naturally in real time and effectively overcome language barriers.

[1074] (Example 1)

[1075] 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".

[1076] Conventional speech recognition technologies are highly dependent on specific languages ​​and environments, making it difficult to achieve smooth communication for travelers or in situations requiring cross-language communication. In particular, existing systems have limitations in response speed and accuracy when real-time, accurate translation and responses are required. Furthermore, the process of converting speech data to text data and then converting the response back to speech data is complex and time-consuming, posing a significant challenge.

[1077] 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.

[1078] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to a natural language generation algorithm and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. This enables users to communicate smoothly and accurately in different languages ​​in real time, and facilitates smooth conversations even when used by travelers or in multilingual environments.

[1079] "User" refers to an individual or group that uses this system.

[1080] "Audio data" refers to digital data that records the content of a user's speech.

[1081] "Text data" refers to data obtained by analyzing audio data and representing it as textual information.

[1082] A "natural language generation algorithm" refers to a computer algorithm used to generate natural-sounding language text based on text data.

[1083] A "speech recognition interface" refers to a program or service used to convert speech data into text data.

[1084] An "Application Programming Interface Request" refers to a request used for different software systems to communicate with each other.

[1085] A "communication system" refers to a collection of devices and programs that enable users and systems to interact through voice.

[1086] This invention is a system that integrates speech recognition, natural language generation, and speech synthesis technologies to enable travelers and others to communicate smoothly across language barriers. This system is implemented based on the roles of server, terminal, and user.

[1087] 1. User

[1088] The user inputs their questions and requests into the device using voice. For example, consider a scenario where the user asks, "Where is the nearest pharmacy?"

[1089] 2. Terminal

[1090] The device records the user's voice using its built-in microphone. After recording, the audio data is sent to the server using an HTTP request. At this time, the audio data is saved in a format such as WAV or MP3, and the HTTP POST protocol is used for transmission.

[1091] 3. Server

[1092] The server receives audio data sent from the terminal. Using a speech recognition API (e.g., a speech recognition interface), it converts this audio data into text data. The converted text data is then sent to a natural language generation AI (e.g., a natural language generation algorithm) to generate an appropriate response. Examples of natural language generation algorithms include widely used models.

[1093] The generated response text is sent to a speech synthesis API (e.g., a speech synthesis interface) and converted into audio data. The final generated audio data is then sent to the terminal as an HTTP response.

[1094] 4. Terminal

[1095] The terminal receives audio data sent from the server and plays it for the user. For example, the terminal uses its built-in speaker to convey the message, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[1096] Specific example

[1097] As a concrete example, let's explain a scenario where a user asks, "Where is the nearest pharmacy?" When the user speaks the question into the terminal, the terminal records the voice and sends it to the server. The server uses speech recognition to convert this voice into text data: "Where is the nearest pharmacy?"

[1098] Next, the server sends this text data to a natural language generation AI to request an appropriate response. For example, it might generate a response like, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The generated text data is converted into audio data using a speech synthesis API and sent to the device. The device then plays this audio data to provide the user with the answer.

[1099] Examples of prompt statements to input into a generative AI model include the following:

[1100] "The user wants to know the nearest pharmacy. Question: Where is the nearest pharmacy? Please provide the appropriate address and travel time as your answer."

[1101] This allows users to receive real-time audio answers to their questions, enabling smooth communication across language barriers.

[1102] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1103] Processing steps

[1104] Step 1:

[1105] The user performs voice input. The user speaks their questions or requests into the terminal. For example, they might say, "Where is the nearest pharmacy?" The input data is the voice data of the user's question or request. The output data is the voice data recorded on the terminal.

[1106] Step 2:

[1107] The device records audio and sends it to the server. The device uses its built-in microphone to record audio and saves the audio data in a format such as WAV or MP3. Then, it sends the audio data to the server using an HTTP POST request. The input data is the user's audio data, and the output data is the audio data sent to the server.

[1108] Step 3:

[1109] The server receives the audio data and converts it to text. The server receives an HTTP request and retrieves the audio data. Next, it uses a speech recognition interface (e.g., a speech recognition API) to convert the audio data into text data. The input data is the audio data sent from the terminal, and the output data is the text "Where is the nearest pharmacy?".

[1110] Step 4:

[1111] The server sends the converted text to a natural language generation algorithm to generate a response. The server analyzes the generated text data and sends it as a prompt to the natural language generation algorithm (e.g., a generative AI model) to generate an appropriate response. The input data is the text "Where is the nearest pharmacy?", and the output data is the response text "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[1112] Step 5:

[1113] The server converts the generated response text into audio data. The server sends the generated response text data to a speech synthesis interface (e.g., a speech synthesis API) and converts it into audio data. The input data is the generated response text data, and the output data is audio data.

[1114] Step 6:

[1115] The server sends the final audio data to the terminal. The server sends the generated audio data to the terminal as an HTTP response. The input data is the audio data obtained from the speech synthesis API, and the output data is the audio data sent to the terminal.

[1116] Step 7:

[1117] The terminal plays audio data and provides answers to the user. The terminal plays audio data received from the server through its built-in speaker and provides answers to the user. For example, it might tell the user, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The input data is audio data received from the server, and the output data is the audio information the user hears.

[1118] (Application Example 1)

[1119] 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".

[1120] There is a problem in that users of shopping facilities lack the means to quickly and accurately obtain information within the facility. In particular, users who speak different languages ​​often have difficulty obtaining the necessary information, so effective communication methods are needed.

[1121] 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.

[1122] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to a natural language generation AI and receiving the generated response text data, means for converting the response text data into voice data, and means for guiding the user to information within the shopping facility in order to provide appropriate answers based on the user's questions. This enables the user to obtain answers to their questions in real time and to smoothly obtain user-oriented information, overcoming language barriers.

[1123] A "user" is someone who operates the system and uses voice input.

[1124] "Audio data" refers to data that records user-generated audio in digital format.

[1125] "Text data" refers to the transcribed text data of speech converted by speech recognition.

[1126] "Natural language generation AI" refers to artificial intelligence that generates appropriate responses based on text data entered by the user.

[1127] A "speech recognition API" is an application programming interface for converting speech data into text data.

[1128] "Answer text data" refers to the text data of the answer generated by the natural language generation AI.

[1129] A "speech synthesis API" is an application programming interface for converting text data into speech data.

[1130] A "shopping facility" refers to a commercial area where many shops and facilities are gathered together.

[1131] An "information guidance means" is a means of providing relevant information within a shopping facility based on a user's question.

[1132] A "server" is a central computer system that processes, converts, generates, and redistributes audio data.

[1133] A specific system for implementing this invention provides voice-based information guidance to users within a shopping facility. The embodiments of this system are described in detail below.

[1134] Hardware and software configuration

[1135] This system uses terminals, servers, and multiple APIs. A terminal refers to a device where the user performs voice input and receives the results as voice. The server is a central computer system that processes voice data, converts it to text data, generates natural language, and synthesizes speech.

[1136] Hardware: A device requiring a microphone, speaker, and internet connection (e.g., smartphone, dedicated information terminal).

[1137] software:

[1138] Speech recognition API (e.g., Google Speech Recognition API)

[1139] Natural language generation AI (e.g. OpenAI GPT-3)

[1140] Text-to-speech APIs (e.g., Google Text-to-Speech, gTTS)

[1141] Data processing and data calculation

[1142] 1. Acquisition of audio data:

[1143] The device acquires voice data from the user and stores it in digital format.

[1144] 2. Speech recognition:

[1145] The device uses a speech recognition API to convert the acquired audio data into text data. For example, it uses the Google Speech Recognition API.

[1146] 3. Natural Language Processing:

[1147] The server receives text data and sends it to a natural language generation AI (e.g., GPT-3) to generate appropriate response text data.

[1148] 4. Speech synthesis:

[1149] The generated response text data is converted into speech data using a speech synthesis API, such as Google Text-to-Speech or gTTS.

[1150] 5. Audio Playback:

[1151] The terminal plays audio data transmitted from the server and provides it to the user.

[1152] Specific example

[1153] When a user asks the terminal, "Where is the nearest cafe?", the terminal will act as follows:

[1154] 1. The device receives voice input and saves it as voice data.

[1155] 2. Use a speech recognition API to convert the text into "Where is the nearest cafe?".

[1156] 3. This text data is sent to the server, which then passes it to a natural language generation AI to generate a response. For example, the response might be, "There is a cafe on the east side of the second floor."

[1157] 4. The generated text data is converted into speech data using a speech synthesis API.

[1158] 5. Finally, the device plays this audio data and provides it to the user.

[1159] Example of a prompt

[1160] The prompt message when a user asks "Where is the restroom?" is as follows:

[1161] A user is asking about the location of restrooms in a shopping mall. Please generate appropriate answers to the following questions.

[1162] User question: Where is the restroom?

[1163] In this way, a system is realized that allows users to receive necessary information via voice within shopping facilities without experiencing language barriers. This invention significantly improves the user experience within shopping facilities.

[1164] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1165] Step 1:

[1166] The device acquires voice data from the user.

[1167] When a user speaks into the device, the voice data is saved digitally via the microphone. The input is the user's voice question, and the output is digital voice data.

[1168] Step 2:

[1169] The device uses a speech recognition API to convert the audio data into text data.

[1170] The device sends the stored audio data to a speech recognition API, which, for example, uses the Google Speech Recognition API to convert the audio to text. The input is audio data in digital format, and the output is question text as character data.

[1171] Step 3:

[1172] The terminal sends text data to the server.

[1173] The converted text data is sent to the server via an HTTP request. The input is text data, and the output is the text data sent to the server.

[1174] Step 4:

[1175] The server sends text data to the natural language generation AI and receives the generated response text data.

[1176] The server sends the received text data to a natural language generation AI (e.g., GPT-3) to generate an appropriate response. The input is the text data of the user's question, and the output is the generated response text data.

[1177] Step 5:

[1178] The server sends the response text data to a speech synthesis API, where it is converted into speech data.

[1179] The generated response text data is sent to a speech synthesis API (e.g., Google Text-to-Speech or gTTS) and converted into speech data. The input is the response text data, and the output is the synthesized speech data.

[1180] Step 6:

[1181] The server sends the audio data to the terminal.

[1182] The generated audio data is sent from the server to the terminal in the form of an HTTP response or similar. The input is synthesized speech data, and the output is the transmission of audio data to the terminal.

[1183] Step 7:

[1184] The device plays audio data and provides it to the user.

[1185] The device plays the received audio data through its speaker and provides the user with an answer. The input is the received audio data, and the output is the audio answer that the user can hear.

[1186] 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.

[1187] This invention is a system that integrates speech recognition, natural language generation, speech synthesis, and emotion recognition technologies to enable users to communicate smoothly across language barriers. In particular, by adding a function that recognizes the user's emotions and generates the optimal response based on them, it achieves more natural and effective communication.

[1188] User

[1189] The user inputs their questions or requests into the device using voice. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as audio data.

[1190] terminal

[1191] The user's voice data is sent to the server. Specifically, the voice file is uploaded to the server using an HTTP POST request.

[1192] server

[1193] Receive and temporarily store audio data.

[1194] This process uses a speech recognition API to convert received audio data into text data. For example, it uses Google's speech recognition API to convert the audio into the text "Where is the nearest pharmacy?".

[1195] The voice data is passed to an emotion engine, which recognizes the user's emotions. For example, the emotion engine can detect emotions such as "excited" or "tired."

[1196] The recognized text data and user sentiment information are sent to a natural language generation AI to generate appropriate response text data. For example, OpenAI's GPT-3 is used. In this process, the content and tone of the response are adjusted according to the user's sentiment.

[1197] The generated response text data is sent to a speech synthesis API and converted into speech data. For example, using the Google Text-to-Speech API or gTTS, the speech data "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated.

[1198] The final audio data is sent to the terminal. This audio data is sent in the form of an HTTP response or similar.

[1199] terminal

[1200] Receive and save audio data sent from the server.

[1201] The system plays back the received audio data and provides the user with an answer. This allows the user to hear the answer, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[1202] Specific example

[1203] Let's take a scenario where a user asks, "Where is the nearest pharmacy?" The user inputs the question by voice, and the device records the voice and sends it to the server. The server uses speech recognition to convert it into text, "Where is the nearest pharmacy?" Next, the emotion engine detects the user's level of excitement. Based on this emotional information, the natural language generation AI generates a calming response in a soothing tone: "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." The generated response text is converted into audio data and sent to the device. The device plays the received audio data and provides it to the user.

[1204] This system allows users to receive real-time voice responses to their questions, and these responses are appropriately adjusted according to the user's emotions. This enables more natural and effective communication.

[1205] The following describes the processing flow.

[1206] Step 1:

[1207] The user makes a voice input to the device. For example, they might say, "Where is the nearest pharmacy?" The device records this voice and saves it as an audio file.

[1208] Step 2:

[1209] The device sends the recorded audio file to the server. Specifically, it uploads the audio data to the server using an HTTP POST request.

[1210] Step 3:

[1211] The server receives the audio file and uses a speech recognition API to convert the audio data into text data. The received audio data is passed to Google's speech recognition API to obtain the text data "Where is the nearest pharmacy?".

[1212] Step 4:

[1213] The server passes the voice data to the emotion engine, which then recognizes the user's emotions. The emotion engine analyzes the voice to extract emotional information, such as "excited" or "tired."

[1214] Step 5:

[1215] The server sends the converted text data and sentiment information obtained by the sentiment engine to the natural language generation AI to generate response text data. Using OpenAI's GPT-3 API, the response is generated based on the text data "Where is the nearest pharmacy?" and the sentiment information. Based on the sentiment information, for example, a response such as "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated in a calming tone.

[1216] Step 6:

[1217] The server sends the generated response text data to a text-to-speech API to convert it into speech data. For example, it uses the Google Text-to-Speech API to generate speech data such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[1218] Step 7:

[1219] The server sends the generated audio data to the terminal as an HTTP response. The audio data is converted to the appropriate format so that the terminal can receive it.

[1220] Step 8:

[1221] The device saves the received audio data and plays it back to the user. The device plays the audio data, and the user can hear the response, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[1222] This allows users to engage in natural, real-time interactions and receive voice responses, as well as emotionally sensitive responses.

[1223] (Example 2)

[1224] 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".

[1225] Conventional speech recognition systems have a problem in that they cannot achieve effective and natural communication with users because they convert speech data into text data and provide generated responses without considering the user's emotions. In particular, it has been difficult to provide responses with appropriate tone and content that correspond to the user's emotional state.

[1226] 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.

[1227] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to an emotion recognition engine and acquiring the user's emotion information, means for transmitting the text data to a natural language generation device based on the emotion information and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. This enables natural and effective communication that takes the user's emotions into consideration.

[1228] A "user" refers to a person who uses the system to input information using voice commands.

[1229] "Audio data" refers to data that records the voice spoken by a user in digital format.

[1230] "Text data" refers to data obtained by converting audio data into a string format.

[1231] An "emotion recognition engine" refers to a device or program that analyzes and determines a user's emotional state from text data or audio data.

[1232] A "natural language generation device" refers to a device or program that generates responses in natural language that is easy for humans to understand, based on input text data and emotional information.

[1233] A "data transmission device" refers to a device or program for transmitting text data or audio data to other devices or systems over a network.

[1234] A "speech recognition device" refers to a device or program that analyzes speech data and converts it into text data.

[1235] "Answer text data" refers to data that represents, in string format, the answers to user questions and requests generated by a natural language generation device.

[1236] This invention is a system that integrates speech recognition, natural language generation, speech synthesis, and emotion recognition technologies to enable users to communicate smoothly across language barriers. This system acquires voice data from the user, converts that voice data into text data, further recognizes the user's emotions to generate appropriate responses, and finally provides them to the user as voice data.

[1237] The user inputs questions and requests by voice into the terminal. For example, they might say, "Where is the nearest pharmacy?" The terminal uses a microphone and a voice recording API to record the voice and save it as audio data. The recorded audio data is sent to the server using an HTTP POST request.

[1238] The server temporarily stores the received audio data and converts it into text data using a speech recognition API (for example, a commonly used speech recognition service). The converted text data is then transformed into a string of text, such as "Where is the nearest pharmacy?".

[1239] Next, the server passes text and audio data to the emotion recognition engine to recognize the user's emotions. The emotion recognition engine analyzes the tone of voice, speaking speed, volume, etc., to determine the user's emotional state. For example, the emotion recognition engine might detect that the user is "excited."

[1240] Next, the server sends the recognized text data and user sentiment information to a natural language generator. Here, the natural language generator uses, for example, a commonly available natural language generation technology (AI model). The generated response is adjusted in content and tone according to the user's sentiment. For example, a prompt might be given stating, "The user is agitated, please generate a calming response."

[1241] The generated response text data is converted into speech data using a speech synthesis API (e.g., a commonly available speech synthesis service). For example, the speech data "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated.

[1242] Finally, the server sends the generated audio data to the terminal. The terminal plays the received audio data and provides it to the user. This allows the user to receive the response in audio format.

[1243] This system allows users to receive real-time audio answers to their questions, with the responses appropriately adjusted according to the user's emotions. This enables more natural and effective communication.

[1244] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1245] Step 1:

[1246] The user performs voice input.

[1247] Users input questions and requests into the terminal using voice. For example, they might say, "Where is the nearest pharmacy?"

[1248] Input: User's spoken voice

[1249] The device uses a microphone and an audio recording API to record this audio and save it as audio data.

[1250] Output: Audio data stored on the device

[1251] Step 2:

[1252] The device sends voice data to the server.

[1253] The device sends the recorded audio data to the server using an HTTP POST request.

[1254] Input: Audio data stored on the device

[1255] Specifically, the request contains binary data of the audio.

[1256] Output: Audio data embedded in the HTTP request

[1257] Step 3:

[1258] The server receives the audio data.

[1259] The server receives audio data via an API that has an endpoint for receiving audio data, and stores it temporarily.

[1260] Input: Audio data embedded in an HTTP request

[1261] Output: Audio data temporarily stored on the server

[1262] Step 4:

[1263] The server converts the audio data into text data.

[1264] The server uses a speech recognition API (assuming a typical speech recognition service) to convert the audio data into text data.

[1265] Input: Audio data stored on the server

[1266] A request is sent to a speech recognition API, and text data is received as a response from the API. For example, the string data "Where is the nearest pharmacy?" is obtained.

[1267] Output: Converted text data

[1268] Step 5:

[1269] The server performs emotion recognition.

[1270] The server passes text and audio data to the emotion recognition engine to recognize the user's emotions.

[1271] Input: Converted text data and audio data

[1272] The emotion recognition engine analyzes the tone, speed, and volume of the voice to determine the user's emotional state. For example, it might detect the emotion of "excitement."

[1273] Output: User sentiment information

[1274] Step 6:

[1275] The server generates the appropriate answer.

[1276] The server sends emotional information and text data to a natural language generator to produce an appropriate response.

[1277] Input: Converted text data and user sentiment information

[1278] The natural language generator uses emotional information to generate responses, for example, according to a prompt such as "The user is agitated, please generate a calming response."

[1279] Output: Generated response text data

[1280] Step 7:

[1281] The server converts the response text data into audio data.

[1282] The server sends the generated response text data to a speech synthesis API (using a common speech synthesis service) and converts it into speech data.

[1283] Input: Generated response text data

[1284] Text data is sent to a speech synthesis API, and speech data is received as a response from the API. For example, the speech data "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk." is generated.

[1285] Output: Generated response audio data

[1286] Step 8:

[1287] The server sends the final audio data to the terminal.

[1288] The server sends the generated audio data to the terminal as an HTTP response.

[1289] Input: Generated response audio data

[1290] The audio data is sent to the terminal as an HTTP response.

[1291] Output: Audio data sent to the terminal

[1292] Step 9:

[1293] The device receives and plays the audio data.

[1294] The terminal receives and stores the audio data sent from the server.

[1295] Input: Sent audio data

[1296] Using the built-in voice playback function, the system plays audio data to the user and provides an answer. This allows the user to hear an answer such as, "The nearest pharmacy is in front of JR Koenji Station. It's about a 5-minute walk."

[1297] Output: Played response audio

[1298] The above describes the specific processing steps of the program.

[1299] (Application Example 2)

[1300] 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".

[1301] Conventional speech recognition systems have struggled to generate natural responses that take into account the user's emotions. Furthermore, there has been a lack of technology to automatically provide optimal responses based on the user's emotions. This has hindered improvements in the user experience, and flexible responses to diverse situations are particularly needed in services where emotion recognition is crucial, such as food delivery.

[1302] 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.

[1303] In this invention, the server includes means for acquiring voice data from a user, means for converting the voice data into text data, means for transmitting the text data to means for recognizing emotions and recognizing the user's emotions, means for transmitting the text data to a natural language generation AI based on the emotion information and receiving the generated response text data, means for converting the response text data into voice data, and means for providing the voice data to the user. This makes it possible to provide natural and appropriate responses that correspond to the user's emotions.

[1304] A "user" refers to anyone who uses the system to make inquiries or give instructions via voice.

[1305] "Audio data" refers to sound information recorded from the user's speech.

[1306] "Text data" refers to data that results from converting audio data into text information.

[1307] "Means of recognizing emotions" refers to algorithms and systems that determine a user's emotions from text data.

[1308] "Natural language generation AI" refers to a type of artificial intelligence that automatically generates documents and texts based on given input data.

[1309] "Response text data" refers to character data generated as a response by a natural language generation AI.

[1310] A "speech recognition API" refers to an application programming interface for converting speech data into text data.

[1311] "Means of converting to audio data" refers to the process or system of converting text data into audio.

[1312] An "API request" refers to the act of software sending a request to access an API (Application Programming Interface) and use its services.

[1313] A "prompt sentence" refers to a predefined input sentence used to give instructions to a natural language generation AI.

[1314] This invention provides a system for food delivery services that integrates speech recognition, emotion recognition, natural language generation, and speech synthesis technologies to provide an optimal response tailored to the user's emotions when the user places an order or makes an inquiry using their voice.

[1315] Users place orders or make inquiries by voice into their smartphone's microphone. This voice data is acquired by the device and sent to the server. The server is configured with various APIs and algorithms, and performs the following processing.

[1316] 1. Speech recognition:

[1317] The server converts the acquired audio data into text data using Google's speech recognition API. This converted text data represents the user's spoken content.

[1318] 2. Emotion recognition:

[1319] The converted text data is sent to an emotion recognition system to determine the user's emotional state. For example, if a user says, "I'm really hungry right now, and I want to order a pizza," the emotion recognition engine will determine that the user's emotion is "excited."

[1320] 3. Natural language generation:

[1321] The emotional information and text data obtained through emotion recognition are sent to a natural language generation AI such as OpenAI. In this process, appropriate response text data that takes the user's emotions into account is generated. For example, if the prompt is given as "The customer is feeling excited. Respond to their query: 'I'm really hungry right now, but I'd like to order a pizza'", a response will be generated.

[1322] 4. Speech synthesis:

[1323] The generated response text data is converted into audio data using a speech synthesis API such as Google's Text-to-Speech API. This results in the response being produced as spoken audio.

[1324] 5. Providing a response:

[1325] The audio data generated in the process described above is sent to the device, which then plays this audio data. The user receives an audio response such as, "You ordered pizza, right? We recommend the Margherita. We'll deliver it within 30 minutes."

[1326] The key to this system is its ability to respond flexibly while taking into account the user's emotional state. A concrete example of the invention is a user utterance such as, "I'm really hungry right now, and I'd like to order a pizza." In this case, the emotion of "being hungry" is recognized, and a response in a friendly tone appropriate to that emotion is generated.

[1327] This will improve the user experience in food delivery services and enable more natural and effective communication.

[1328] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1329] Step 1:

[1330] The user speaks their order or inquiry into the microphone of their smartphone. For example, they might say, "I'd like to order a pizza." This voice data is then captured by the device.

[1331] Input: User's voice data

[1332] Output: Audio data stored on the device

[1333] Step 2:

[1334] The device sends the user's voice data to the server. Specifically, it uploads the voice file using an HTTP POST request.

[1335] Input: Audio data stored on the device

[1336] Output: Audio data sent to the server

[1337] Step 3:

[1338] The server converts the received audio data into text data using a speech recognition API (for example, Google's speech recognition API).

[1339] Input: Audio data stored on the server

[1340] Output: Converted text data (e.g., "I would like to order a pizza")

[1341] Step 4:

[1342] The server sends the text data to an emotion recognition engine (for example, the emotion recognition model in Hugging Face) to recognize the user's emotions.

[1343] Input: Converted text data

[1344] Output: Recognized emotion information (e.g., "excited")

[1345] Step 5:

[1346] The server sends text data and recognized sentiment information to a natural language generation AI (e.g., OpenAI's GPT-3) to generate appropriate response text data. The prompt is "The customer is feeling excited. Respond to their query: 'I would like to order a pizza.'"

[1347] Input: Text data and sentiment information

[1348] Output: Generated response text data (Example: "Your order is pizza. We recommend the Margherita. We will deliver it within 30 minutes.")

[1349] Step 6:

[1350] The server passes the generated response text data to a speech synthesis API (for example, Google Text-to-Speech API) to generate the audio data.

[1351] Input: Generated response text data

[1352] Output: Generated audio data (Example: "You ordered a pizza, correct? We recommend the Margherita. We'll deliver it within 30 minutes." - audio recording)

[1353] Step 7:

[1354] The server sends the generated audio data to the terminal. The terminal receives this audio data and plays it.

[1355] Input: Generated audio data

[1356] Output: Voice response provided to the user

[1357] The above outlines the process for generating voice responses that respond to user emotions in a food delivery service. Through this process, users can receive natural and appropriate voice responses to their utterances.

[1358] 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.

[1359] 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 those described above. 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 shown 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.

[1360] 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.

[1361] 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.

[1362] 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.

[1363] 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.

[1364] 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.

[1365] 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 based, for example, 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.

[1366] 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."

[1367] 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.

[1368] 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.

[1369] 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.

[1370] 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.

[1371] 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.

[1372] 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.

[1373] 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.

[1374] 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.

[1375] 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.

[1376] 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.

[1377] 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.

[1378] 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.

[1379] The following is further disclosed regarding the embodiments described above.

[1380] (Claim 1)

[1381] A means of obtaining voice data from the user,

[1382] Means for converting the aforementioned audio data into text data,

[1383] A means for transmitting the aforementioned text data to a natural language generation AI and receiving the generated response text data,

[1384] A means for converting the aforementioned response text data into audio data,

[1385] Means for providing the aforementioned audio data to the user,

[1386] A system that includes this.

[1387] (Claim 2)

[1388] The system according to claim 1, wherein the means for converting the audio data into text data uses a speech recognition API.

[1389] (Claim 3)

[1390] The system according to claim 1, wherein the means for transmitting the text data to the natural language generation AI is to use an API request.

[1391] "Example 1"

[1392] (Claim 1)

[1393] A means of obtaining voice data from the user,

[1394] Means for converting the aforementioned audio data into text data,

[1395] The means for transmitting the aforementioned text data to a natural language generation algorithm and receiving the generated response text data,

[1396] A means for converting the aforementioned response text data into audio data,

[1397] Means for providing the aforementioned audio data to the user,

[1398] A communication system that includes this.

[1399] (Claim 2)

[1400] The communication system according to claim 1, wherein the means for converting the voice data into text data uses a voice recognition interface.

[1401] (Claim 3)

[1402] The communication system according to claim 1, wherein the means for transmitting the text data to a natural language generation algorithm uses an application program interface request.

[1403] "Application Example 1"

[1404] (Claim 1)

[1405] A means of obtaining voice data from the user,

[1406] Means for converting the aforementioned audio data into text data,

[1407] A means for transmitting the aforementioned text data to a natural language generation AI and receiving the generated response text data,

[1408] A means for converting the aforementioned response text data into audio data,

[1409] Means for providing the aforementioned audio data to the user,

[1410] In order to provide appropriate answers based on user questions, a means of guiding users to information within the shopping facility,

[1411] A system that includes this.

[1412] (Claim 2)

[1413] The system according to claim 1, wherein the means for converting the aforementioned audio data into text data is to use a speech recognition API.

[1414] (Claim 3)

[1415] The system according to claim 1, wherein the means for transmitting the text data to the natural language generation AI is to use an API request.

[1416] "Example 2 of combining an emotion engine"

[1417] (Claim 1)

[1418] A means of obtaining voice data from the user,

[1419] Means for converting the aforementioned audio data into text data,

[1420] A means for transmitting the aforementioned text data to an emotion recognition engine and obtaining the user's emotion information,

[1421] A means for transmitting text data to a natural language generation device based on the aforementioned sentiment information and receiving the generated response text data,

[1422] A means for converting the aforementioned response text data into audio data,

[1423] Means for providing the aforementioned audio data to the user,

[1424] A system that includes this.

[1425] (Claim 2)

[1426] The system according to claim 1, wherein the means for converting the aforementioned audio data into text data is a speech recognition device.

[1427] (Claim 3)

[1428] The system according to claim 1, wherein the means for transmitting the text data to the natural language generation device is a data transmission device.

[1429] "Application example 2 when combining with an emotional engine"

[1430] (Claim 1)

[1431] A means of obtaining voice data from the user,

[1432] Means for converting the aforementioned audio data into text data,

[1433] The aforementioned text data is transmitted to a means for recognizing emotions, and the means for recognizing the user's emotions,

[1434] A means for transmitting the text data based on the aforementioned emotional information to a natural language generation AI and receiving the generated response text data,

[1435] A means for converting the aforementioned response text data into audio data,

[1436] Means for providing the aforementioned audio data to the user,

[1437] A system that includes this.

[1438] (Claim 2)

[1439] The system according to claim 1, wherein the means for converting the aforementioned audio data into text data is to use a speech recognition API.

[1440] (Claim 3)

[1441] The system according to claim 1, wherein the means for sending the aforementioned text data and sentiment information to a natural language generation AI and generating an appropriate response in a prompt sentence uses an API request. [Explanation of Symbols]

[1442] 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 obtaining voice data from the user, Means for converting the aforementioned audio data into text data, A means for transmitting the aforementioned text data to a natural language generation AI and receiving the generated response text data, A means for converting the aforementioned response text data into audio data, Means for providing the aforementioned audio data to the user, A system that includes this.

2. The system according to claim 1, wherein the means for converting the audio data into text data is a speech recognition API.

3. The system according to claim 1, wherein the means for transmitting the text data to the natural language generation AI is to use an API request.

Images