system

Abstract

We provide the system. [Solution] A means of analyzing user conversation data using natural language processing methods and extracting keywords and context, A means for automatically classifying conversation data into predefined categories based on extracted information, A means of recording classified conversation data in a user history database and organizing it for searchability, A means to identify time-sensitive tasks from conversations and set reminders for users, A means of providing a dashboard for users to access and search categorized conversation history, 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 persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a character of the chatbot, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Modern AI agents can support interactions with users, but the mechanisms for properly managing conversation history and effectively utilizing that information are not well established. In particular, users may require a great deal of time and effort to search for past conversations, and it is difficult to efficiently manage tasks and important information that emerged during conversations. Furthermore, there is a lack of support for users to extract useful information in future conversations. It is required to solve such problems.

Means for Solving the Problems

[0005] This invention uses natural language processing to analyze conversation data with a user, extracting keywords and context. This allows for the automatic classification of conversation data into predefined categories and recording them in the user's history database. Furthermore, the classified data is organized in a searchable format, enabling users to easily access their past conversation history. Additionally, it can identify time-sensitive tasks from conversations and set reminders for the user. Moreover, the natural language processing means perform tokenization, part-of-speech tagging, and named entity recognition, and based on the classified conversation data, it provides category suggestions for future conversations, enabling users to utilize the AI ​​agent more efficiently.

[0006] "Natural language processing means" refers to technologies for analyzing conversation data with users and extracting keywords and context.

[0007] "Conversation data" refers to the record of dialogue between the user and the AI ​​agent, and includes text data and audio data.

[0008] A "category" is a criterion or label used to classify conversational data, such as indicating topics like "shopping" or "asset management."

[0009] A "reminder" is a function or mechanism that notifies users to help them remember deadline-based tasks they have set.

[0010] A "dashboard" is a visual interface that allows users to access and search their conversation history, and it is a screen that organizes and presents information.

[0011] "Tokenization" in natural language processing refers to the process of dividing text into its smallest units, such as words or phrases.

[0012] "Part-of-speech tagging" is a technique in natural language processing that assigns a part of speech (noun, verb, adjective, etc.) to each word.

[0013] "Named entity recognition" is a natural language processing technique that extracts proper nouns such as names of people, places, and organizations from text.

[0014] A "history database" is a database system for recording and managing past conversation data and related information.

[0015] "Category suggestion" is a feature that uses past conversation data to suggest categories that are likely to interest the user in future conversations. [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] Shows an emotion map to which a plurality of emotions are mapped. [Figure 11] It is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Example 2 when the emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when the emotion engine is combined.

Mode for Carrying Out the Invention

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

[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 a plurality of arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of a plurality of types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[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] As an embodiment of the present invention, a system that utilizes an AI agent to efficiently classify and manage conversation data with users will be described.

[0038] Start interaction with the user

[0039] The interaction begins when the user speaks to the AI ​​agent through the device. Input can be provided as either text or voice. In the case of voice input, the device uses speech recognition technology to convert the voice data into text data.

[0040] Processing conversation data

[0041] The terminal sends the converted text data and its metadata to the server. The server receives this data and applies natural language processing (NLP) to it. In the analysis, the text is tokenized and broken down into individual words and phrases, each tagged with its part of speech. Then, named entity recognition is used to extract specific proper nouns.

[0042] Data classification and organization

[0043] Based on the analysis results, the server classifies the conversation data into predefined categories. This classification utilizes automated machine learning algorithms to determine the appropriate category based on the similarity of keywords and context within the conversation. The classified data is then recorded in a searchable historical database.

[0044] Implementing the reminder function

[0045] If a time-sensitive task is detected in the conversation, the server will suggest setting a reminder to the user. This reminder will be managed to ensure that the user receives timely notifications about the scheduled task.

[0046] Access to history

[0047] Users can access the dashboard from their devices and view their past conversation history. The history is organized by category, and users can easily search based on specific keywords or topics. This allows users to quickly access the information they need and support their current and future decision-making by referring to past conversations.

[0048] Specific example

[0049] For example, if a user asks their device, "Please recommend some slide designs for my presentation next week," this conversation is categorized as "business support." The server records this information in the conversation history database, and when the user asks about "slide designs" again later, it can refer to the relevant history and provide suggestions based on past information. In this way, users can efficiently utilize the AI ​​agent while maintaining the context of the conversation.

[0050] The following describes the processing flow.

[0051] Step 1:

[0052] The user uses their device to speak to the AI ​​agent and initiate a conversation. The user can input information via voice or text.

[0053] Step 2:

[0054] When the device receives voice input, it uses speech recognition technology to convert the speech into text, and then generates text data.

[0055] Step 3:

[0056] The terminal sends text data from the user to the server. This data includes metadata such as input timestamps and device information.

[0057] Step 4:

[0058] The server analyzes the received text data. It uses natural language processing algorithms to perform tokenization, part-of-speech tagging, and named entity recognition.

[0059] Step 5:

[0060] The server identifies the subject of the conversation data based on the analysis results and classifies it into predefined categories. This process is automated using machine learning models.

[0061] Step 6:

[0062] The server records the classified conversation data in a history database and organizes the data in a way that makes it easily searchable.

[0063] Step 7:

[0064] The server extracts time-sensitive tasks identified during the conversation and sends reminder suggestions for these tasks to the user.

[0065] Step 8:

[0066] The user reviews the reminder suggestions sent from the server and sets the reminder content and notification time as needed.

[0067] Step 9:

[0068] Based on the user's selection, the server completes the reminder settings and prepares notifications according to the set time.

[0069] Step 10:

[0070] Users can access the dashboard via their device to view and search organized conversation history. They can quickly find past information by specific categories or keywords.

[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] In recent years, there has been a growing demand for users to utilize generative AI models to efficiently extract necessary information from conversational data. However, conventional technologies have often been cumbersome in terms of data management and information retrieval. Furthermore, there has been a lack of automated systems for maintaining the context of conversations and setting appropriate reminders. To address these challenges, a method is needed that combines more advanced natural language processing and speech recognition to enable efficient data classification and access.

[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 analyzing user dialogue data using natural language processing means and extracting words and contexts, means for using speech recognition technology to convert speech into text information, and means for using a machine learning algorithm to determine classifications based on the analysis results. This enables users to efficiently analyze and organize dialogue data and utilize appropriate reminder functions.

[0076] "Natural language processing methods" refer to a series of techniques that use computer technology to analyze human language and extract meaning from text data.

[0077] A "user" is the entity that operates this system and receives the input and output dialogue data.

[0078] "Dialogue data" refers to communication data in voice or text format that users provide to the system.

[0079] A "word or phrase" refers to a word or short expression contained within a text, and is the basic unit of semantic analysis.

[0080] "Context" refers to the information surrounding a particular word or phrase, determining how it fits into a dialogue and thus its meaning.

[0081] "Means of extraction" refers to the technical process of extracting necessary information from dialogue data.

[0082] "Speech recognition technology" is a technology that processes human speech as a digital signal and converts it into text data.

[0083] "Character information" refers to string data expressed in a format that can be processed by a computer.

[0084] A "machine learning algorithm" is a method in which a computer learns patterns from large amounts of data and uses that knowledge to make predictions and classifications on new data.

[0085] A "reminder" is a function that notifies users based on specific times or conditions to remind them of appointments or tasks.

[0086] An "operation screen" is an interface that allows users to operate the system and view information.

[0087] This invention provides a system that utilizes an AI agent to efficiently manage user interaction data. This system mainly consists of a server, a terminal, and a user interface.

[0088] The device receives voice or text input from the user. In the case of voice input, the device uses speech recognition technology to convert the voice data into text. This conversion utilizes commonly used speech recognition services.

[0089] The converted text information and its metadata are sent from the terminal to the server. The server analyzes the received data using natural language processing (NLP) techniques. NLP includes tokenization, part-of-speech tagging, and named entity recognition. To perform these tasks, natural language processing libraries such as Python's NLTK and spaCy are utilized.

[0090] Based on the analysis results, the server classifies the dialogue data into predefined categories using a machine learning algorithm. This classification process evaluates the words and context within the data to determine the appropriate category. The classification results are recorded in a history database and organized into a format that can be searched by users at a later date.

[0091] Additionally, if the interaction data includes time-sensitive tasks, the server will suggest setting a reminder for the user. The reminder will be sent at an appropriate time using a standard calendar API.

[0092] Users can access their past conversation history through the terminal's interface and easily search for it as needed. The history is organized by category for user convenience, allowing for efficient information retrieval using specific keywords or topics.

[0093] For example, if a user says to their device, "Please prepare the materials for next week's meeting," this conversation is analyzed by the server and classified as a task with a deadline. Based on this information, the server can suggest that the user set a reminder. Examples of prompts include "next week's meeting," "prepare materials," and "I want to set a reminder."

[0094] In this way, this system allows users to efficiently manage conversational data and quickly access the information they need.

[0095] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0096] Step 1:

[0097] The user speaks to the AI ​​agent through the device. Input can be voice or text. In the case of voice input, the device captures the voice using its built-in microphone. This voice data is the initial input. The device converts this voice data into text information using speech recognition technology. This technology is performed by a speech recognition service. The output is text information.

[0098] Step 2:

[0099] The terminal sends text information and associated metadata (such as a time stamp and user ID) to the server. In this process, the terminal connects to the server using data communication technology and sends data using a secure protocol (e.g., HTTPS). The input is the text information and metadata from the terminal, and the output is the data received by the server.

[0100] Step 3:

[0101] The server applies natural language processing to the received text information. The input is text information. The server tokenizes this text data, tags it with parts of speech, and applies named entity recognition to identify proper nouns. Specifically, the server uses Python's natural language processing library. The output is the parsed text data.

[0102] Step 4:

[0103] The server executes a machine learning algorithm based on the analyzed text data to classify the dialogue data into predefined categories. The input is the server's analysis result. Using a machine learning library, the server evaluates the words and context within the data and determines the appropriate category using an algorithm. The output is the classified data.

[0104] Step 5:

[0105] The server records the classified data in a historical information base. The input is the classified data, and the output is the data stored in the historical information base. The server uses a database management system to perform the specific actions of storing the data in the appropriate format.

[0106] Step 6:

[0107] The server suggests setting a reminder to the user if the dialogue data includes time-sensitive tasks. The input is categorized data. The output is a reminder suggestion. The server uses the Calendar API to manage reminders and performs the specific action of sending configurable suggestions to the user.

[0108] Step 7:

[0109] The user searches and views categorized past conversation history on the terminal's operation screen. Input consists of search criteria entered through the operation screen. Output is the conversation history displayed as search results. The terminal performs specific actions, such as retrieving and displaying information from the database, via the user interface.

[0110] (Application Example 1)

[0111] 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."

[0112] In modern households, managing household chores and schedules has become increasingly complex, requiring users to efficiently organize vast amounts of information and smoothly complete daily tasks. However, doing this manually is time-consuming, laborious, and inefficient. Furthermore, there is a lack of readily available voice-activated tools for easily managing information and adjusting schedules within the home. Therefore, there is a need for a system that efficiently supports household chores and organizes information through voice interaction within the home.

[0113] 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.

[0114] In this invention, the server includes means for analyzing conversation information with the user using natural language processing means and extracting keywords and context; means for automatically classifying the conversation information into predefined classification groups based on the extracted information; means for recording the classified conversation information in the user's history information storage unit and organizing it for searchability; and means for receiving user instructions in the home environment through speech recognition and providing household support. This enables efficient home management by allowing the user to give instructions for household chores and daily tasks by voice, easily manage necessary information, and automatically set reminders.

[0115] "Natural language processing" refers to technologies that enable computers to understand and analyze human language, extracting keywords and context through the analysis of text data.

[0116] "Conversational information" refers to the content of communication exchanged between the user and the system via voice or text, and is data that is subject to natural language processing.

[0117] A "classification group" refers to a set of predefined categories used to categorize conversational information based on its content and context.

[0118] The "history information storage unit" refers to a database or data storage area that stores classified conversation information and organizes it so that it can be easily searched later.

[0119] "Speech recognition" is a technology that converts voice input into text data, a means of changing human speech into a format that computers can understand.

[0120] "Household support" refers to the act of providing assistance to efficiently carry out various tasks and management duties that occur within the home, and this invention achieves this using voice recognition.

[0121] "User instructions" refer to commands or inquiries made by the user to the system, which in this invention are given via voice or text.

[0122] A "notification" is a means of conveying information to a user that needs to be communicated at a certain point in time, and it functions as a reminder.

[0123] To implement this invention, a robot or smart device with home communication capabilities is required. When a user provides voice input, the device uses speech recognition technology to convert the speech into text data. Specific technologies that can be used include APIs such as Google® Cloud Speech-to-Text. The converted text data is sent to a server and analyzed using a natural language processing (NLP) library (e.g., spaCy). NLP tokenizes the text, performs part-of-speech tagging, and recognizes named entities.

[0124] The server classifies conversational information based on the analyzed data and automatically places it into predefined classification groups. This effectively organizes information based on similar contexts. The classified information is stored in a history information storage unit, making it easily accessible and searchable by the user as needed.

[0125] In a home environment, users can give instructions to the robot via voice recognition, for example, "Set a reminder to put the laundry in the dryer on Saturday." The invention then automatically sets the reminder and sends a notification to the user at the specified date and time. This enables task management based on voice commands as support for daily household chores, allowing users to efficiently continue their daily tasks.

[0126] As a concrete example, suppose a user speaks the following prompt to the terminal:

[0127] "Tell me your house cleaning list."

[0128] "Please set a reminder for taking out the trash next Tuesday."

[0129] "Make a shopping list for the next groceries."

[0130] The information obtained through these interactions is organized by category and used as foundational data to support the performance of necessary tasks. This embodiment enables efficient information management and household support through voice control within the home.

[0131] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0132] Step 1:

[0133] The device receives voice input. When a user speaks to a robot or smart device, the voice data is captured by the microphone in the device.

[0134] Step 2:

[0135] This process converts audio data into text data using a speech recognition API. The audio data is sent to a speech recognition API such as Google Cloud Speech-to-Text, and the output is obtained as text data.

[0136] Step 3:

[0137] The terminal sends the text data to the server. The terminal then sends the converted text data to the server via the internet and prepares it for analysis.

[0138] Step 4:

[0139] The server performs natural language processing. The server uses a natural language processing library such as spaCy to tokenize, tag parts of speech, and recognize named entities in the received text data. Keywords and context are extracted as part of the analysis results.

[0140] Step 5:

[0141] The server automatically classifies conversational information based on the analysis results. Based on the extracted keywords and context, it sorts the conversational information into predefined classification groups and outputs organized information.

[0142] Step 6:

[0143] The classified conversation information is stored in the history information storage unit. The server records the classification results in the history information storage unit to prepare for future searches and references.

[0144] Step 7:

[0145] If the task has a deadline, a reminder will be set. The server identifies the deadline from the conversation information and automatically sets a reminder for the date and time specified by the user.

[0146] Step 8:

[0147] Users can view conversation history and reminders through the dashboard. Through the interface, users can review past conversations and set reminders, and manage and manipulate information as needed.

[0148] 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.

[0149] This invention provides a system for analyzing conversations with users using an AI agent, recognizing not only the content of the conversation but also the user's emotions. This system combines natural language processing means and an emotion engine to analyze and record conversations while taking the user's emotional state into account.

[0150] Start interaction with the user

[0151] The user speaks to the AI ​​agent via voice or text through their device, initiating a conversation. In the case of voice input, the voice data is converted into text data using speech recognition technology.

[0152] Recognition and analysis of emotions

[0153] The device analyzes the user's emotions using an emotion engine along with voice and text data. This engine infers emotions from voice tone, speaking patterns, and text expressions.

[0154] Processing conversation data

[0155] The terminal sends the generated text data and sentiment data to the server. The server performs analysis on the received data using natural language processing. This involves tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context.

[0156] Category classification and emotional integration

[0157] The server classifies conversation data into predefined categories based on the analysis results. It also records recognized emotion data in association with the conversation history. This ensures that data reflecting the user's emotional state is stored in the history database.

[0158] User Feedback

[0159] Based on the recognized emotions, the server adjusts the content of the conversation and suggestions. For example, if the emotion data indicates that the user has a question, it will provide detailed information addressing that question.

[0160] Access to history

[0161] Users can access a dashboard on their device to view and search organized conversation history. Emotion-based filtering is also available, allowing them to view conversation history associated with specific emotional states.

[0162] Specific example

[0163] For example, if a user tells the AI ​​agent, "Yesterday's meeting was exhausting," the emotion engine recognizes the feeling of fatigue. This conversation is categorized as "work-related" and saved in the history database along with the emotion data. Later, when the user searches for "fatigue" as a keyword on the dashboard, a list of related past conversations is displayed, allowing them to see the connection to their emotions at the time. This feature enables users to analyze and utilize past conversations from a deeper perspective, including their emotional state.

[0164] The following describes the processing flow.

[0165] Step 1:

[0166] The user uses a device to initiate a conversation with the AI ​​agent. In the case of voice input, speech recognition is utilized, and the voice data is converted into text data.

[0167] Step 2:

[0168] The device sends voice and text data to an emotion engine to analyze the user's emotional state. This includes voice tone analysis and text keyword analysis.

[0169] Step 3:

[0170] The device sends text data and sentiment data to the server. This includes metadata such as conversation timestamps and user identification information.

[0171] Step 4:

[0172] The server uses natural language processing algorithms to analyze text data. Specifically, it performs tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context from the conversation.

[0173] Step 5:

[0174] Based on the analysis results, the server classifies the conversation data into predefined categories. Machine learning models are used to ensure appropriate classification based on the context and sentiment data of the conversation.

[0175] Step 6:

[0176] The server records classified conversation data and emotion data in the user's history database. This creates a history that takes emotional states into account.

[0177] Step 7:

[0178] The server generates feedback for the user based on the conversation history. Based on sentiment data, it prepares content to provide more detailed information and appropriate suggestions.

[0179] Step 8:

[0180] Users can access the dashboard through their device to view and search past conversation history along with emotional states. This allows them to quickly find conversations associated with specific emotions.

[0181] Step 9:

[0182] Users can input specific conditions or keywords to filter past conversations and perform quantitative or qualitative analysis. This feature allows users to gain insights into emotions and how they change.

[0183] (Example 2)

[0184] 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".

[0185] Conventional conversation analysis systems focus on analyzing the content of user interactions, but they fail to adequately provide information that takes into account the user's emotional state. Furthermore, searching and analyzing past conversation history based on emotions was difficult, making it challenging to provide feedback tailored to user needs. This highlighted the challenge of providing a personalized user experience.

[0186] 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.

[0187] In this invention, the server includes means for analyzing conversation information with the user using natural language processing means and extracting keywords and context; means for recognizing the user's emotional state using emotion analysis means and storing the recognized emotion in association with the conversation information; and means for providing a display function for the user to access classified conversation history and search for history related to a specific emotional state. This enables detailed information provision that takes the user's emotional state into consideration and advanced analysis of past conversations.

[0188] "Natural language processing methods" are technologies that analyze conversational information in the form of speech or text and extract important keywords and context.

[0189] "Emotional analysis methods" refer to technologies that recognize and analyze a user's emotional state based on their voice tone and text content.

[0190] "Conversational information" refers to audio or text data exchanged between the user and the system.

[0191] "Classification" is the process of organizing analyzed conversational information into predefined categories.

[0192] A "history medium" is a database or storage system that accumulates user conversation information and emotional states so that they can be referenced and analyzed later.

[0193] The "display function" is a function that allows users to access accumulated historical information and display that information based on specific conditions or filters.

[0194] This system analyzes conversations between users and AI agents, taking emotions into consideration when providing information. The system primarily consists of terminals and servers.

[0195] The terminal accepts voice or text input from the user, and in the case of voice input, it uses speech recognition software to convert the voice data into text data. A speech recognition API is generally considered the speech recognition technology used here. The user accesses the AI ​​agent through the terminal and initiates a conversation.

[0196] The converted text data is analyzed for the user's emotions using sentiment analysis tools. This analysis utilizes libraries and emotion recognition engines for natural language processing. This allows the user's emotional state to be inferred from their voice tone and text content.

[0197] The server receives text and sentiment data sent from the terminal and performs automatic analysis. Using natural language processing, the server tokenizes the text data, tags parts of speech, and recognizes named entities, extracting keywords and context from the analysis results. This allows the conversational information to be classified into predefined categories and stored in the user's history storage system.

[0198] This system has the ability to adjust user feedback according to perceived emotions. For example, if a user is showing signs of anxiety, the server provides specific information to reassure them. Users can access historical information through the device's dashboard, filter it based on specific emotional states, and investigate past conversations in detail.

[0199] For example, if a user tells the AI ​​agent, "Yesterday's meeting was exhausting," the emotion engine recognizes "fatigue." This conversation is categorized as work-related and stored in the history media along with the emotion data. Later, when the user searches for "fatigue" on the dashboard, the relevant conversation is displayed, allowing them to analyze their emotional state at the time.

[0200] Examples of prompts include, "What are your thoughts on yesterday's project meeting?" or "Tell me about any stress you've experienced at work recently." These prompts serve as a starting point for user interaction and improve the accuracy of the AI ​​system's sentiment analysis and information provision.

[0201] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0202] Step 1:

[0203] The user speaks to the AI ​​agent via the device in either voice or text format. The device receives voice or text data as input. In the case of voice input, the device uses speech recognition technology to convert the voice data into text data. A speech recognition API is utilized in this process. The output is text data.

[0204] Step 2:

[0205] The terminal sends the converted text data to the sentiment analysis system. It receives text data as input and uses a sentiment analysis engine to identify the user's emotions. This process infers emotions based on tone and expression. The output is analyzed data including emotion labels.

[0206] Step 3:

[0207] The terminal sends text data and sentiment labels to the server. The server receives this data as input and performs detailed analysis using natural language processing techniques. It performs processes such as tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context. The output is structured data containing the analysis results.

[0208] Step 4:

[0209] The server classifies conversation information into predefined categories based on the analysis results. This process uses extracted keywords to identify the conversation's theme. It accepts structured data as input and outputs data with category labels.

[0210] Step 5:

[0211] The server integrates recognized sentiment and category labels into the conversation history and stores them in the user's history medium. The input is sentiment and category-labeled data, and the output is an updated history database.

[0212] Step 6:

[0213] The server generates feedback for the user based on the recognized emotion and category. It references the current emotional state and historical data as input and outputs information and advice tailored to the user's situation.

[0214] Step 7:

[0215] Users can view categorized conversation history through the device's dashboard and filter information based on specific emotional states. The input is an emotional filter condition, and the output is the filtered conversation history. This allows users to analyze past conversations from an emotional perspective.

[0216] (Application Example 2)

[0217] 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".

[0218] In communicating with the elderly and individuals requiring care, it is essential to accurately understand their emotional state and respond appropriately based on that understanding. However, conventional technologies have made it difficult to accurately recognize the emotions of users during communication and respond in real time based on those emotions. Therefore, there is a need for technology that can effectively apply emotion recognition to care settings and provide appropriate support in accordance with the emotions of users.

[0219] 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.

[0220] In this invention, the server includes means for analyzing user interaction data using natural language processing means and extracting keywords and context; means for automatically classifying the interaction data into predefined categories based on the extracted information; and means for recognizing and recording the user's emotional state from speech and text using an emotion analysis engine. This makes it possible to accurately grasp the user's emotional state and quickly provide appropriate responses accordingly.

[0221] "Natural language processing" refers to technologies that analyze speech and text data from users to understand language structure.

[0222] An "emotion analysis engine" is a technology that infers and analyzes a user's emotional state from their voice characteristics and textual expression.

[0223] "Users" refers to the individuals who use the system, particularly the elderly and those requiring care.

[0224] "Dialogue data" refers to data that constitutes the content of conversations exchanged between users and the system.

[0225] A "category" refers to a predefined field or theme used when classifying dialogue data based on extracted information.

[0226] A "history information system" is a database that records analyzed and classified dialogue data and organizes a user's past conversation history.

[0227] "Display interface" refers to the screen or dashboard that allows users to access and search historical information.

[0228] The system implementing this invention is configured around a user terminal and a server. Users interact through a terminal such as a smartphone or smart glasses. This terminal analyzes speech or text data using natural language processing means and extracts keywords and context. In the case of speech input, the terminal converts speech data into text data using speech recognition technology. Google Cloud Speech-to-Text is used in this process.

[0229] The server receives data transmitted from the terminal and uses an emotion analysis engine to recognize and record the user's emotional state from the speech and text. IBM Watson® Tone Analyzer is used for emotion analysis, and spaCy is also used as part of natural language processing technology. This automatically classifies the conversation data into predefined categories.

[0230] Based on the recognized emotions, the server makes appropriate response suggestions. The feedback mechanism for this purpose generates suggestions and reminders tailored to the user's emotions and notifies care staff and the user themselves.

[0231] For example, if a user says, "I'm feeling a little lonely today," the system will detect "loneliness" as an emotion and provide staff with a suggestion such as, "The user is feeling lonely, so we recommend you talk to them." An example of a prompt for the generating AI model would be, "Perform sentiment analysis based on the following conversation data and generate appropriate care suggestions. Conversation: 'I'm feeling a little lonely today.'"

[0232] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0233] Step 1:

[0234] The user speaks to the device via voice or text, which initiates the interaction. If voice input is provided, the device uses Google Cloud Speech-to-Text to convert the voice data into text data. The input is the user's voice data, and the output is text data. Specifically, the device's microphone captures the voice, which is then processed by the cloud service.

[0235] Step 2:

[0236] The terminal performs natural language processing based on the acquired text data to extract keywords and context. This process uses spaCy for tokenization and part-of-speech tagging. The input is the text data obtained in step 1, and the output is the analyzed keywords and context information. Specifically, the analyzed data is stored in memory and sent to subsequent processing.

[0237] Step 3:

[0238] The terminal uses an emotion analysis engine to infer emotional states from text and voice characteristics. Input is user text data and voice characteristic information, and output is user emotional state data. Specifically, IBM Watson Tone Analyzer analyzes the emotion of the text and saves the result as a status.

[0239] Step 4:

[0240] The server receives keywords, contextual information, and sentiment data sent from the terminal and automatically classifies the dialogue data into predefined categories. The input is the analyzed keywords and sentiment data, and the output is the classification result. The specific operation is the process of executing the classification algorithm and saving the data to the database.

[0241] Step 5:

[0242] The server generates appropriate response suggestions based on classified data and sentiment data, and notifies the user or care staff. The input is the classification results and sentiment state from step 4, and the output is the response suggestion message. The specific operation is to execute the logic for generating response suggestions and send the message to the relevant terminal via the notification function.

[0243] 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.

[0244] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0245] 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.

[0246] [Second Embodiment]

[0247] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0248] 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.

[0249] 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).

[0250] 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.

[0251] 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.

[0252] 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).

[0253] 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.

[0254] 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.

[0255] 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.

[0256] 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.

[0257] 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.

[0258] 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".

[0259] As an embodiment of the present invention, a system that utilizes an AI agent to efficiently classify and manage conversation data with users will be described.

[0260] Start interaction with the user

[0261] The interaction begins when the user speaks to the AI ​​agent through the device. Input can be provided as either text or voice. In the case of voice input, the device uses speech recognition technology to convert the voice data into text data.

[0262] Processing conversation data

[0263] The terminal sends the converted text data and its metadata to the server. The server receives this data and applies natural language processing (NLP) to it. In the analysis, the text is tokenized and broken down into individual words and phrases, each tagged with its part of speech. Then, named entity recognition is used to extract specific proper nouns.

[0264] Data classification and organization

[0265] Based on the analysis results, the server classifies the conversation data into predefined categories. This classification utilizes automated machine learning algorithms to determine the appropriate category based on the similarity of keywords and context within the conversation. The classified data is then recorded in a searchable historical database.

[0266] Implementing the reminder function

[0267] If a time-sensitive task is detected in the conversation, the server will suggest setting a reminder to the user. This reminder will be managed to ensure that the user receives timely notifications about the scheduled task.

[0268] Access to history

[0269] Users can access the dashboard from their devices and view their past conversation history. The history is organized by category, and users can easily search based on specific keywords or topics. This allows users to quickly access the information they need and support their current and future decision-making by referring to past conversations.

[0270] Specific example

[0271] For example, if a user asks their device, "Please recommend some slide designs for my presentation next week," this conversation is categorized as "business support." The server records this information in the conversation history database, and when the user asks about "slide designs" again later, it can refer to the relevant history and provide suggestions based on past information. In this way, users can efficiently utilize the AI ​​agent while maintaining the context of the conversation.

[0272] The following describes the processing flow.

[0273] Step 1:

[0274] The user uses their device to speak to the AI ​​agent and initiate a conversation. The user can input information via voice or text.

[0275] Step 2:

[0276] When the device receives voice input, it uses speech recognition technology to convert the speech into text, and then generates text data.

[0277] Step 3:

[0278] The terminal sends text data from the user to the server. This data includes metadata such as input timestamps and device information.

[0279] Step 4:

[0280] The server analyzes the received text data. It uses natural language processing algorithms to perform tokenization, part-of-speech tagging, and named entity recognition.

[0281] Step 5:

[0282] The server identifies the theme of the conversation data based on the analysis results and classifies it into pre-defined categories. This process is automated using a machine learning model.

[0283] Step 6:

[0284] The server records the classified conversation data in the history database and organizes the data in an easily searchable form.

[0285] Step 7:

[0286] The server extracts the time-limited tasks identified during the conversation. A reminder proposal for this task is sent to the user.

[0287] Step 8:

[0288] The user checks the reminder proposal sent from the server and sets the content and notification time of the reminder as needed.

[0289] Step 9:

[0290] Based on the user's selection, the server completes the reminder settings and prepares the notification according to the set time.

[0291] Step 10:

[0292] The user can access the dashboard via the terminal and view / search the organized conversation history. The user can quickly find past information using specific categories or keywords.

[0293] (Example 1)

[0294] Next, Example 1 will be described. In the following description, the data processing device 12 is referred to as the "server", and the smart glasses 214 are referred to as the "terminal".

[0295] In recent years, there has been a growing demand for users to utilize generative AI models to efficiently extract necessary information from conversational data. However, conventional technologies have often been cumbersome in terms of data management and information retrieval. Furthermore, there has been a lack of automated systems for maintaining the context of conversations and setting appropriate reminders. To address these challenges, a method is needed that combines more advanced natural language processing and speech recognition to enable efficient data classification and access.

[0296] 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.

[0297] In this invention, the server includes means for analyzing user dialogue data using natural language processing means and extracting words and contexts, means for using speech recognition technology to convert speech into text information, and means for using a machine learning algorithm to determine classifications based on the analysis results. This enables users to efficiently analyze and organize dialogue data and utilize appropriate reminder functions.

[0298] "Natural language processing methods" refer to a series of techniques that use computer technology to analyze human language and extract meaning from text data.

[0299] A "user" is the entity that operates this system and receives the input and output dialogue data.

[0300] "Dialogue data" refers to communication data in voice or text format that users provide to the system.

[0301] A "word or phrase" refers to a word or short expression contained within a text, and is the basic unit of semantic analysis.

[0302] "Context" refers to the information surrounding a particular word or phrase, determining how it fits into a dialogue and thus its meaning.

[0303] The "extraction means" is a technical process for extracting necessary information from the dialogue data.

[0304] The "speech recognition technology" is a technology that processes human speech as a digital signal and converts it into character data.

[0305] The "character information" is string data represented in a form that can be processed on a computer.

[0306] The "machine learning algorithm" is a method by which a computer learns patterns from a large amount of data and makes predictions or classifications for new data.

[0307] The "reminder" is a function for notifying the user based on specific times or conditions and reminding them of schedules or tasks.

[0308] The "operation screen" is an interface for the user to operate the system and view information.

[0309] The present invention provides a system that utilizes an AI agent to efficiently manage dialogue data with the user. This system is mainly composed of a server, a terminal, and the user interface.

[0310] The terminal receives voice or text input from the user. In the case of voice input, the terminal uses speech recognition technology to convert the voice data into character information. Generally widely used speech recognition services are utilized for this conversion.

[0311] The converted character information and its metadata are transmitted from the terminal to the server. The server analyzes the received data using natural language processing means. Natural language processing includes tokenization, part-of-speech tagging, and named entity recognition. Natural language processing libraries such as Python's NLTK and spaCy are utilized to execute these.

[0312] Based on the analysis results, the server classifies the dialogue data into predefined categories using a machine learning algorithm. This classification process evaluates the words and context within the data to determine the appropriate category. The classification results are recorded in a history database and organized into a format that can be searched by users at a later date.

[0313] Additionally, if the interaction data includes time-sensitive tasks, the server will suggest setting a reminder for the user. The reminder will be sent at an appropriate time using a standard calendar API.

[0314] Users can access their past conversation history through the terminal's interface and easily search for it as needed. The history is organized by category for user convenience, allowing for efficient information retrieval using specific keywords or topics.

[0315] For example, if a user says to their device, "Please prepare the materials for next week's meeting," this conversation is analyzed by the server and classified as a task with a deadline. Based on this information, the server can suggest that the user set a reminder. Examples of prompts include "next week's meeting," "prepare materials," and "I want to set a reminder."

[0316] In this way, this system allows users to efficiently manage conversational data and quickly access the information they need.

[0317] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0318] Step 1:

[0319] The user speaks to the AI ​​agent through the device. Input can be voice or text. In the case of voice input, the device captures the voice using its built-in microphone. This voice data is the initial input. The device converts this voice data into text information using speech recognition technology. This technology is performed by a speech recognition service. The output is text information.

[0320] Step 2:

[0321] The terminal sends text information and associated metadata (such as a time stamp and user ID) to the server. In this process, the terminal connects to the server using data communication technology and sends data using a secure protocol (e.g., HTTPS). The input is the text information and metadata from the terminal, and the output is the data received by the server.

[0322] Step 3:

[0323] The server applies natural language processing to the received text information. The input is text information. The server tokenizes this text data, tags it with parts of speech, and applies named entity recognition to identify proper nouns. Specifically, the server uses Python's natural language processing library. The output is the parsed text data.

[0324] Step 4:

[0325] The server executes a machine learning algorithm based on the analyzed text data to classify the dialogue data into predefined categories. The input is the server's analysis result. Using a machine learning library, the server evaluates the words and context within the data and determines the appropriate category using an algorithm. The output is the classified data.

[0326] Step 5:

[0327] The server records the classified data in a historical information base. The input is the classified data, and the output is the data stored in the historical information base. The server uses a database management system to perform the specific actions of storing the data in the appropriate format.

[0328] Step 6:

[0329] The server suggests setting a reminder to the user if the dialogue data includes time-sensitive tasks. The input is categorized data. The output is a reminder suggestion. The server uses the Calendar API to manage reminders and performs the specific action of sending configurable suggestions to the user.

[0330] Step 7:

[0331] The user searches and views categorized past conversation history on the terminal's operation screen. Input consists of search criteria entered through the operation screen. Output is the conversation history displayed as search results. The terminal performs specific actions, such as retrieving and displaying information from the database, via the user interface.

[0332] (Application Example 1)

[0333] 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."

[0334] In modern households, managing household chores and schedules has become increasingly complex, requiring users to efficiently organize vast amounts of information and smoothly complete daily tasks. However, doing this manually is time-consuming, laborious, and inefficient. Furthermore, there is a lack of readily available voice-activated tools for easily managing information and adjusting schedules within the home. Therefore, there is a need for a system that efficiently supports household chores and organizes information through voice interaction within the home.

[0335] 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.

[0336] In this invention, the server includes means for analyzing conversation information with the user using natural language processing means and extracting keywords and context; means for automatically classifying the conversation information into predefined classification groups based on the extracted information; means for recording the classified conversation information in the user's history information storage unit and organizing it for searchability; and means for receiving user instructions in the home environment through speech recognition and providing household support. This enables efficient home management by allowing the user to give instructions for household chores and daily tasks by voice, easily manage necessary information, and automatically set reminders.

[0337] "Natural language processing" refers to technologies that enable computers to understand and analyze human language, extracting keywords and context through the analysis of text data.

[0338] "Conversational information" refers to the content of communication exchanged between the user and the system via voice or text, and is data that is subject to natural language processing.

[0339] A "classification group" refers to a set of predefined categories used to categorize conversational information based on its content and context.

[0340] The "history information storage unit" refers to a database or data storage area that stores classified conversation information and organizes it so that it can be easily searched later.

[0341] "Speech recognition" is a technology that converts voice input into text data, a means of changing human speech into a format that computers can understand.

[0342] "Household support" refers to the act of providing assistance to efficiently carry out various tasks and management duties that occur within the home, and this invention achieves this using voice recognition.

[0343] "User instructions" refer to commands or inquiries made by the user to the system, which in this invention are given via voice or text.

[0344] A "notification" is a means of conveying information to a user that needs to be communicated at a certain point in time, and it functions as a reminder.

[0345] To implement this invention, a robot or smart device with home communication capabilities is required. When a user provides voice input, the device uses speech recognition technology to convert the speech into text data. Specific technologies that can be used include APIs such as Google Cloud Speech-to-Text. The converted text data is sent to a server and analyzed using a natural language processing (NLP) library (e.g., spaCy). NLP tokenizes the text, performs part-of-speech tagging, and recognizes named entities.

[0346] The server classifies conversational information based on the analyzed data and automatically places it into predefined classification groups. This effectively organizes information based on similar contexts. The classified information is stored in a history information storage unit, making it easily accessible and searchable by the user as needed.

[0347] In a home environment, users can give instructions to the robot via voice recognition, for example, "Set a reminder to put the laundry in the dryer on Saturday." The invention then automatically sets the reminder and sends a notification to the user at the specified date and time. This enables task management based on voice commands as support for daily household chores, allowing users to efficiently continue their daily tasks.

[0348] As a concrete example, suppose a user speaks the following prompt to the terminal:

[0349] "Tell me your house cleaning list."

[0350] "Please set a reminder for taking out the trash next Tuesday."

[0351] "Make a shopping list for the next groceries."

[0352] The information obtained through these interactions is organized by category and used as foundational data to support the performance of necessary tasks. This embodiment enables efficient information management and household support through voice control within the home.

[0353] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0354] Step 1:

[0355] The device receives voice input. When a user speaks to a robot or smart device, the voice data is captured by the microphone in the device.

[0356] Step 2:

[0357] This process converts audio data into text data using a speech recognition API. The audio data is sent to a speech recognition API such as Google Cloud Speech-to-Text, and the output is obtained as text data.

[0358] Step 3:

[0359] The terminal sends the text data to the server. The terminal then sends the converted text data to the server via the internet and prepares it for analysis.

[0360] Step 4:

[0361] The server performs natural language processing. The server uses a natural language processing library such as spaCy to tokenize, tag parts of speech, and recognize named entities in the received text data. Keywords and context are extracted as part of the analysis results.

[0362] Step 5:

[0363] The server automatically classifies conversational information based on the analysis results. Based on the extracted keywords and context, it sorts the conversational information into predefined classification groups and outputs organized information.

[0364] Step 6:

[0365] The classified conversation information is stored in the history information storage unit. The server records the classification results in the history information storage unit to prepare for future searches and references.

[0366] Step 7:

[0367] If the task has a deadline, a reminder will be set. The server identifies the deadline from the conversation information and automatically sets a reminder for the date and time specified by the user.

[0368] Step 8:

[0369] Users can view conversation history and reminders through the dashboard. Through the interface, users can review past conversations and set reminders, and manage and manipulate information as needed.

[0370] 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.

[0371] This invention provides a system for analyzing conversations with users using an AI agent, recognizing not only the content of the conversation but also the user's emotions. This system combines natural language processing means and an emotion engine to analyze and record conversations while taking the user's emotional state into account.

[0372] Start interaction with the user

[0373] The user speaks to the AI ​​agent via voice or text through their device, initiating a conversation. In the case of voice input, the voice data is converted into text data using speech recognition technology.

[0374] Recognition and analysis of emotions

[0375] The device analyzes the user's emotions using an emotion engine along with voice and text data. This engine infers emotions from voice tone, speaking patterns, and text expressions.

[0376] Processing conversation data

[0377] The terminal sends the generated text data and sentiment data to the server. The server performs analysis on the received data using natural language processing. This involves tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context.

[0378] Category classification and emotional integration

[0379] The server classifies conversation data into predefined categories based on the analysis results. It also records recognized emotion data in association with the conversation history. This ensures that data reflecting the user's emotional state is stored in the history database.

[0380] User Feedback

[0381] Based on the recognized emotions, the server adjusts the content of the conversation and suggestions. For example, if the emotion data indicates that the user has a question, it will provide detailed information addressing that question.

[0382] Access to history

[0383] Users can access a dashboard on their device to view and search organized conversation history. Emotion-based filtering is also available, allowing them to view conversation history associated with specific emotional states.

[0384] Specific example

[0385] For example, if a user tells the AI ​​agent, "Yesterday's meeting was exhausting," the emotion engine recognizes the feeling of fatigue. This conversation is categorized as "work-related" and saved in the history database along with the emotion data. Later, when the user searches for "fatigue" as a keyword on the dashboard, a list of related past conversations is displayed, allowing them to see the connection to their emotions at the time. This feature enables users to analyze and utilize past conversations from a deeper perspective, including their emotional state.

[0386] The following describes the processing flow.

[0387] Step 1:

[0388] The user uses a device to initiate a conversation with the AI ​​agent. In the case of voice input, speech recognition is utilized, and the voice data is converted into text data.

[0389] Step 2:

[0390] The device sends voice and text data to an emotion engine to analyze the user's emotional state. This includes voice tone analysis and text keyword analysis.

[0391] Step 3:

[0392] The device sends text data and sentiment data to the server. This includes metadata such as conversation timestamps and user identification information.

[0393] Step 4:

[0394] The server uses natural language processing algorithms to analyze text data. Specifically, it performs tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context from the conversation.

[0395] Step 5:

[0396] Based on the analysis results, the server classifies the conversation data into predefined categories. Machine learning models are used to ensure appropriate classification based on the context and sentiment data of the conversation.

[0397] Step 6:

[0398] The server records classified conversation data and emotion data in the user's history database. This creates a history that takes emotional states into account.

[0399] Step 7:

[0400] The server generates feedback for the user based on the conversation history. Based on sentiment data, it prepares content to provide more detailed information and appropriate suggestions.

[0401] Step 8:

[0402] Users can access the dashboard through their device to view and search past conversation history along with emotional states. This allows them to quickly find conversations associated with specific emotions.

[0403] Step 9:

[0404] Users can input specific conditions or keywords to filter past conversations and perform quantitative or qualitative analysis. This feature allows users to gain insights into emotions and how they change.

[0405] (Example 2)

[0406] 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".

[0407] Conventional conversation analysis systems focus on analyzing the content of user interactions, but they fail to adequately provide information that takes into account the user's emotional state. Furthermore, searching and analyzing past conversation history based on emotions was difficult, making it challenging to provide feedback tailored to user needs. This highlighted the challenge of providing a personalized user experience.

[0408] 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.

[0409] In this invention, the server includes means for analyzing conversation information with the user using natural language processing means and extracting keywords and context; means for recognizing the user's emotional state using emotion analysis means and storing the recognized emotion in association with the conversation information; and means for providing a display function for the user to access classified conversation history and search for history related to a specific emotional state. This enables detailed information provision that takes the user's emotional state into consideration and advanced analysis of past conversations.

[0410] "Natural language processing methods" are technologies that analyze conversational information in the form of speech or text and extract important keywords and context.

[0411] "Emotional analysis methods" refer to technologies that recognize and analyze a user's emotional state based on their voice tone and text content.

[0412] "Conversational information" refers to audio or text data exchanged between the user and the system.

[0413] "Classification" is the process of organizing analyzed conversational information into predefined categories.

[0414] A "history medium" is a database or storage system that accumulates user conversation information and emotional states so that they can be referenced and analyzed later.

[0415] The "display function" is a function that allows users to access accumulated historical information and display that information based on specific conditions or filters.

[0416] This system analyzes conversations between users and AI agents, taking emotions into consideration when providing information. The system primarily consists of terminals and servers.

[0417] The terminal accepts voice or text input from the user, and in the case of voice input, it uses speech recognition software to convert the voice data into text data. A speech recognition API is generally considered the speech recognition technology used here. The user accesses the AI ​​agent through the terminal and initiates a conversation.

[0418] The converted text data is analyzed for the user's emotions using sentiment analysis tools. This analysis utilizes libraries and emotion recognition engines for natural language processing. This allows the user's emotional state to be inferred from their voice tone and text content.

[0419] The server receives text and sentiment data sent from the terminal and performs automatic analysis. Using natural language processing, the server tokenizes the text data, tags parts of speech, and recognizes named entities, extracting keywords and context from the analysis results. This allows the conversational information to be classified into predefined categories and stored in the user's history storage system.

[0420] This system has the ability to adjust user feedback according to perceived emotions. For example, if a user is showing signs of anxiety, the server provides specific information to reassure them. Users can access historical information through the device's dashboard, filter it based on specific emotional states, and investigate past conversations in detail.

[0421] For example, if a user tells the AI ​​agent, "Yesterday's meeting was exhausting," the emotion engine recognizes "fatigue." This conversation is categorized as work-related and stored in the history media along with the emotion data. Later, when the user searches for "fatigue" on the dashboard, the relevant conversation is displayed, allowing them to analyze their emotional state at the time.

[0422] Examples of prompts include, "What are your thoughts on yesterday's project meeting?" or "Tell me about any stress you've experienced at work recently." These prompts serve as a starting point for user interaction and improve the accuracy of the AI ​​system's sentiment analysis and information provision.

[0423] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0424] Step 1:

[0425] The user speaks to the AI ​​agent via the device in either voice or text format. The device receives voice or text data as input. In the case of voice input, the device uses speech recognition technology to convert the voice data into text data. A speech recognition API is utilized in this process. The output is text data.

[0426] Step 2:

[0427] The terminal sends the converted text data to the sentiment analysis system. It receives text data as input and uses a sentiment analysis engine to identify the user's emotions. This process infers emotions based on tone and expression. The output is analyzed data including emotion labels.

[0428] Step 3:

[0429] The terminal sends text data and sentiment labels to the server. The server receives this data as input and performs detailed analysis using natural language processing techniques. It performs processes such as tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context. The output is structured data containing the analysis results.

[0430] Step 4:

[0431] The server classifies conversation information into predefined categories based on the analysis results. This process uses extracted keywords to identify the conversation's theme. It accepts structured data as input and outputs data with category labels.

[0432] Step 5:

[0433] The server integrates recognized sentiment and category labels into the conversation history and stores them in the user's history medium. The input is sentiment and category-labeled data, and the output is an updated history database.

[0434] Step 6:

[0435] The server generates feedback for the user based on the recognized emotion and category. It references the current emotional state and historical data as input and outputs information and advice tailored to the user's situation.

[0436] Step 7:

[0437] Users can view categorized conversation history through the device's dashboard and filter information based on specific emotional states. The input is an emotional filter condition, and the output is the filtered conversation history. This allows users to analyze past conversations from an emotional perspective.

[0438] (Application Example 2)

[0439] 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."

[0440] In communicating with the elderly and individuals requiring care, it is essential to accurately understand their emotional state and respond appropriately based on that understanding. However, conventional technologies have made it difficult to accurately recognize the emotions of users during communication and respond in real time based on those emotions. Therefore, there is a need for technology that can effectively apply emotion recognition to care settings and provide appropriate support in accordance with the emotions of users.

[0441] 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.

[0442] In this invention, the server includes means for analyzing user interaction data using natural language processing means and extracting keywords and context; means for automatically classifying the interaction data into predefined categories based on the extracted information; and means for recognizing and recording the user's emotional state from speech and text using an emotion analysis engine. This makes it possible to accurately grasp the user's emotional state and quickly provide appropriate responses accordingly.

[0443] "Natural language processing" refers to technologies that analyze speech and text data from users to understand language structure.

[0444] An "emotion analysis engine" is a technology that infers and analyzes a user's emotional state from their voice characteristics and textual expression.

[0445] "Users" refers to the individuals who use the system, particularly the elderly and those requiring care.

[0446] "Dialogue data" refers to data that constitutes the content of conversations exchanged between users and the system.

[0447] A "category" refers to a predefined field or theme used when classifying dialogue data based on extracted information.

[0448] A "history information system" is a database that records analyzed and classified dialogue data and organizes a user's past conversation history.

[0449] "Display interface" refers to the screen or dashboard that allows users to access and search historical information.

[0450] The system implementing this invention is configured around a user terminal and a server. Users interact through a terminal such as a smartphone or smart glasses. This terminal analyzes speech or text data using natural language processing means and extracts keywords and context. In the case of speech input, the terminal converts speech data into text data using speech recognition technology. Google Cloud Speech-to-Text is used in this process.

[0451] The server receives data transmitted from the terminal and uses an emotion analysis engine to recognize and record the user's emotional state from the speech and text. IBM Watson Tone Analyzer is used for emotion analysis, and spaCy is also used as part of natural language processing technology. This automatically classifies the conversation data into predefined categories.

[0452] Based on the recognized emotions, the server makes appropriate response suggestions. The feedback mechanism for this purpose generates suggestions and reminders tailored to the user's emotions and notifies care staff and the user themselves.

[0453] For example, if a user says, "I'm feeling a little lonely today," the system will detect "loneliness" as an emotion and provide staff with a suggestion such as, "The user is feeling lonely, so we recommend you talk to them." An example of a prompt for the generating AI model would be, "Perform sentiment analysis based on the following conversation data and generate appropriate care suggestions. Conversation: 'I'm feeling a little lonely today.'"

[0454] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0455] Step 1:

[0456] The user speaks to the device via voice or text, which initiates the interaction. If voice input is provided, the device uses Google Cloud Speech-to-Text to convert the voice data into text data. The input is the user's voice data, and the output is text data. Specifically, the device's microphone captures the voice, which is then processed by the cloud service.

[0457] Step 2:

[0458] The terminal performs natural language processing based on the acquired text data to extract keywords and context. This process uses spaCy for tokenization and part-of-speech tagging. The input is the text data obtained in step 1, and the output is the analyzed keywords and context information. Specifically, the analyzed data is stored in memory and sent to subsequent processing.

[0459] Step 3:

[0460] The terminal uses an emotion analysis engine to infer emotional states from text and voice characteristics. Input is user text data and voice characteristic information, and output is user emotional state data. Specifically, IBM Watson Tone Analyzer analyzes the emotion of the text and saves the result as a status.

[0461] Step 4:

[0462] The server receives keywords, contextual information, and sentiment data sent from the terminal and automatically classifies the dialogue data into predefined categories. The input is the analyzed keywords and sentiment data, and the output is the classification result. The specific operation is the process of executing the classification algorithm and saving the data to the database.

[0463] Step 5:

[0464] The server generates appropriate response suggestions based on classified data and sentiment data, and notifies the user or care staff. The input is the classification results and sentiment state from step 4, and the output is the response suggestion message. The specific operation is to execute the logic for generating response suggestions and send the message to the relevant terminal via the notification function.

[0465] 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.

[0466] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0467] 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.

[0468] [Third Embodiment]

[0469] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0470] 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.

[0471] 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).

[0472] 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.

[0473] 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.

[0474] 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).

[0475] 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.

[0476] 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.

[0477] 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.

[0478] 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.

[0479] 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.

[0480] 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".

[0481] As an embodiment of the present invention, a system that utilizes an AI agent to efficiently classify and manage conversation data with users will be described.

[0482] Start interaction with the user

[0483] The interaction begins when the user speaks to the AI ​​agent through the device. Input can be provided as either text or voice. In the case of voice input, the device uses speech recognition technology to convert the voice data into text data.

[0484] Processing conversation data

[0485] The terminal sends the converted text data and its metadata to the server. The server receives this data and applies natural language processing (NLP) to it. In the analysis, the text is tokenized and broken down into individual words and phrases, each tagged with its part of speech. Then, named entity recognition is used to extract specific proper nouns.

[0486] Data classification and organization

[0487] Based on the analysis results, the server classifies the conversation data into predefined categories. This classification utilizes automated machine learning algorithms to determine the appropriate category based on the similarity of keywords and context within the conversation. The classified data is then recorded in a searchable historical database.

[0488] Implementing the reminder function

[0489] If a time-sensitive task is detected in the conversation, the server will suggest setting a reminder to the user. This reminder will be managed to ensure that the user receives timely notifications about the scheduled task.

[0490] Access to history

[0491] Users can access the dashboard from their devices and view their past conversation history. The history is organized by category, and users can easily search based on specific keywords or topics. This allows users to quickly access the information they need and support their current and future decision-making by referring to past conversations.

[0492] Specific example

[0493] For example, if a user asks their device, "Please recommend some slide designs for my presentation next week," this conversation is categorized as "business support." The server records this information in the conversation history database, and when the user asks about "slide designs" again later, it can refer to the relevant history and provide suggestions based on past information. In this way, users can efficiently utilize the AI ​​agent while maintaining the context of the conversation.

[0494] The following describes the processing flow.

[0495] Step 1:

[0496] The user uses their device to speak to the AI ​​agent and initiate a conversation. The user can input information via voice or text.

[0497] Step 2:

[0498] When the device receives voice input, it uses speech recognition technology to convert the speech into text, and then generates text data.

[0499] Step 3:

[0500] The terminal sends text data from the user to the server. This data includes metadata such as input timestamps and device information.

[0501] Step 4:

[0502] The server analyzes the received text data. It uses natural language processing algorithms to perform tokenization, part-of-speech tagging, and named entity recognition.

[0503] Step 5:

[0504] The server identifies the subject of the conversation data based on the analysis results and classifies it into predefined categories. This process is automated using machine learning models.

[0505] Step 6:

[0506] The server records the classified conversation data in a history database and organizes the data in a way that makes it easily searchable.

[0507] Step 7:

[0508] The server extracts time-sensitive tasks identified during the conversation and sends reminder suggestions for these tasks to the user.

[0509] Step 8:

[0510] The user reviews the reminder suggestions sent from the server and sets the reminder content and notification time as needed.

[0511] Step 9:

[0512] Based on the user's selection, the server completes the reminder settings and prepares notifications according to the set time.

[0513] Step 10:

[0514] Users can access the dashboard via their device to view and search organized conversation history. They can quickly find past information by specific categories or keywords.

[0515] (Example 1)

[0516] 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."

[0517] In recent years, there has been a growing demand for users to utilize generative AI models to efficiently extract necessary information from conversational data. However, conventional technologies have often been cumbersome in terms of data management and information retrieval. Furthermore, there has been a lack of automated systems for maintaining the context of conversations and setting appropriate reminders. To address these challenges, a method is needed that combines more advanced natural language processing and speech recognition to enable efficient data classification and access.

[0518] 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.

[0519] In this invention, the server includes means for analyzing user dialogue data using natural language processing means and extracting words and contexts, means for using speech recognition technology to convert speech into text information, and means for using a machine learning algorithm to determine classifications based on the analysis results. This enables users to efficiently analyze and organize dialogue data and utilize appropriate reminder functions.

[0520] "Natural language processing methods" refer to a series of techniques that use computer technology to analyze human language and extract meaning from text data.

[0521] A "user" is the entity that operates this system and receives the input and output dialogue data.

[0522] "Dialogue data" refers to communication data in voice or text format that users provide to the system.

[0523] A "word or phrase" refers to a word or short expression contained within a text, and is the basic unit of semantic analysis.

[0524] "Context" refers to the information surrounding a particular word or phrase, determining how it fits into a dialogue and thus its meaning.

[0525] "Means of extraction" refers to the technical process of extracting necessary information from dialogue data.

[0526] "Speech recognition technology" is a technology that processes human speech as a digital signal and converts it into text data.

[0527] "Character information" refers to string data expressed in a format that can be processed by a computer.

[0528] A "machine learning algorithm" is a method in which a computer learns patterns from large amounts of data and uses that knowledge to make predictions and classifications on new data.

[0529] A "reminder" is a function that notifies users based on specific times or conditions to remind them of appointments or tasks.

[0530] An "operation screen" is an interface that allows users to operate the system and view information.

[0531] This invention provides a system that utilizes an AI agent to efficiently manage user interaction data. This system mainly consists of a server, a terminal, and a user interface.

[0532] The device receives voice or text input from the user. In the case of voice input, the device uses speech recognition technology to convert the voice data into text. This conversion utilizes commonly used speech recognition services.

[0533] The converted text information and its metadata are sent from the terminal to the server. The server analyzes the received data using natural language processing (NLP) techniques. NLP includes tokenization, part-of-speech tagging, and named entity recognition. To perform these tasks, natural language processing libraries such as Python's NLTK and spaCy are utilized.

[0534] Based on the analysis results, the server classifies the dialogue data into predefined categories using a machine learning algorithm. This classification process evaluates the words and context within the data to determine the appropriate category. The classification results are recorded in a history database and organized into a format that can be searched by users at a later date.

[0535] Additionally, if the interaction data includes time-sensitive tasks, the server will suggest setting a reminder for the user. The reminder will be sent at an appropriate time using a standard calendar API.

[0536] Users can access their past conversation history through the terminal's interface and easily search for it as needed. The history is organized by category for user convenience, allowing for efficient information retrieval using specific keywords or topics.

[0537] For example, if a user says to their device, "Please prepare the materials for next week's meeting," this conversation is analyzed by the server and classified as a task with a deadline. Based on this information, the server can suggest that the user set a reminder. Examples of prompts include "next week's meeting," "prepare materials," and "I want to set a reminder."

[0538] In this way, this system allows users to efficiently manage conversational data and quickly access the information they need.

[0539] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0540] Step 1:

[0541] The user speaks to the AI ​​agent through the device. Input can be voice or text. In the case of voice input, the device captures the voice using its built-in microphone. This voice data is the initial input. The device converts this voice data into text information using speech recognition technology. This technology is performed by a speech recognition service. The output is text information.

[0542] Step 2:

[0543] The terminal sends text information and associated metadata (such as a time stamp and user ID) to the server. In this process, the terminal connects to the server using data communication technology and sends data using a secure protocol (e.g., HTTPS). The input is the text information and metadata from the terminal, and the output is the data received by the server.

[0544] Step 3:

[0545] The server applies natural language processing to the received text information. The input is text information. The server tokenizes this text data, tags it with parts of speech, and applies named entity recognition to identify proper nouns. Specifically, the server uses Python's natural language processing library. The output is the parsed text data.

[0546] Step 4:

[0547] The server executes a machine learning algorithm based on the analyzed text data to classify the dialogue data into predefined categories. The input is the server's analysis result. Using a machine learning library, the server evaluates the words and context within the data and determines the appropriate category using an algorithm. The output is the classified data.

[0548] Step 5:

[0549] The server records the classified data in a historical information base. The input is the classified data, and the output is the data stored in the historical information base. The server uses a database management system to perform the specific actions of storing the data in the appropriate format.

[0550] Step 6:

[0551] The server suggests setting a reminder to the user if the dialogue data includes time-sensitive tasks. The input is categorized data. The output is a reminder suggestion. The server uses the Calendar API to manage reminders and performs the specific action of sending configurable suggestions to the user.

[0552] Step 7:

[0553] The user searches and views categorized past conversation history on the terminal's operation screen. Input consists of search criteria entered through the operation screen. Output is the conversation history displayed as search results. The terminal performs specific actions, such as retrieving and displaying information from the database, via the user interface.

[0554] (Application Example 1)

[0555] 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."

[0556] In modern households, managing household chores and schedules has become increasingly complex, requiring users to efficiently organize vast amounts of information and smoothly complete daily tasks. However, doing this manually is time-consuming, laborious, and inefficient. Furthermore, there is a lack of readily available voice-activated tools for easily managing information and adjusting schedules within the home. Therefore, there is a need for a system that efficiently supports household chores and organizes information through voice interaction within the home.

[0557] 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.

[0558] In this invention, the server includes means for analyzing conversation information with the user using natural language processing means and extracting keywords and context; means for automatically classifying the conversation information into predefined classification groups based on the extracted information; means for recording the classified conversation information in the user's history information storage unit and organizing it for searchability; and means for receiving user instructions in the home environment through speech recognition and providing household support. This enables efficient home management by allowing the user to give instructions for household chores and daily tasks by voice, easily manage necessary information, and automatically set reminders.

[0559] "Natural language processing" refers to technologies that enable computers to understand and analyze human language, extracting keywords and context through the analysis of text data.

[0560] "Conversational information" refers to the content of communication exchanged between the user and the system via voice or text, and is data that is subject to natural language processing.

[0561] A "classification group" refers to a set of predefined categories used to categorize conversational information based on its content and context.

[0562] The "history information storage unit" refers to a database or data storage area that stores classified conversation information and organizes it so that it can be easily searched later.

[0563] "Speech recognition" is a technology that converts voice input into text data, a means of changing human speech into a format that computers can understand.

[0564] "Household support" refers to the act of providing assistance to efficiently carry out various tasks and management duties that occur within the home, and this invention achieves this using voice recognition.

[0565] "User instructions" refer to commands or inquiries made by the user to the system, which in this invention are given via voice or text.

[0566] A "notification" is a means of conveying information to a user that needs to be communicated at a certain point in time, and it functions as a reminder.

[0567] To implement this invention, a robot or smart device with home communication capabilities is required. When a user provides voice input, the device uses speech recognition technology to convert the speech into text data. Specific technologies that can be used include APIs such as Google Cloud Speech-to-Text. The converted text data is sent to a server and analyzed using a natural language processing (NLP) library (e.g., spaCy). NLP tokenizes the text, performs part-of-speech tagging, and recognizes named entities.

[0568] The server classifies conversational information based on the analyzed data and automatically places it into predefined classification groups. This effectively organizes information based on similar contexts. The classified information is stored in a history information storage unit, making it easily accessible and searchable by the user as needed.

[0569] In a home environment, users can give instructions to the robot via voice recognition, for example, "Set a reminder to put the laundry in the dryer on Saturday." The invention then automatically sets the reminder and sends a notification to the user at the specified date and time. This enables task management based on voice commands as support for daily household chores, allowing users to efficiently continue their daily tasks.

[0570] As a concrete example, suppose a user speaks the following prompt to the terminal:

[0571] "Tell me your house cleaning list."

[0572] "Please set a reminder for taking out the trash next Tuesday."

[0573] "Make a shopping list for the next groceries."

[0574] The information obtained through these interactions is organized by category and used as foundational data to support the performance of necessary tasks. This embodiment enables efficient information management and household support through voice control within the home.

[0575] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0576] Step 1:

[0577] The device receives voice input. When a user speaks to a robot or smart device, the voice data is captured by the microphone in the device.

[0578] Step 2:

[0579] This process converts audio data into text data using a speech recognition API. The audio data is sent to a speech recognition API such as Google Cloud Speech-to-Text, and the output is obtained as text data.

[0580] Step 3:

[0581] The terminal sends the text data to the server. The terminal then sends the converted text data to the server via the internet and prepares it for analysis.

[0582] Step 4:

[0583] The server performs natural language processing. The server uses a natural language processing library such as spaCy to tokenize, tag parts of speech, and recognize named entities in the received text data. Keywords and context are extracted as part of the analysis results.

[0584] Step 5:

[0585] The server automatically classifies conversational information based on the analysis results. Based on the extracted keywords and context, it sorts the conversational information into predefined classification groups and outputs organized information.

[0586] Step 6:

[0587] The classified conversation information is stored in the history information storage unit. The server records the classification results in the history information storage unit to prepare for future searches and references.

[0588] Step 7:

[0589] If the task has a deadline, a reminder will be set. The server identifies the deadline from the conversation information and automatically sets a reminder for the date and time specified by the user.

[0590] Step 8:

[0591] Users can view conversation history and reminders through the dashboard. Through the interface, users can review past conversations and set reminders, and manage and manipulate information as needed.

[0592] 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.

[0593] This invention provides a system for analyzing conversations with users using an AI agent, recognizing not only the content of the conversation but also the user's emotions. This system combines natural language processing means and an emotion engine to analyze and record conversations while taking the user's emotional state into account.

[0594] Start interaction with the user

[0595] The user speaks to the AI ​​agent via voice or text through their device, initiating a conversation. In the case of voice input, the voice data is converted into text data using speech recognition technology.

[0596] Recognition and analysis of emotions

[0597] The device analyzes the user's emotions using an emotion engine along with voice and text data. This engine infers emotions from voice tone, speaking patterns, and text expressions.

[0598] Processing conversation data

[0599] The terminal sends the generated text data and sentiment data to the server. The server performs analysis on the received data using natural language processing. This involves tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context.

[0600] Category classification and emotional integration

[0601] The server classifies conversation data into predefined categories based on the analysis results. It also records recognized emotion data in association with the conversation history. This ensures that data reflecting the user's emotional state is stored in the history database.

[0602] User Feedback

[0603] Based on the recognized emotions, the server adjusts the content of the conversation and suggestions. For example, if the emotion data indicates that the user has a question, it will provide detailed information addressing that question.

[0604] Access to history

[0605] Users can access a dashboard on their device to view and search organized conversation history. Emotion-based filtering is also available, allowing them to view conversation history associated with specific emotional states.

[0606] Specific example

[0607] For example, if a user tells the AI ​​agent, "Yesterday's meeting was exhausting," the emotion engine recognizes the feeling of fatigue. This conversation is categorized as "work-related" and saved in the history database along with the emotion data. Later, when the user searches for "fatigue" as a keyword on the dashboard, a list of related past conversations is displayed, allowing them to see the connection to their emotions at the time. This feature enables users to analyze and utilize past conversations from a deeper perspective, including their emotional state.

[0608] The following describes the processing flow.

[0609] Step 1:

[0610] The user uses a device to initiate a conversation with the AI ​​agent. In the case of voice input, speech recognition is utilized, and the voice data is converted into text data.

[0611] Step 2:

[0612] The device sends voice and text data to an emotion engine to analyze the user's emotional state. This includes voice tone analysis and text keyword analysis.

[0613] Step 3:

[0614] The device sends text data and sentiment data to the server. This includes metadata such as conversation timestamps and user identification information.

[0615] Step 4:

[0616] The server uses natural language processing algorithms to analyze text data. Specifically, it performs tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context from the conversation.

[0617] Step 5:

[0618] Based on the analysis results, the server classifies the conversation data into predefined categories. Machine learning models are used to ensure appropriate classification based on the context and sentiment data of the conversation.

[0619] Step 6:

[0620] The server records classified conversation data and emotion data in the user's history database. This creates a history that takes emotional states into account.

[0621] Step 7:

[0622] The server generates feedback for the user based on the conversation history. Based on sentiment data, it prepares content to provide more detailed information and appropriate suggestions.

[0623] Step 8:

[0624] Users can access the dashboard through their device to view and search past conversation history along with emotional states. This allows them to quickly find conversations associated with specific emotions.

[0625] Step 9:

[0626] Users can input specific conditions or keywords to filter past conversations and perform quantitative or qualitative analysis. This feature allows users to gain insights into emotions and how they change.

[0627] (Example 2)

[0628] 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."

[0629] Conventional conversation analysis systems focus on analyzing the content of user interactions, but they fail to adequately provide information that takes into account the user's emotional state. Furthermore, searching and analyzing past conversation history based on emotions was difficult, making it challenging to provide feedback tailored to user needs. This highlighted the challenge of providing a personalized user experience.

[0630] 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.

[0631] In this invention, the server includes means for analyzing conversation information with the user using natural language processing means and extracting keywords and context; means for recognizing the user's emotional state using emotion analysis means and storing the recognized emotion in association with the conversation information; and means for providing a display function for the user to access classified conversation history and search for history related to a specific emotional state. This enables detailed information provision that takes the user's emotional state into consideration and advanced analysis of past conversations.

[0632] "Natural language processing methods" are technologies that analyze conversational information in the form of speech or text and extract important keywords and context.

[0633] "Emotional analysis methods" refer to technologies that recognize and analyze a user's emotional state based on their voice tone and text content.

[0634] "Conversational information" refers to audio or text data exchanged between the user and the system.

[0635] "Classification" is the process of organizing analyzed conversational information into predefined categories.

[0636] A "history medium" is a database or storage system that accumulates user conversation information and emotional states so that they can be referenced and analyzed later.

[0637] The "display function" is a function that allows users to access accumulated historical information and display that information based on specific conditions or filters.

[0638] This system analyzes conversations between users and AI agents, taking emotions into consideration when providing information. The system primarily consists of terminals and servers.

[0639] The terminal accepts voice or text input from the user, and in the case of voice input, it uses speech recognition software to convert the voice data into text data. A speech recognition API is generally considered the speech recognition technology used here. The user accesses the AI ​​agent through the terminal and initiates a conversation.

[0640] The converted text data is analyzed for the user's emotions using sentiment analysis tools. This analysis utilizes libraries and emotion recognition engines for natural language processing. This allows the user's emotional state to be inferred from their voice tone and text content.

[0641] The server receives text and sentiment data sent from the terminal and performs automatic analysis. Using natural language processing, the server tokenizes the text data, tags parts of speech, and recognizes named entities, extracting keywords and context from the analysis results. This allows the conversational information to be classified into predefined categories and stored in the user's history storage system.

[0642] This system has the ability to adjust user feedback according to perceived emotions. For example, if a user is showing signs of anxiety, the server provides specific information to reassure them. Users can access historical information through the device's dashboard, filter it based on specific emotional states, and investigate past conversations in detail.

[0643] For example, if a user tells the AI ​​agent, "Yesterday's meeting was exhausting," the emotion engine recognizes "fatigue." This conversation is categorized as work-related and stored in the history media along with the emotion data. Later, when the user searches for "fatigue" on the dashboard, the relevant conversation is displayed, allowing them to analyze their emotional state at the time.

[0644] Examples of prompts include, "What are your thoughts on yesterday's project meeting?" or "Tell me about any stress you've experienced at work recently." These prompts serve as a starting point for user interaction and improve the accuracy of the AI ​​system's sentiment analysis and information provision.

[0645] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0646] Step 1:

[0647] The user speaks to the AI ​​agent via the device in either voice or text format. The device receives voice or text data as input. In the case of voice input, the device uses speech recognition technology to convert the voice data into text data. A speech recognition API is utilized in this process. The output is text data.

[0648] Step 2:

[0649] The terminal sends the converted text data to the sentiment analysis system. It receives text data as input and uses a sentiment analysis engine to identify the user's emotions. This process infers emotions based on tone and expression. The output is analyzed data including emotion labels.

[0650] Step 3:

[0651] The terminal sends text data and sentiment labels to the server. The server receives this data as input and performs detailed analysis using natural language processing techniques. It performs processes such as tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context. The output is structured data containing the analysis results.

[0652] Step 4:

[0653] The server classifies conversation information into predefined categories based on the analysis results. This process uses extracted keywords to identify the conversation's theme. It accepts structured data as input and outputs data with category labels.

[0654] Step 5:

[0655] The server integrates recognized sentiment and category labels into the conversation history and stores them in the user's history medium. The input is sentiment and category-labeled data, and the output is an updated history database.

[0656] Step 6:

[0657] The server generates feedback for the user based on the recognized emotion and category. It references the current emotional state and historical data as input and outputs information and advice tailored to the user's situation.

[0658] Step 7:

[0659] Users can view categorized conversation history through the device's dashboard and filter information based on specific emotional states. The input is an emotional filter condition, and the output is the filtered conversation history. This allows users to analyze past conversations from an emotional perspective.

[0660] (Application Example 2)

[0661] 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."

[0662] In communicating with the elderly and individuals requiring care, it is essential to accurately understand their emotional state and respond appropriately based on that understanding. However, conventional technologies have made it difficult to accurately recognize the emotions of users during communication and respond in real time based on those emotions. Therefore, there is a need for technology that can effectively apply emotion recognition to care settings and provide appropriate support in accordance with the emotions of users.

[0663] 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.

[0664] In this invention, the server includes means for analyzing user interaction data using natural language processing means and extracting keywords and context; means for automatically classifying the interaction data into predefined categories based on the extracted information; and means for recognizing and recording the user's emotional state from speech and text using an emotion analysis engine. This makes it possible to accurately grasp the user's emotional state and quickly provide appropriate responses accordingly.

[0665] "Natural language processing" refers to technologies that analyze speech and text data from users to understand language structure.

[0666] An "emotion analysis engine" is a technology that infers and analyzes a user's emotional state from their voice characteristics and textual expression.

[0667] "Users" refers to the individuals who use the system, particularly the elderly and those requiring care.

[0668] "Dialogue data" refers to data that constitutes the content of conversations exchanged between users and the system.

[0669] A "category" refers to a predefined field or theme used when classifying dialogue data based on extracted information.

[0670] A "history information system" is a database that records analyzed and classified dialogue data and organizes a user's past conversation history.

[0671] "Display interface" refers to the screen or dashboard that allows users to access and search historical information.

[0672] The system implementing this invention is configured around a user terminal and a server. Users interact through a terminal such as a smartphone or smart glasses. This terminal analyzes speech or text data using natural language processing means and extracts keywords and context. In the case of speech input, the terminal converts speech data into text data using speech recognition technology. Google Cloud Speech-to-Text is used in this process.

[0673] The server receives data transmitted from the terminal and uses an emotion analysis engine to recognize and record the user's emotional state from the speech and text. IBM Watson Tone Analyzer is used for emotion analysis, and spaCy is also used as part of natural language processing technology. This automatically classifies the conversation data into predefined categories.

[0674] Based on the recognized emotions, the server makes appropriate response suggestions. The feedback mechanism for this purpose generates suggestions and reminders tailored to the user's emotions and notifies care staff and the user themselves.

[0675] For example, if a user says, "I'm feeling a little lonely today," the system will detect "loneliness" as an emotion and provide staff with a suggestion such as, "The user is feeling lonely, so we recommend you talk to them." An example of a prompt for the generating AI model would be, "Perform sentiment analysis based on the following conversation data and generate appropriate care suggestions. Conversation: 'I'm feeling a little lonely today.'"

[0676] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0677] Step 1:

[0678] The user speaks to the device via voice or text, which initiates the interaction. If voice input is provided, the device uses Google Cloud Speech-to-Text to convert the voice data into text data. The input is the user's voice data, and the output is text data. Specifically, the device's microphone captures the voice, which is then processed by the cloud service.

[0679] Step 2:

[0680] The terminal performs natural language processing based on the acquired text data to extract keywords and context. This process uses spaCy for tokenization and part-of-speech tagging. The input is the text data obtained in step 1, and the output is the analyzed keywords and context information. Specifically, the analyzed data is stored in memory and sent to subsequent processing.

[0681] Step 3:

[0682] The terminal uses an emotion analysis engine to infer emotional states from text and voice characteristics. Input is user text data and voice characteristic information, and output is user emotional state data. Specifically, IBM Watson Tone Analyzer analyzes the emotion of the text and saves the result as a status.

[0683] Step 4:

[0684] The server receives keywords, contextual information, and sentiment data sent from the terminal and automatically classifies the dialogue data into predefined categories. The input is the analyzed keywords and sentiment data, and the output is the classification result. The specific operation is the process of executing the classification algorithm and saving the data to the database.

[0685] Step 5:

[0686] The server generates appropriate response suggestions based on classified data and sentiment data, and notifies the user or care staff. The input is the classification results and sentiment state from step 4, and the output is the response suggestion message. The specific operation is to execute the logic for generating response suggestions and send the message to the relevant terminal via the notification function.

[0687] 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.

[0688] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0689] 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.

[0690] [Fourth Embodiment]

[0691] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0692] 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.

[0693] 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).

[0694] 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.

[0695] 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.

[0696] 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).

[0697] 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.

[0698] 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.

[0699] 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.

[0700] 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.

[0701] 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.

[0702] 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.

[0703] 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".

[0704] As an embodiment of the present invention, a system that utilizes an AI agent to efficiently classify and manage conversation data with users will be described.

[0705] Start interaction with the user

[0706] The interaction begins when the user speaks to the AI ​​agent through the device. Input can be provided as either text or voice. In the case of voice input, the device uses speech recognition technology to convert the voice data into text data.

[0707] Processing conversation data

[0708] The terminal sends the converted text data and its metadata to the server. The server receives this data and applies natural language processing (NLP) to it. In the analysis, the text is tokenized and broken down into individual words and phrases, each tagged with its part of speech. Then, named entity recognition is used to extract specific proper nouns.

[0709] Data classification and organization

[0710] Based on the analysis results, the server classifies the conversation data into predefined categories. This classification utilizes automated machine learning algorithms to determine the appropriate category based on the similarity of keywords and context within the conversation. The classified data is then recorded in a searchable historical database.

[0711] Implementing the reminder function

[0712] If a time-sensitive task is detected in the conversation, the server will suggest setting a reminder to the user. This reminder will be managed to ensure that the user receives timely notifications about the scheduled task.

[0713] Access to history

[0714] Users can access the dashboard from their devices and view their past conversation history. The history is organized by category, and users can easily search based on specific keywords or topics. This allows users to quickly access the information they need and support their current and future decision-making by referring to past conversations.

[0715] Specific example

[0716] For example, if a user asks their device, "Please recommend some slide designs for my presentation next week," this conversation is categorized as "business support." The server records this information in the conversation history database, and when the user asks about "slide designs" again later, it can refer to the relevant history and provide suggestions based on past information. In this way, users can efficiently utilize the AI ​​agent while maintaining the context of the conversation.

[0717] The following describes the processing flow.

[0718] Step 1:

[0719] The user uses their device to speak to the AI ​​agent and initiate a conversation. The user can input information via voice or text.

[0720] Step 2:

[0721] When the device receives voice input, it uses speech recognition technology to convert the speech into text, and then generates text data.

[0722] Step 3:

[0723] The terminal sends text data from the user to the server. This data includes metadata such as input timestamps and device information.

[0724] Step 4:

[0725] The server analyzes the received text data. It uses natural language processing algorithms to perform tokenization, part-of-speech tagging, and named entity recognition.

[0726] Step 5:

[0727] The server identifies the subject of the conversation data based on the analysis results and classifies it into predefined categories. This process is automated using machine learning models.

[0728] Step 6:

[0729] The server records the classified conversation data in a history database and organizes the data in a way that makes it easily searchable.

[0730] Step 7:

[0731] The server extracts time-sensitive tasks identified during the conversation and sends reminder suggestions for these tasks to the user.

[0732] Step 8:

[0733] The user reviews the reminder suggestions sent from the server and sets the reminder content and notification time as needed.

[0734] Step 9:

[0735] Based on the user's selection, the server completes the reminder settings and prepares notifications according to the set time.

[0736] Step 10:

[0737] Users can access the dashboard via their device to view and search organized conversation history. They can quickly find past information by specific categories or keywords.

[0738] (Example 1)

[0739] 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".

[0740] In recent years, there has been a growing demand for users to utilize generative AI models to efficiently extract necessary information from conversational data. However, conventional technologies have often been cumbersome in terms of data management and information retrieval. Furthermore, there has been a lack of automated systems for maintaining the context of conversations and setting appropriate reminders. To address these challenges, a method is needed that combines more advanced natural language processing and speech recognition to enable efficient data classification and access.

[0741] 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.

[0742] In this invention, the server includes means for analyzing user dialogue data using natural language processing means and extracting words and contexts, means for using speech recognition technology to convert speech into text information, and means for using a machine learning algorithm to determine classifications based on the analysis results. This enables users to efficiently analyze and organize dialogue data and utilize appropriate reminder functions.

[0743] "Natural language processing methods" refer to a series of techniques that use computer technology to analyze human language and extract meaning from text data.

[0744] A "user" is the entity that operates this system and receives the input and output dialogue data.

[0745] "Dialogue data" refers to communication data in voice or text format that users provide to the system.

[0746] A "word or phrase" refers to a word or short expression contained within a text, and is the basic unit of semantic analysis.

[0747] "Context" refers to the information surrounding a particular word or phrase, determining how it fits into a dialogue and thus its meaning.

[0748] "Means of extraction" refers to the technical process of extracting necessary information from dialogue data.

[0749] "Speech recognition technology" is a technology that processes human speech as a digital signal and converts it into text data.

[0750] "Character information" refers to string data expressed in a format that can be processed by a computer.

[0751] A "machine learning algorithm" is a method in which a computer learns patterns from large amounts of data and uses that knowledge to make predictions and classifications on new data.

[0752] A "reminder" is a function that notifies users based on specific times or conditions to remind them of appointments or tasks.

[0753] An "operation screen" is an interface that allows users to operate the system and view information.

[0754] This invention provides a system that utilizes an AI agent to efficiently manage user interaction data. This system mainly consists of a server, a terminal, and a user interface.

[0755] The device receives voice or text input from the user. In the case of voice input, the device uses speech recognition technology to convert the voice data into text. This conversion utilizes commonly used speech recognition services.

[0756] The converted text information and its metadata are sent from the terminal to the server. The server analyzes the received data using natural language processing (NLP) techniques. NLP includes tokenization, part-of-speech tagging, and named entity recognition. To perform these tasks, natural language processing libraries such as Python's NLTK and spaCy are utilized.

[0757] Based on the analysis results, the server classifies the dialogue data into predefined categories using a machine learning algorithm. This classification process evaluates the words and context within the data to determine the appropriate category. The classification results are recorded in a history database and organized into a format that can be searched by users at a later date.

[0758] Additionally, if the interaction data includes time-sensitive tasks, the server will suggest setting a reminder for the user. The reminder will be sent at an appropriate time using a standard calendar API.

[0759] Users can access their past conversation history through the terminal's interface and easily search for it as needed. The history is organized by category for user convenience, allowing for efficient information retrieval using specific keywords or topics.

[0760] For example, if a user says to their device, "Please prepare the materials for next week's meeting," this conversation is analyzed by the server and classified as a task with a deadline. Based on this information, the server can suggest that the user set a reminder. Examples of prompts include "next week's meeting," "prepare materials," and "I want to set a reminder."

[0761] In this way, this system allows users to efficiently manage conversational data and quickly access the information they need.

[0762] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0763] Step 1:

[0764] The user speaks to the AI ​​agent through the device. Input can be voice or text. In the case of voice input, the device captures the voice using its built-in microphone. This voice data is the initial input. The device converts this voice data into text information using speech recognition technology. This technology is performed by a speech recognition service. The output is text information.

[0765] Step 2:

[0766] The terminal sends text information and associated metadata (such as a time stamp and user ID) to the server. In this process, the terminal connects to the server using data communication technology and sends data using a secure protocol (e.g., HTTPS). The input is the text information and metadata from the terminal, and the output is the data received by the server.

[0767] Step 3:

[0768] The server applies natural language processing to the received text information. The input is text information. The server tokenizes this text data, tags it with parts of speech, and applies named entity recognition to identify proper nouns. Specifically, the server uses Python's natural language processing library. The output is the parsed text data.

[0769] Step 4:

[0770] The server executes a machine learning algorithm based on the analyzed text data to classify the dialogue data into predefined categories. The input is the server's analysis result. Using a machine learning library, the server evaluates the words and context within the data and determines the appropriate category using an algorithm. The output is the classified data.

[0771] Step 5:

[0772] The server records the classified data in a historical information base. The input is the classified data, and the output is the data stored in the historical information base. The server uses a database management system to perform the specific actions of storing the data in the appropriate format.

[0773] Step 6:

[0774] The server suggests setting a reminder to the user if the dialogue data includes time-sensitive tasks. The input is categorized data. The output is a reminder suggestion. The server uses the Calendar API to manage reminders and performs the specific action of sending configurable suggestions to the user.

[0775] Step 7:

[0776] The user searches and views categorized past conversation history on the terminal's operation screen. Input consists of search criteria entered through the operation screen. Output is the conversation history displayed as search results. The terminal performs specific actions, such as retrieving and displaying information from the database, via the user interface.

[0777] (Application Example 1)

[0778] 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".

[0779] In modern households, managing household chores and schedules has become increasingly complex, requiring users to efficiently organize vast amounts of information and smoothly complete daily tasks. However, doing this manually is time-consuming, laborious, and inefficient. Furthermore, there is a lack of readily available voice-activated tools for easily managing information and adjusting schedules within the home. Therefore, there is a need for a system that efficiently supports household chores and organizes information through voice interaction within the home.

[0780] 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.

[0781] In this invention, the server includes means for analyzing conversation information with the user using natural language processing means and extracting keywords and context; means for automatically classifying the conversation information into predefined classification groups based on the extracted information; means for recording the classified conversation information in the user's history information storage unit and organizing it for searchability; and means for receiving user instructions in the home environment through speech recognition and providing household support. This enables efficient home management by allowing the user to give instructions for household chores and daily tasks by voice, easily manage necessary information, and automatically set reminders.

[0782] "Natural language processing" refers to technologies that enable computers to understand and analyze human language, extracting keywords and context through the analysis of text data.

[0783] "Conversational information" refers to the content of communication exchanged between the user and the system via voice or text, and is data that is subject to natural language processing.

[0784] A "classification group" refers to a set of predefined categories used to categorize conversational information based on its content and context.

[0785] The "history information storage unit" refers to a database or data storage area that stores classified conversation information and organizes it so that it can be easily searched later.

[0786] "Speech recognition" is a technology that converts voice input into text data, a means of changing human speech into a format that computers can understand.

[0787] "Household support" refers to the act of providing assistance to efficiently carry out various tasks and management duties that occur within the home, and this invention achieves this using voice recognition.

[0788] "User instructions" refer to commands or inquiries made by the user to the system, which in this invention are given via voice or text.

[0789] A "notification" is a means of conveying information to a user that needs to be communicated at a certain point in time, and it functions as a reminder.

[0790] To implement this invention, a robot or smart device with home communication capabilities is required. When a user provides voice input, the device uses speech recognition technology to convert the speech into text data. Specific technologies that can be used include APIs such as Google Cloud Speech-to-Text. The converted text data is sent to a server and analyzed using a natural language processing (NLP) library (e.g., spaCy). NLP tokenizes the text, performs part-of-speech tagging, and recognizes named entities.

[0791] The server classifies conversational information based on the analyzed data and automatically places it into predefined classification groups. This effectively organizes information based on similar contexts. The classified information is stored in a history information storage unit, making it easily accessible and searchable by the user as needed.

[0792] In a home environment, users can give instructions to the robot via voice recognition, for example, "Set a reminder to put the laundry in the dryer on Saturday." The invention then automatically sets the reminder and sends a notification to the user at the specified date and time. This enables task management based on voice commands as support for daily household chores, allowing users to efficiently continue their daily tasks.

[0793] As a concrete example, suppose a user speaks the following prompt to the terminal:

[0794] "Tell me your house cleaning list."

[0795] "Please set a reminder for taking out the trash next Tuesday."

[0796] "Make a shopping list for the next groceries."

[0797] The information obtained through these interactions is organized by category and used as foundational data to support the performance of necessary tasks. This embodiment enables efficient information management and household support through voice control within the home.

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

[0799] Step 1:

[0800] The device receives voice input. When a user speaks to a robot or smart device, the voice data is captured by the microphone in the device.

[0801] Step 2:

[0802] This process converts audio data into text data using a speech recognition API. The audio data is sent to a speech recognition API such as Google Cloud Speech-to-Text, and the output is obtained as text data.

[0803] Step 3:

[0804] The terminal sends the text data to the server. The terminal then sends the converted text data to the server via the internet and prepares it for analysis.

[0805] Step 4:

[0806] The server performs natural language processing. The server uses a natural language processing library such as spaCy to tokenize, tag parts of speech, and recognize named entities in the received text data. Keywords and context are extracted as part of the analysis results.

[0807] Step 5:

[0808] The server automatically classifies conversational information based on the analysis results. Based on the extracted keywords and context, it sorts the conversational information into predefined classification groups and outputs organized information.

[0809] Step 6:

[0810] The classified conversation information is stored in the history information storage unit. The server records the classification results in the history information storage unit to prepare for future searches and references.

[0811] Step 7:

[0812] If the task has a deadline, a reminder will be set. The server identifies the deadline from the conversation information and automatically sets a reminder for the date and time specified by the user.

[0813] Step 8:

[0814] Users can view conversation history and reminders through the dashboard. Through the interface, users can review past conversations and set reminders, and manage and manipulate information as needed.

[0815] 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.

[0816] This invention provides a system for analyzing conversations with users using an AI agent, recognizing not only the content of the conversation but also the user's emotions. This system combines natural language processing means and an emotion engine to analyze and record conversations while taking the user's emotional state into account.

[0817] Start interaction with the user

[0818] The user speaks to the AI ​​agent via voice or text through their device, initiating a conversation. In the case of voice input, the voice data is converted into text data using speech recognition technology.

[0819] Recognition and analysis of emotions

[0820] The device analyzes the user's emotions using an emotion engine along with voice and text data. This engine infers emotions from voice tone, speaking patterns, and text expressions.

[0821] Processing conversation data

[0822] The terminal sends the generated text data and sentiment data to the server. The server performs analysis on the received data using natural language processing. This involves tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context.

[0823] Category classification and emotional integration

[0824] The server classifies conversation data into predefined categories based on the analysis results. It also records recognized emotion data in association with the conversation history. This ensures that data reflecting the user's emotional state is stored in the history database.

[0825] User Feedback

[0826] Based on the recognized emotions, the server adjusts the content of the conversation and suggestions. For example, if the emotion data indicates that the user has a question, it will provide detailed information addressing that question.

[0827] Access to history

[0828] Users can access a dashboard on their device to view and search organized conversation history. Emotion-based filtering is also available, allowing them to view conversation history associated with specific emotional states.

[0829] Specific example

[0830] For example, if a user tells the AI ​​agent, "Yesterday's meeting was exhausting," the emotion engine recognizes the feeling of fatigue. This conversation is categorized as "work-related" and saved in the history database along with the emotion data. Later, when the user searches for "fatigue" as a keyword on the dashboard, a list of related past conversations is displayed, allowing them to see the connection to their emotions at the time. This feature enables users to analyze and utilize past conversations from a deeper perspective, including their emotional state.

[0831] The following describes the processing flow.

[0832] Step 1:

[0833] The user uses a device to initiate a conversation with the AI ​​agent. In the case of voice input, speech recognition is utilized, and the voice data is converted into text data.

[0834] Step 2:

[0835] The device sends voice and text data to an emotion engine to analyze the user's emotional state. This includes voice tone analysis and text keyword analysis.

[0836] Step 3:

[0837] The device sends text data and sentiment data to the server. This includes metadata such as conversation timestamps and user identification information.

[0838] Step 4:

[0839] The server uses natural language processing algorithms to analyze text data. Specifically, it performs tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context from the conversation.

[0840] Step 5:

[0841] Based on the analysis results, the server classifies the conversation data into predefined categories. Machine learning models are used to ensure appropriate classification based on the context and sentiment data of the conversation.

[0842] Step 6:

[0843] The server records classified conversation data and emotion data in the user's history database. This creates a history that takes emotional states into account.

[0844] Step 7:

[0845] The server generates feedback for the user based on the conversation history. Based on sentiment data, it prepares content to provide more detailed information and appropriate suggestions.

[0846] Step 8:

[0847] Users can access the dashboard through their device to view and search past conversation history along with emotional states. This allows them to quickly find conversations associated with specific emotions.

[0848] Step 9:

[0849] Users can input specific conditions or keywords to filter past conversations and perform quantitative or qualitative analysis. This feature allows users to gain insights into emotions and how they change.

[0850] (Example 2)

[0851] 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".

[0852] Conventional conversation analysis systems focus on analyzing the content of user interactions, but they fail to adequately provide information that takes into account the user's emotional state. Furthermore, searching and analyzing past conversation history based on emotions was difficult, making it challenging to provide feedback tailored to user needs. This highlighted the challenge of providing a personalized user experience.

[0853] 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.

[0854] In this invention, the server includes means for analyzing conversation information with the user using natural language processing means and extracting keywords and context; means for recognizing the user's emotional state using emotion analysis means and storing the recognized emotion in association with the conversation information; and means for providing a display function for the user to access classified conversation history and search for history related to a specific emotional state. This enables detailed information provision that takes the user's emotional state into consideration and advanced analysis of past conversations.

[0855] "Natural language processing methods" are technologies that analyze conversational information in the form of speech or text and extract important keywords and context.

[0856] "Emotional analysis methods" refer to technologies that recognize and analyze a user's emotional state based on their voice tone and text content.

[0857] "Conversational information" refers to audio or text data exchanged between the user and the system.

[0858] "Classification" is the process of organizing analyzed conversational information into predefined categories.

[0859] A "history medium" is a database or storage system that accumulates user conversation information and emotional states so that they can be referenced and analyzed later.

[0860] The "display function" is a function that allows users to access accumulated historical information and display that information based on specific conditions or filters.

[0861] This system analyzes conversations between users and AI agents, taking emotions into consideration when providing information. The system primarily consists of terminals and servers.

[0862] The terminal accepts voice or text input from the user, and in the case of voice input, it uses speech recognition software to convert the voice data into text data. A speech recognition API is generally considered the speech recognition technology used here. The user accesses the AI ​​agent through the terminal and initiates a conversation.

[0863] The converted text data is analyzed for the user's emotions using sentiment analysis tools. This analysis utilizes libraries and emotion recognition engines for natural language processing. This allows the user's emotional state to be inferred from their voice tone and text content.

[0864] The server receives text and sentiment data sent from the terminal and performs automatic analysis. Using natural language processing, the server tokenizes the text data, tags parts of speech, and recognizes named entities, extracting keywords and context from the analysis results. This allows the conversational information to be classified into predefined categories and stored in the user's history storage system.

[0865] This system has the ability to adjust user feedback according to perceived emotions. For example, if a user is showing signs of anxiety, the server provides specific information to reassure them. Users can access historical information through the device's dashboard, filter it based on specific emotional states, and investigate past conversations in detail.

[0866] For example, if a user tells the AI ​​agent, "Yesterday's meeting was exhausting," the emotion engine recognizes "fatigue." This conversation is categorized as work-related and stored in the history media along with the emotion data. Later, when the user searches for "fatigue" on the dashboard, the relevant conversation is displayed, allowing them to analyze their emotional state at the time.

[0867] Examples of prompts include, "What are your thoughts on yesterday's project meeting?" or "Tell me about any stress you've experienced at work recently." These prompts serve as a starting point for user interaction and improve the accuracy of the AI ​​system's sentiment analysis and information provision.

[0868] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0869] Step 1:

[0870] The user speaks to the AI ​​agent via the device in either voice or text format. The device receives voice or text data as input. In the case of voice input, the device uses speech recognition technology to convert the voice data into text data. A speech recognition API is utilized in this process. The output is text data.

[0871] Step 2:

[0872] The terminal sends the converted text data to the sentiment analysis system. It receives text data as input and uses a sentiment analysis engine to identify the user's emotions. This process infers emotions based on tone and expression. The output is analyzed data including emotion labels.

[0873] Step 3:

[0874] The terminal sends text data and sentiment labels to the server. The server receives this data as input and performs detailed analysis using natural language processing techniques. It performs processes such as tokenization, part-of-speech tagging, and named entity recognition to extract keywords and context. The output is structured data containing the analysis results.

[0875] Step 4:

[0876] The server classifies conversation information into predefined categories based on the analysis results. This process uses extracted keywords to identify the conversation's theme. It accepts structured data as input and outputs data with category labels.

[0877] Step 5:

[0878] The server integrates recognized sentiment and category labels into the conversation history and stores them in the user's history medium. The input is sentiment and category-labeled data, and the output is an updated history database.

[0879] Step 6:

[0880] The server generates feedback for the user based on the recognized emotion and category. It references the current emotional state and historical data as input and outputs information and advice tailored to the user's situation.

[0881] Step 7:

[0882] Users can view categorized conversation history through the device's dashboard and filter information based on specific emotional states. The input is an emotional filter condition, and the output is the filtered conversation history. This allows users to analyze past conversations from an emotional perspective.

[0883] (Application Example 2)

[0884] 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".

[0885] In communicating with the elderly and individuals requiring care, it is essential to accurately understand their emotional state and respond appropriately based on that understanding. However, conventional technologies have made it difficult to accurately recognize the emotions of users during communication and respond in real time based on those emotions. Therefore, there is a need for technology that can effectively apply emotion recognition to care settings and provide appropriate support in accordance with the emotions of users.

[0886] 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.

[0887] In this invention, the server includes means for analyzing user interaction data using natural language processing means and extracting keywords and context; means for automatically classifying the interaction data into predefined categories based on the extracted information; and means for recognizing and recording the user's emotional state from speech and text using an emotion analysis engine. This makes it possible to accurately grasp the user's emotional state and quickly provide appropriate responses accordingly.

[0888] "Natural language processing" refers to technologies that analyze speech and text data from users to understand language structure.

[0889] An "emotion analysis engine" is a technology that infers and analyzes a user's emotional state from their voice characteristics and textual expression.

[0890] "Users" refers to the individuals who use the system, particularly the elderly and those requiring care.

[0891] "Dialogue data" refers to data that constitutes the content of conversations exchanged between users and the system.

[0892] A "category" refers to a predefined field or theme used when classifying dialogue data based on extracted information.

[0893] A "history information system" is a database that records analyzed and classified dialogue data and organizes a user's past conversation history.

[0894] "Display interface" refers to the screen or dashboard that allows users to access and search historical information.

[0895] The system implementing this invention is configured around a user terminal and a server. Users interact through a terminal such as a smartphone or smart glasses. This terminal analyzes speech or text data using natural language processing means and extracts keywords and context. In the case of speech input, the terminal converts speech data into text data using speech recognition technology. Google Cloud Speech-to-Text is used in this process.

[0896] The server receives data transmitted from the terminal and uses an emotion analysis engine to recognize and record the user's emotional state from the speech and text. IBM Watson Tone Analyzer is used for emotion analysis, and spaCy is also used as part of natural language processing technology. This automatically classifies the conversation data into predefined categories.

[0897] Based on the recognized emotions, the server makes appropriate response suggestions. The feedback mechanism for this purpose generates suggestions and reminders tailored to the user's emotions and notifies care staff and the user themselves.

[0898] For example, if a user says, "I'm feeling a little lonely today," the system will detect "loneliness" as an emotion and provide staff with a suggestion such as, "The user is feeling lonely, so we recommend you talk to them." An example of a prompt for the generating AI model would be, "Perform sentiment analysis based on the following conversation data and generate appropriate care suggestions. Conversation: 'I'm feeling a little lonely today.'"

[0899] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0900] Step 1:

[0901] The user speaks to the device via voice or text, which initiates the interaction. If voice input is provided, the device uses Google Cloud Speech-to-Text to convert the voice data into text data. The input is the user's voice data, and the output is text data. Specifically, the device's microphone captures the voice, which is then processed by the cloud service.

[0902] Step 2:

[0903] The terminal performs natural language processing based on the acquired text data to extract keywords and context. This process uses spaCy for tokenization and part-of-speech tagging. The input is the text data obtained in step 1, and the output is the analyzed keywords and context information. Specifically, the analyzed data is stored in memory and sent to subsequent processing.

[0904] Step 3:

[0905] The terminal uses an emotion analysis engine to infer emotional states from text and voice characteristics. Input is user text data and voice characteristic information, and output is user emotional state data. Specifically, IBM Watson Tone Analyzer analyzes the emotion of the text and saves the result as a status.

[0906] Step 4:

[0907] The server receives keywords, contextual information, and sentiment data sent from the terminal and automatically classifies the dialogue data into predefined categories. The input is the analyzed keywords and sentiment data, and the output is the classification result. The specific operation is the process of executing the classification algorithm and saving the data to the database.

[0908] Step 5:

[0909] The server generates appropriate response suggestions based on classified data and sentiment data, and notifies the user or care staff. The input is the classification results and sentiment state from step 4, and the output is the response suggestion message. The specific operation is to execute the logic for generating response suggestions and send the message to the relevant terminal via the notification function.

[0910] 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.

[0911] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0912] 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.

[0913] 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.

[0914] 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.

[0915] 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.

[0916] 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.

[0917] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[0918] 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."

[0919] 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.

[0920] 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.

[0921] 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.

[0922] 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.

[0923] 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.

[0924] 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.

[0925] 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.

[0926] 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.

[0927] 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.

[0928] 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.

[0929] 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.

[0930] 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.

[0931] The following is further disclosed regarding the embodiments described above.

[0932] (Claim 1)

[0933] A means of analyzing user conversation data using natural language processing methods and extracting keywords and context,

[0934] A means for automatically classifying conversation data into predefined categories based on extracted information,

[0935] A means of recording classified conversation data in a user history database and organizing it for searchability,

[0936] A means to identify time-sensitive tasks from conversations and set reminders for users,

[0937] A means of providing a dashboard for users to access and search categorized conversation history,

[0938] A system that includes this.

[0939] (Claim 2)

[0940] The system according to claim 1, comprising means for performing tokenization, part-of-speech tagging, and named entity recognition as natural language processing means.

[0941] (Claim 3)

[0942] The system according to claim 1, further comprising means for learning from past conversations based on classified conversation data and making category suggestions in future conversations.

[0943] "Example 1"

[0944] (Claim 1)

[0945] A means of analyzing user interaction data using natural language processing methods and extracting words and contexts,

[0946] A means for automatically classifying dialogue data into predefined categories based on extracted information,

[0947] A means of recording classified dialogue data in a user history information base and organizing it in a searchable manner,

[0948] A means to identify time-sensitive tasks through dialogue and suggest notification settings to the user,

[0949] A means of providing a user interface for accessing and exploring categorized conversation history,

[0950] A means of using speech recognition technology to convert speech into text information,

[0951] A means of using a machine learning algorithm to determine classification based on the analysis results,

[0952] A system that includes this.

[0953] (Claim 2)

[0954] The system according to claim 1, comprising means for performing tokenization, part-of-speech tagging, and naming entity recognition.

[0955] (Claim 3)

[0956] The system according to claim 1, further comprising means for learning from past dialogues based on classified dialogue data and making classification suggestions in future dialogues.

[0957] "Application Example 1"

[0958] (Claim 1)

[0959] A means of analyzing conversational information with a user using natural language processing methods and extracting keywords and context,

[0960] A means for automatically classifying conversation information into predefined classification groups based on extracted information,

[0961] A means for recording classified conversation information in the user's history information storage unit and organizing it in a searchable manner,

[0962] A means to identify time-sensitive tasks from conversations and set up notifications for users,

[0963] A means for providing a display unit for users to access and search categorized conversation history,

[0964] A means of receiving user instructions in a home environment via voice recognition and providing assistance with household chores,

[0965] A system that includes this.

[0966] (Claim 2)

[0967] The system according to claim 1, comprising means for performing tokenization, part-of-speech tagging, and specific noun recognition as natural language processing means.

[0968] (Claim 3)

[0969] The system according to claim 1, further comprising means for learning from past conversations based on classified conversation information and making classification group suggestions in future conversations.

[0970] "Example 2 of combining an emotion engine"

[0971] (Claim 1)

[0972] A means of analyzing conversational information with a user using natural language processing methods and extracting keywords and context,

[0973] A means for automatically classifying conversation information into predefined categories based on extracted information,

[0974] A means for recording classified conversation information on the user's history medium and organizing it in a searchable manner,

[0975] A means for recognizing the user's emotional state using emotion analysis means, and storing the recognized emotion in association with conversational information,

[0976] A means to provide a display function for users to access categorized conversation history and search for history related to a specific emotional state,

[0977] A system that includes this.

[0978] (Claim 2)

[0979] The system according to claim 1, comprising natural language processing means for performing tokenization, part-of-speech tagging, and named entity recognition.

[0980] (Claim 3)

[0981] The system according to claim 1, comprising means for adjusting the content of a conversation in accordance with recognized emotions and providing feedback to the user.

[0982] "Application example 2 when combining with an emotional engine"

[0983] (Claim 1)

[0984] A means of analyzing user interaction data using natural language processing methods and extracting keywords and context,

[0985] A means for automatically classifying dialogue data into predefined categories based on extracted information,

[0986] A means for recording classified dialogue data in the user history information system and organizing it in a searchable manner,

[0987] A means for recognizing and recording a user's emotional state from speech and text using an emotion analysis engine,

[0988] A means of adjusting user interactions and providing feedback based on recognized emotions,

[0989] A means of providing a display interface for users to access and search categorized conversation history,

[0990] A system that includes this.

[0991] (Claim 2)

[0992] The system according to claim 1, comprising, as an emotion analysis engine, means for estimating an emotional state using voice characteristics and text representation.

[0993] (Claim 3)

[0994] The system according to claim 1, further comprising means for making response suggestions appropriate to the user's emotional state based on classified dialogue data and emotional data. [Explanation of Symbols]

[0995] 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 analyzing user conversation data using natural language processing methods and extracting keywords and context, A means for automatically classifying conversation data into predefined categories based on extracted information, A means of recording classified conversation data in a user history database and organizing it for searchability, A means to identify time-sensitive tasks from conversations and set reminders for users, A means of providing a dashboard for users to access and search categorized conversation history, A system that includes this.

2. The system according to claim 1, comprising means for performing tokenization, part-of-speech tagging, and named entity recognition as natural language processing means.

3. The system according to claim 1, further comprising means for learning from past conversations based on classified conversation data and making category suggestions in future conversations.

Images