system

Abstract

We provide the system. [Solution] An interface means that the user can access, A means of collecting audio data in real time, A means of converting collected audio data into text data, A method for extracting and organizing the key points of a discussion from text data, A means of generating and displaying meeting minutes based on key points, A means of proposing the direction of future discussions, Means for collecting and storing feedback, A system that includes this.

Description

Technical Field

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

Background Art

[0002] Patent Document 1 discloses a method for controlling a persona chatbot, which is performed by at least one processor, including 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] In meetings and discussions, manually recording each statement one by one or summarizing the key points later is a very time-consuming and laborious task. Also, during the progress of the discussion, it is easy to lose sight of the direction of what to discuss next, and as a result, the discussion often does not proceed efficiently and effectively. Furthermore, since it also takes time to check the content of the minutes and reflect feedback, it is difficult for all participants to grasp the progress of the discussion in real time.

Means for Solving the Problems

[0005] The present invention solves the above problems by providing a system that includes an interface means accessible to the user, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for proposing the direction of future discussions, and means for collecting and saving feedback. This system improves the efficiency and depth of discussions by summarizing the main points of the discussion in real time and proposing the content to be discussed next. Furthermore, since the system makes self-improvements based on the collected feedback, it is possible to continuously improve the quality of discussions.

[0006] A "user" is a person who accesses a system and participates in meetings and discussions through its interface.

[0007] "Interface means" refers to a device or software that allows a user to access and operate a system.

[0008] "Means for collecting audio data in real time" refers to devices or software that instantly collect audio spoken during meetings or discussions as digital data.

[0009] "Means for converting collected audio data into text data" refers to a device or software for converting collected audio data into text information. This primarily utilizes speech recognition technology.

[0010] "Means for extracting and organizing the main points of an argument from text data" refers to a device or software that identifies important information and keywords contained in text data and organizes them into an easily understandable format.

[0011] "Means for generating and displaying meeting minutes based on key points" refers to a device or software that creates meeting minutes based on extracted key points and displays them in an easy-to-understand manner for the user.

[0012] "Means for proposing the direction of future discussions" refers to a device or software that suggests the next items or topics to be discussed, based on the progress of the discussion.

[0013] "Means for collecting and storing feedback" refers to devices or software that receive opinions and requests for corrections from users and store them in a database or similar format. [Brief explanation of the drawing]

[0014] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12]It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Example 2 when combined with an emotion engine. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when combined with an emotion engine.

Mode for Carrying Out the Invention

[0015] 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.

[0016] First, the language used in the following description will be explained.

[0017] In the following embodiments, the labeled processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), APU (Accelerated Processing Unit), etc.

[0018] In the following embodiments, the labeled RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0019] 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.

[0020] 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).

[0021] 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."

[0022] [First Embodiment]

[0023] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0024] 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.

[0025] 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).

[0026] 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.

[0027] 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.

[0028] 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.

[0029] 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.

[0030] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0031] 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.

[0032] 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.

[0033] 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.

[0034] 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".

[0035] This invention is a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggesting directions for future discussions. This system is realized through the cooperation of the user, server, and terminal.

[0036] System Configuration

[0037] The system consists of the following main components:

[0038] User interface means: A device (e.g., personal computer, tablet, smartphone) that allows a user to access and operate a system.

[0039] Audio data collection method: This refers to software or hardware (e.g., ZOOM API) that collects meeting audio data in real time.

[0040] Audio data conversion means: This refers to software (e.g., Google® Speech-to-Text API) for converting collected audio data into text data.

[0041] Key Point Extraction Method: This is a natural language processing (NLP) model for extracting and organizing the key points of an argument from converted text data.

[0042] Meeting minutes generation method: This is software that generates meeting minutes based on key points and displays them to the user.

[0043] A means of suggesting the direction of discussion: This is an analytical function for suggesting the next topics to be discussed.

[0044] Feedback collection method: This is software for collecting and saving user feedback and requests for improvements.

[0045] System operation

[0046] 1. User login and meeting setup

[0047] The user accesses the system from their device and enters their username and password on the login page.

[0048] The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[0049] Once logged in, users can select either "Create a new meeting" or "Join an existing meeting" on the interface.

[0050] When creating a meeting, the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user can join by entering the meeting ID.

[0051] 2. Collection of audio data and sequential transcription

[0052] The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[0053] The acquired audio data is sent to a speech recognition API and converted into text data. The server then sequentially saves the converted text data to a database.

[0054] 3. Extracting the key points of the discussion

[0055] The server sends the stored text data to a natural language processing (NLP) model at regular intervals to extract important keywords and phrases.

[0056] The extracted key points are organized by the server and saved as data for meeting minutes.

[0057] 4. Creating and sharing meeting minutes

[0058] The server generates meeting minutes summarizing the discussion based on the key points.

[0059] The generated meeting minutes are converted into text files or HTML format and displayed on the user's device via Zoom's chat function or a dedicated interface.

[0060] 5. Proposal for the direction of the next discussion

[0061] The server analyzes the content of the discussion and uses an NLP model to identify the next topic to discuss.

[0062] The proposal will be notified to the user's device, supporting the progress of the discussion.

[0063] 6. Gathering Feedback

[0064] Users can provide feedback on meeting minutes and proposals from their devices.

[0065] The feedback provided will be collected and stored by the server and used to improve the system.

[0066] Specific example

[0067] For example, consider a scenario where users A, B, and C discuss the marketing strategy for a new product via Zoom. Using this system, the following operations can be performed:

[0068] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[0069] The server collects audio data from the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[0070] The system extracts key points such as "identifying the target market" and "competitor analysis" from text data, and then automatically generates meeting minutes based on this information.

[0071] The generated meeting minutes are displayed in the Zoom chat, allowing all users to view them in real time.

[0072] The server proposes "Pricing for the new product" as the next topic to discuss and notifies the terminal of this.

[0073] If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server saves this and uses it for future analyses.

[0074] In this way, the system of the present invention enables efficient and in-depth discussions by summarizing the key points of the discussion in real time and suggesting the next topics to be discussed.

[0075] The following describes the processing flow.

[0076] Step 1:

[0077] The user accesses the system's login page from their terminal. The user enters their username and password and completes the login process.

[0078] Step 2:

[0079] The server compares the user's authentication information with the database, and if correct, creates a session and returns a session ID to the user.

[0080] Step 3:

[0081] The user selects either "Create a new meeting" or "Join an existing meeting" on the terminal interface. If the user selects "Create a new meeting," the server generates a meeting ID and notifies the user of it.

[0082] Step 4:

[0083] When joining an existing meeting, the user enters the meeting ID, and the server authenticates that ID. Once the user joins the meeting, real-time audio data collection begins.

[0084] Step 5:

[0085] The server uses the ZOOM API to acquire audio data during the meeting in real time. The collected audio data is then sent to the server.

[0086] Step 6:

[0087] The server continuously sends audio data to a speech recognition API (e.g., Google Speech-to-Text API), converting the audio data into text data in real time. The converted text data is then stored in a database.

[0088] Step 7:

[0089] The server sends the stored text data to a natural language processing (NLP) model at regular intervals to extract important keywords and phrases.

[0090] Step 8:

[0091] The server organizes the extracted key points chronologically and saves them as data for meeting minutes.

[0092] Step 9:

[0093] The server generates meeting minutes in real time based on the stored key points data. The generated minutes are then instructed to be displayed on the device via Zoom's chat function or a dedicated interface.

[0094] Step 10:

[0095] The server analyzes the discussion and uses an NLP model to identify the next topic to discuss. The identified topic is generated as a suggestion and notified to the user's terminal.

[0096] Step 11:

[0097] Users input feedback on meeting minutes and proposals through the terminal's interface.

[0098] Step 12:

[0099] The server stores the collected feedback in a database. This stored feedback is then used as training data to improve the system's natural language processing model.

[0100] (Example 1)

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

[0102] In traditional meetings and discussions, creating meeting minutes was cumbersome, making it difficult to summarize key points of important discussions in real time. Furthermore, it was impossible to suggest future directions for discussions, and it was difficult to effectively collect participant feedback and use it for improvement. This resulted in decreased meeting efficiency and the risk of important discussions being overlooked.

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

[0104] In this invention, the server includes an interface means accessible to users, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for analyzing the content of the discussion and suggesting what should be discussed next, means for collecting and saving user feedback, means for using a communication service application programming interface for collecting audio data, means for using a speech recognition application programming interface to convert the audio data into text data, and means for using a natural language processing model for extracting main points. This makes it possible to automatically generate meeting minutes in real time, support the progress of meetings by suggesting what should be discussed next, and effectively collect feedback from participants to improve the system.

[0105] "Interface means" refers to devices and software that allow users to access and operate a system.

[0106] "Audio data" refers to the digital data of audio signals collected during meetings or discussions.

[0107] "Means of real-time collection" refers to equipment and software for instantly collecting audio data during a meeting.

[0108] "Text data" refers to character data generated by transcribing audio data.

[0109] "Methods for extracting and organizing the key points of an argument" refers to natural language processing models and algorithms that find important information from text data and organize it systematically.

[0110] "Means for generating and displaying meeting minutes" refers to software that automatically creates meeting minutes based on extracted key points and presents them to the user.

[0111] "Means of proposing a direction for discussion" refers to a function that analyzes the content of the current discussion and indicates the points that should be discussed next.

[0112] "Means for collecting and saving feedback" refers to a system for collecting user opinions and correction requests and saving them for later use.

[0113] "Application programming interface for communication services" refers to the means of utilizing functions through the programming interface provided by communication services.

[0114] A "Speech Recognition Application Programming Interface" refers to a programming interface that provides speech recognition functionality for converting speech data into text.

[0115] A "natural language processing model" refers to machine learning models and algorithms used to understand and analyze human language.

[0116] The system of this invention collects audio data in real time during meetings and discussions and transcribes it sequentially. It then extracts and organizes the key points of the discussion to automatically generate meeting minutes. Furthermore, it proposes directions for future discussions and collects and stores user feedback. This system is realized through the cooperation of the user, server, and terminal.

[0117] Key components of the system

[0118] Interface means: A device that allows a user to access and operate a system (e.g., a personal computer, tablet, or smartphone).

[0119] Audio data collection means: A communication service application programming interface (e.g., a communication service API) for collecting conference audio data in real time.

[0120] Audio data conversion means: A speech recognition application programming interface (e.g., speech recognition API) for converting collected audio data into text data.

[0121] Key point extraction method: A natural language processing model for extracting and organizing the key points of a discussion from text data.

[0122] Meeting minutes generation method: Software for generating and displaying meeting minutes based on key points.

[0123] A means of suggesting the direction of discussion: A function for analyzing the content of a discussion and suggesting what should be discussed next.

[0124] Feedback collection method: Software for collecting and storing user feedback and correction requests.

[0125] System operation

[0126] The user accesses the system from their terminal and enters their username and password on the login page. The server verifies the authentication information against the database, and if successful, creates a session and returns a session ID to the user. After logging in, the user selects either "Create a new meeting" or "Join an existing meeting." The server generates a meeting ID and notifies the user of it.

[0127] During the meeting, the server uses a communication service API to acquire audio data in real time. The acquired audio data is sent to a speech recognition API and converted into text data. The server then sequentially saves the converted text data to a database.

[0128] The server sends the stored text data to a natural language processing model at regular intervals to extract important keywords and phrases. The extracted key points are then organized by the server and stored as data for meeting minutes.

[0129] The server generates meeting minutes summarizing the discussion based on key points. The generated minutes are converted into text files or HTML format and displayed on the user's device through the communication service's chat function or a dedicated interface.

[0130] Furthermore, the server analyzes the content of the discussion and uses a natural language processing model to identify the next topic to discuss. This suggestion is then communicated to the user's terminal to support the progress of the discussion.

[0131] Users can provide feedback on meeting minutes and proposals from their devices. The feedback provided is collected and stored by the server and used to improve the system.

[0132] Specific example

[0133] For example, consider a scenario where users A, B, and C discuss the marketing strategy for a new product through a communication service. Using this system, the following operations can be performed:

[0134] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[0135] The server collects audio data from the meeting via a communication service API and converts statements such as "We will discuss the target market for the new product" into text in real time.

[0136] The system extracts key points such as "identifying the target market" and "competitor analysis" from text data, and then automatically generates meeting minutes based on this information.

[0137] The generated meeting minutes are displayed in the communication service's chat, allowing all users to view them in real time.

[0138] The server proposes "Pricing for the new product" as the next topic to discuss and notifies the terminal of this.

[0139] If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server saves this and uses it for future analyses.

[0140] Example of a prompt

[0141] "We've developed a system that collects meeting audio data in real time and transcribes it sequentially. It also has a function to identify and suggest the next topic of discussion. For example, if we're discussing the marketing strategy for a new product, and topics like the target market and competitor analysis come up, how can you tell me how to suggest the next topic to discuss?"

[0142] In this way, the system of the present invention enables efficient and in-depth discussions by summarizing the key points of the discussion in real time and suggesting the next topics to be discussed.

[0143] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0144] Step 1: User login and meeting setup

[0145] Specific operation: The user accesses the system from their terminal and enters their username and password on the login page.

[0146] Enter: Username and password.

[0147] Data processing / calculation: The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[0148] Output: Session ID.

[0149] After the user logs in, they can select either "Create a New Meeting" or "Join an Existing Meeting" on the interface. If a meeting is created, the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user enters the meeting ID.

[0150] Step 2: Audio data collection and sequential transcription

[0151] Specific operation: The server uses a communication service API to acquire audio data during the meeting in real time.

[0152] Input: Voice data obtained from a communication service API.

[0153] Data processing / calculation: The server sends the acquired audio data to a speech recognition API and converts it into text data. The server then sequentially saves the converted text data to a database.

[0154] Output: Text data.

[0155] Step 3: Extract the key points of the discussion

[0156] Specific operation: The server sends the stored text data to a natural language processing model at regular intervals, and extracts important keywords and phrases.

[0157] Input: Text data.

[0158] Data Processing / Calculation: A natural language processing model analyzes the text data and extracts important keywords and phrases. The server organizes the extracted key points and saves them as data for meeting minutes.

[0159] Output: Key data.

[0160] Step 4: Generate and share meeting minutes

[0161] Specific operation: The server generates meeting minutes summarizing the discussion based on key points.

[0162] Input: Key data.

[0163] Data Processing / Calculation: Automatically creates meeting minutes based on key data and converts them to text files or HTML format.

[0164] Output: Meeting minutes file (text file / HTML format).

[0165] The generated meeting minutes are displayed on the user's device through the communication service's chat function or a dedicated interface.

[0166] Step 5: Propose the direction of the next discussion

[0167] Specific operation: The server analyzes the content of the discussion and uses a natural language processing model to identify the next topic to discuss.

[0168] Input: Meeting minutes data.

[0169] Data processing / computation: Using a natural language processing model, identify the next topic to discuss. The server notifies the user's terminal of the identified topic as a suggestion.

[0170] Output: Suggestions for the next topic to discuss.

[0171] Step 6: Gathering Feedback

[0172] Specific functionality: Users can provide feedback on meeting minutes and proposals from their devices.

[0173] Input: User feedback.

[0174] Data processing / calculation: The server collects and stores user feedback, which is later used to improve the system.

[0175] Output: Saved feedback data.

[0176] In this way, a series of processes can be realized in which the server, terminal, and user cooperate with each other to summarize the key points of the discussion in real time and propose the next topics to be discussed.

[0177] (Application Example 1)

[0178] 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."

[0179] In modern meetings and discussions, creating meeting minutes, summarizing key points, and proposing directions for future discussions are crucial, yet extremely time-consuming. Therefore, there is a need for a system that collects audio data in real time, automatically generates meeting minutes, and efficiently proposes directions for future discussions. Furthermore, in environments requiring rapid decision-making, such as logistics centers, the implementation of an efficient meeting management system is urgently needed.

[0180] 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.

[0181] In this invention, the server includes an interface means accessible to the user, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for proposing the direction of future discussions, means for collecting and saving feedback, communication means for acquiring audio data during a meeting, means for transmitting the acquired audio data to a speech recognition platform and generating sequential transcript data, means for transmitting the converted text data to a natural language processing model and extracting important keywords and phrases, means for proposing topics to be discussed next based on the content of the discussion, and means for improving the system's performance using feedback. This enables real-time audio collection, automatic generation of meeting minutes on the spot, efficient and rapid progress of discussions and organization of main points, and proposals for topics to be discussed next.

[0182] "Interface means" refers to devices or software that users can access and operate.

[0183] "Means of collecting audio data in real time" refers to devices or software for instantly acquiring audio during meetings or discussions.

[0184] "Means for converting audio data into text data" refers to devices or software that convert collected audio data into textual information.

[0185] "Methods for extracting and organizing the key points of an argument" refer to techniques for finding important keywords and phrases from text data and organizing them systematically.

[0186] "Means for generating and displaying meeting minutes" refers to technology that records the content of a meeting based on extracted key points and provides it to the user.

[0187] "Methods for proposing the direction of future discussions" refers to techniques that suggest topics or themes to be discussed next, based on the current discussion.

[0188] "Means for collecting and storing feedback" refers to technologies that record user opinions and requests for later analysis and improvement.

[0189] "Communication means for acquiring audio data during a meeting" refers to technology for transmitting audio data from a meeting to a server.

[0190] A "speech recognition platform" refers to speech recognition technologies and APIs used to convert speech data into text data.

[0191] A "natural language processing model" is an artificial intelligence technology that analyzes the meaning and content of text data and extracts key points.

[0192] "A means of suggesting the next topic to be discussed" refers to a technology that analyzes the current discussion and suggests themes or topics to be addressed next to the user.

[0193] "Means for improving system performance" refer to technologies that analyze collected feedback and data to improve the overall efficiency and accuracy of the system.

[0194] This invention is a system that automatically generates meeting minutes at sites such as logistics centers, efficiently extracts the key points of the discussion, and suggests topics to be discussed next.

[0195] System Configuration

[0196] This system consists of the following main components:

[0197] User interface means: Devices and software (e.g., personal computers, mobile devices, smartphones) that allow users to access and operate a system.

[0198] Means of collecting audio data in real time: These are devices or software for instantly acquiring audio data from meetings (e.g., software or hardware that collects audio remotely using communication means).

[0199] Means of converting audio data into text data: These are devices or software that convert collected audio data into text information (e.g., Google Speech-to-Text API).

[0200] A means of extracting and organizing the key points of an argument: This refers to techniques for finding important keywords and phrases from text data and organizing them systematically (e.g., Google Natural Language API).

[0201] A means of generating and displaying meeting minutes: This technology records the content of a meeting based on extracted key points and provides it to the user.

[0202] A means of proposing the direction of future discussions: This is a technique that suggests topics or themes to be discussed next, based on the content of the current discussion.

[0203] A means of collecting and storing feedback: This is a technology for recording user opinions and requests and using them later for analysis and improvement.

[0204] System operation

[0205] 1. Starting the meeting and collecting audio data:

[0206] Users access the system through an interface and initiate a new meeting. During the meeting, audio data is collected in real time using communication methods.

[0207] 2. Converting audio to text data:

[0208] The collected audio data is sent to a speech recognition platform (e.g., Google Speech-to-Text API), where sequential transcription data is generated. The server stores the converted text data in a database.

[0209] 3. Extracting the key points of the discussion:

[0210] The server sends the stored text data to a natural language processing model (e.g., Google Natural Language API) at regular intervals to extract important keywords and phrases. The extracted key points are then organized by the server and stored as data for meeting minutes.

[0211] 4. Generating and displaying meeting minutes:

[0212] The server generates meeting minutes summarizing the discussion based on key points. The generated minutes are converted into text files or HTML format and displayed to the user through an interface.

[0213] 5. Proposed direction for the next discussion:

[0214] The server analyzes the discussion content and uses a natural language processing model to identify the next topic to discuss. The suggested topics are communicated to the user to support the progress of the discussion.

[0215] 6. Collecting and saving feedback:

[0216] Users can provide feedback on meeting minutes and proposals through the interface. This feedback is collected by the server and used to improve system performance.

[0217] Specific example

[0218] Consider a scenario where a manager at a logistics center uses this system to hold a meeting. The manager starts the meeting using their smartphone, and on-site staff participate. Audio data is collected in real time and converted into text data using the Google Speech-to-Text API. Key points are extracted using the Google Natural Language API, and meeting minutes are automatically generated and displayed. Furthermore, the system suggests topics to discuss next, allowing the manager to conduct the meeting efficiently. A feedback function collects user opinions, contributing to the improvement of the system's performance.

[0219] Example of a prompt

[0220] "We are developing an application to support logistics management meetings. We would like to implement the following functions using audio data collected during the meetings:

[0221] 1. Collect audio data in real time and convert it into text data.

[0222] 2. Extract important keywords and phrases from the converted text data and automatically generate meeting minutes.

[0223] 3. Based on the extracted data, propose the next topic to discuss.

[0224] The application will be implemented using speech recognition APIs and natural language processing APIs. We would appreciate your suggestions for this system.

[0225] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0226] Step 1:

[0227] The user initiates a new meeting using the interface. A meeting start request is sent to the server. The server generates a new meeting ID and returns it to the user. This completes the meeting preparation. The input is the user's request, and the output is the new meeting ID.

[0228] Step 2:

[0229] When the meeting begins, the terminal starts collecting audio data in real time. The collected audio data is sent to the server via a communication method. The server prepares to process the received audio data sequentially. The input is real-time audio data, and the output is the transmitted audio data.

[0230] Step 3:

[0231] The server sends audio data to a speech recognition platform (e.g., Google Speech-to-Text API) and converts it into text data. Once the audio data is converted to text data, the server stores it in a database. The input is audio data, and the output is text data.

[0232] Step 4:

[0233] The server sends the converted text data to a natural language processing model (e.g., Google Natural Language API) at regular intervals to extract important keywords and phrases. The extracted key points are organized by the server and saved for meeting minutes. The input is text data, and the output is important keywords and phrases.

[0234] Step 5:

[0235] The server automatically generates meeting minutes based on the extracted key points. The generated minutes are converted into text files or HTML format and displayed on the terminal via a user interface. Users can review the minutes in real time. Input consists of important keywords and phrases, while output is the meeting minutes data.

[0236] Step 6:

[0237] The server uses a natural language processing model to suggest the next topic to discuss, based on the generated meeting minutes data. The suggestions are then communicated to the user's terminal to support the discussion. The input is the meeting minutes data, and the output is the next topic to discuss.

[0238] Step 7:

[0239] After the meeting, users provide feedback using an interface. The server collects and stores this feedback. The collected feedback data is later used to improve the system. Input is user feedback, and output is data and suggestions for improvements regarding system performance.

[0240] 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.

[0241] This invention combines a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggests directions for future discussions, with an emotion engine that recognizes the user's emotions. By reflecting the user's emotions in real time, this system supports the progress of deeper and more appropriate discussions.

[0242] System Configuration

[0243] The system consists of the following main components:

[0244] User interface means: A device (e.g., personal computer, tablet, smartphone) that allows a user to access and operate a system.

[0245] Audio data collection method: This refers to software or hardware (e.g., ZOOM API) that collects meeting audio data in real time.

[0246] Audio data conversion means: This refers to software (e.g., Google Speech-to-Text API) for converting collected audio data into text data.

[0247] Emotion recognition means: This is an engine (e.g., emotion recognition API) that analyzes a user's emotions from voice data or text data.

[0248] Key Point Extraction Method: This is a natural language processing (NLP) model for extracting and organizing the key points of an argument from converted text data and sentiment data.

[0249] Meeting minutes generation method: This is software that generates meeting minutes based on key points and sentiment data, and displays them to the user.

[0250] A means of suggesting the direction of discussion: This is an analytical function for suggesting the next topics to be discussed.

[0251] Feedback collection method: This is software for collecting and saving user feedback and requests for improvements.

[0252] System operation

[0253] 1. User login and meeting setup

[0254] The user accesses the system from their device and enters their username and password on the login page.

[0255] The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[0256] Once logged in, users can select either "Create a new meeting" or "Join an existing meeting" on the interface.

[0257] When creating a meeting, the server generates a meeting ID and notifies the user of it. When joining an existing meeting, the user can join by entering the meeting ID.

[0258] 2. Collection of audio data and sequential transcription

[0259] The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[0260] The acquired audio data is sent to a speech recognition API and converted into text data. The server then stores the converted text data in a database.

[0261] 3. Recognition of emotions

[0262] The server sends the collected voice and text data to an emotion recognition API to analyze the user's emotions.

[0263] The emotion recognition API returns the emotional state (e.g., joy, anger, sadness) for each statement, and the server stores this information in a database.

[0264] 4. Extracting the key points of the discussion

[0265] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations.

[0266] The extracted key points and sentiment data are organized by the server and stored as data for meeting minutes.

[0267] 5. Creating and sharing meeting minutes

[0268] The server generates meeting minutes summarizing the discussion based on key points and sentiment data.

[0269] The generated meeting minutes are converted into text files or HTML format and displayed on the user's device via Zoom's chat function or a dedicated interface.

[0270] 6. Proposal for the direction of the next discussion

[0271] The server analyzes the content and sentiment data of the discussion and uses an NLP model to identify the next topic to discuss.

[0272] The proposal will be notified to the user's device, supporting the progress of the discussion.

[0273] 7. Gathering Feedback

[0274] Users can input meeting minutes and feedback on proposals through the terminal's interface.

[0275] The feedback provided will be collected and stored by the server and used to improve the system.

[0276] Specific example

[0277] For example, consider a scenario where users A, B, and C discuss the marketing strategy for a new product via Zoom. Using this system, the following operations can be performed:

[0278] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[0279] The server collects audio data from the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[0280] Simultaneously, the server uses an emotion recognition API to analyze the emotional state of each statement and obtains information such as, "User A is excited when talking about the target market."

[0281] Based on text data and sentiment data, key points such as "identifying the target market" and "competitive analysis" are extracted, and minutes of the meeting are automatically generated based on this.

[0282] The generated meeting minutes are displayed in the ZOOM chat, and all users can view them in real time.

[0283] The server proposes "setting the price of new products" as the content to be discussed next and notifies this to the terminal.

[0284] When user B inputs "want to specifically discuss the age group of the target market" as feedback, the server saves this and uses it for subsequent analysis.

[0285] In this way, the system of the present invention summarizes the key points of the discussion and the sentiment of the users in real time, and proposes the content to be discussed next, enabling an efficient and in-depth discussion.

[0286] The following describes the processing flow.

[0287] Step 1:

[0288] The user accesses the system login page from the terminal. The user enters the username and password and executes the login.

[0289] Step 2:

[0290] The server verifies the user's authentication information with the database. If the authentication is successful, a session is generated and the session ID is returned to the user.

[0291] Step 3:

[0292] The user selects "create a new meeting" or "participate in an existing meeting" through the terminal interface. If the user selects to create a new meeting, the server generates a meeting ID and notifies it to the user.

[0293] Step 4:

[0294] When joining an existing meeting, the user enters the meeting ID, and the server authenticates that ID. Once the user joins the meeting, real-time audio data collection begins.

[0295] Step 5:

[0296] The server uses the ZOOM API to acquire audio data during the meeting in real time. The collected audio data is then sent to the server.

[0297] Step 6:

[0298] The server continuously sends audio data to a speech recognition API (e.g., Google Speech-to-Text API), converting the audio data into text data in real time. The converted text data is then stored in a database.

[0299] Step 7:

[0300] The server simultaneously sends audio and text data to an emotion recognition API to analyze the emotions contained in the user's statements. The emotion recognition API returns the emotional state for each statement (e.g., joy, anger, sadness), and the server stores this in a database.

[0301] Step 8:

[0302] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations.

[0303] Step 9:

[0304] The server organizes the extracted key points and sentiment data chronologically and saves them as data for meeting minutes.

[0305] Step 10:

[0306] The server generates meeting minutes in real time based on the saved key point data and sentiment data. The generated meeting minutes are instructed to be displayed on the terminal through the ZOOM chat function or a dedicated interface. The sentiment information is also included in the meeting minutes.

[0307] Step 11:

[0308] The server analyzes the content of the discussion and the sentiment data, and then uses an NLP model to identify the topics to be discussed next. The identified topics are generated as proposals and notified to the user's terminal.

[0309] Step 12:

[0310] The user inputs feedback on the meeting minutes and proposals through the interface of the terminal.

[0311] Step 13:

[0312] The server saves the collected feedback in the database. The saved feedback is used as learning data to improve the system's natural language processing model.

[0313] (Example 2)

[0314] Next, Example 2 will be described. In the following description, the data processing device 12 is referred to as the "server", and the smart device 14 is referred to as the "terminal".

[0315] In the modern business environment, the effective operation of meetings is very important. However, in many meetings, the creation of meeting minutes is time-consuming and it is difficult to organize the key points of the discussion. In addition, it is difficult to grasp the emotions of the participants and support the progress of the discussion. As a result, there is a problem that meaningful discussions cannot be carried out and the efficiency of the meeting decreases. Furthermore, there is also a problem that the direction for the next discussion cannot be clarified.

[0316] 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.

[0317] In this invention, the server includes an interface means accessible to users, means for collecting audio data in real time, means for converting the collected audio data into text data, means for recognizing emotions from the text data and audio data, means for extracting and organizing the main points of the discussion from the text data and emotion data, means for generating and displaying meeting minutes based on the main points and emotion data, means for proposing the direction of future discussions, and means for collecting and saving feedback. This makes it possible to grasp the main points of the discussion and the emotions of the participants in real time and propose the direction of the next discussion.

[0318] An "interface means" refers to a screen or device that allows a user to access and operate a system.

[0319] "Methods for real-time collection" refers to software and hardware systems for instantly acquiring audio data from meetings.

[0320] "Methods for converting audio data to text data" refers to software that uses speech recognition technology to convert audio data into text data.

[0321] "Means of recognizing emotions" refer to engines and algorithms that analyze voice and text data to identify a user's emotional state.

[0322] "Methods for extracting and organizing the key points of an argument" refer to natural language processing technologies that automatically identify important keywords and phrases and systematically summarize the points of contention.

[0323] "Methods for generating and displaying meeting minutes" refers to software that summarizes the content of a meeting based on extracted key points and sentiment data, and displays it in document format.

[0324] "Methods for suggesting direction" refers to algorithms that automatically determine and suggest the next topic to be discussed based on the content of the discussion and sentiment data.

[0325] A "means for collecting and saving feedback" refers to a system that records user opinions and requests for corrections in a database so that they can be used later.

[0326] This invention combines a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggests directions for future discussions, with an emotion engine that recognizes the user's emotions. By reflecting the user's emotions in real time, this system supports the progress of deeper and more appropriate discussions.

[0327] System Configuration

[0328] This system consists of the following main components:

[0329] 1. Interface means: A device (e.g., personal computer, tablet, smartphone) that allows the user to access and operate the system.

[0330] 2. Means of real-time collection: This refers to software or hardware (e.g., video conferencing APIs) that collect meeting audio data in real time.

[0331] 3. Means for converting audio data to text data: This refers to software (e.g., speech recognition API) for converting collected audio data into text data.

[0332] 4. Means of recognizing emotions: This is an engine that analyzes the user's emotions from voice data or text data (e.g., emotion recognition API).

[0333] 5. Means for extracting and organizing the key points of the discussion: This is a natural language processing (NLP) model for extracting and organizing the key points of the discussion from the converted text data and sentiment data.

[0334] 6. Means for generating and displaying meeting minutes: This is software for generating meeting minutes based on key points and sentiment data, and displaying them to the user.

[0335] 7. Means for proposing direction: This is an analytical function for proposing what should be discussed next.

[0336] 8. Means for collecting and saving feedback: This is software for collecting and saving user feedback and requests for improvements.

[0337] Specific examples of system operation

[0338] For example, consider a scenario where users A, B, and C discuss a new product's marketing strategy via video conference. Using this system, the following operations and actions can be performed:

[0339] 1. Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[0340] 2. The server collects the audio data of the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[0341] 3. Simultaneously, the server uses an emotion recognition API to analyze the emotional state of each statement and obtains information such as, "User A is excited when talking about the target market."

[0342] 4. Key points such as "identifying the target market" and "competitor analysis" are extracted from text data and sentiment data, and meeting minutes are automatically generated based on this.

[0343] 5. The generated meeting minutes are immediately displayed on the user's device via the video conference chat function or a dedicated interface.

[0344] 6. The server proposes "Pricing for the new product" as the next topic to be discussed and notifies the terminal of this.

[0345] 7. If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server will save this and use it for future analyses.

[0346] Example of a prompt

[0347] Examples of prompts to input into a generative AI model include the following:

[0348] Please prepare the minutes for the new product marketing strategy meeting. The topics discussed were as follows: The target market for the new product was discussed, including target market identification and competitive analysis. User emotions such as excitement, anticipation, and concern were observed. Pricing for the new product was proposed as the next topic to be discussed.

[0349] By providing specific context in this way, AI models can generate more appropriate meeting minutes.

[0350] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0351] Step 1: User login and meeting setup

[0352] Enter: Username and password

[0353] Specific actions: The user opens a browser on their device and accesses the system's login page. There, they enter their username and password and click the "Login" button.

[0354] Data processing: The terminal sends the input information to the server. The server compares the authentication information with the database, and if there is a match, authentication is considered successful. At this point, a session ID is generated.

[0355] Output: The server returns the generated session ID to the user's terminal, and the user becomes logged in.

[0356] For example, when a user selects "Create New Meeting," the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user enters the meeting ID to complete the joining process.

[0357] Step 2: Audio data collection and sequential transcription

[0358] Input: Meeting audio data

[0359] Specific operation: When a user starts a meeting, the server collects audio data in real time using the specified API (e.g., a video conferencing API).

[0360] Data processing: The collected audio data is sent to a speech recognition API and converted into text data.

[0361] Output: The converted text data is saved to the database by the server.

[0362] For example, if the audio data is "Discussing the target market for a new product," the speech recognition API will return this as text data.

[0363] Step 3: Recognizing Emotions

[0364] Input: Converted text data and audio data

[0365] Specific operation: The server sends the collected audio and text data to the specified emotion recognition API.

[0366] Data processing: The emotion recognition API analyzes the emotional state (e.g., joy, anger, sadness, etc.) in each statement and returns the result.

[0367] Output: The server saves the emotion data returned from the emotion recognition API to the database.

[0368] Example: If the emotion recognition API analyzes that User A is "excited" while talking about the target market, that information will be saved.

[0369] Step 4: Extract the key points of the discussion

[0370] Input: Saved text data and sentiment data

[0371] Specific operation: The server sends this data to a natural language processing (NLP) model and starts the extraction process.

[0372] Data processing: The NLP model identifies and extracts important keywords and phrases.

[0373] Output: The extracted key points data, along with sentiment data, is organized by the server and saved as data for meeting minutes.

[0374] Example: The NLP model extracts key points such as "identifying the target market" and "conducting competitive analysis."

[0375] Step 5: Generate and share meeting minutes

[0376] Input: Key data and sentiment data

[0377] Specific operation: Based on this data, the server generates meeting minutes summarizing the meeting content.

[0378] Data processing: The generated meeting minutes are converted into text files or HTML format.

[0379] Output: Meeting minutes are displayed on the user's device via the video conference chat function or a dedicated interface.

[0380] Example: All users will be able to view meeting minutes in real time.

[0381] Step 6: Propose the direction of the next discussion

[0382] Input: Discussion content and sentiment data

[0383] Specific operation: The server analyzes this data and uses an NLP model to identify the next topic to discuss.

[0384] Data processing: Determine the topic proposed by the NLP model.

[0385] Output: The proposed topic information will be notified to the user's terminal.

[0386] Example: The server suggests "Pricing for the new product" as the next discussion topic.

[0387] Step 7: Gathering Feedback

[0388] Input: User feedback information

[0389] Specific operation: Users input feedback on meeting minutes and proposals using the terminal's interface.

[0390] Data processing: The provided feedback information is collected and recorded by the server.

[0391] Output: Feedback information is stored in a database and used for future system improvements.

[0392] For example, if User B provides feedback such as "I would like to discuss the age range of the target market in more detail," that information will be saved.

[0393] (Application Example 2)

[0394] 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".

[0395] Traditional methods of recording meetings and discussions often involve manual transcription of audio data and extraction of key points, which is time-consuming and labor-intensive. Furthermore, it is difficult to consider the emotional state of participants during discussions, and there is a lack of means to support the proper progression of the discussion, resulting in inconsistent quality of meeting minutes and insufficient suggestions for the direction of future discussions.

[0396] 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.

[0397] In this invention, the server includes an interface means accessible to the user, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for proposing the direction of future discussions, means for collecting and saving feedback, emotion recognition means for recognizing and displaying the user's emotions in real time, means for identifying the content to be discussed next based on the main points and emotion data, and means for sharing the generated meeting minutes. This makes it possible to efficiently extract and organize the main points of the discussion, automatically generate high-quality meeting minutes, and propose the direction of the next discussion. Furthermore, by reflecting the user's emotional state in real time, it is possible to support the progress of discussions more deeply and appropriately.

[0398] "Interface means" refers to a device or software that allows a user to access and operate a system.

[0399] "Means for collecting audio data in real time" refers to a system component that acquires audio data from meetings and discussions and transmits it to a server in real time.

[0400] "Means for converting audio data to text data" refers to software or algorithms that transcribe collected audio data and convert it into text format.

[0401] "Methods for extracting and organizing the main points of a discussion from text data" refers to system components that analyze converted text data, extract important keywords and phrases, and organize the main points of the discussion.

[0402] "A means of generating and displaying meeting minutes based on key points" refers to a system component that automatically creates meeting minutes based on extracted key points and displays them in an easy-to-understand manner for the user.

[0403] "Means for proposing the direction of future discussions" refers to a system component that analyzes the content of discussions and sentiment data to identify and propose topics that should be discussed next.

[0404] "Means for collecting and storing feedback" refers to system components for collecting feedback information such as opinions and requests for corrections from users and storing it in a database.

[0405] An "emotion recognition tool" is an engine or software that analyzes and recognizes emotions in real time from a user's voice or text data.

[0406] "Means for sharing generated meeting minutes" refers to a system component that provides the functionality to share generated meeting minutes with other users or systems.

[0407] This invention combines a system that collects audio data from meetings and discussions in real time, transcribes it sequentially to extract and organize the key points of the discussion, and automatically generates meeting minutes with an emotion engine that recognizes the user's emotions in real time. This system is composed of the following various means.

[0408] Hardware and software to be used

[0409] User interface means: A smartphone, tablet, or personal computer is used as a device for the user to access and operate the system.

[0410] Audio data collection method: To collect meeting audio data in real time, for example, use the ZOOM API.

[0411] Audio data conversion method: The Google Speech-to-Text API is used to convert the collected audio data into text data.

[0412] Emotion recognition method: To analyze user emotions from voice and text data, an emotion recognition API (e.g., Microsoft® Azure® or AWS® Rekognition) is used.

[0413] Key points of the discussion: Natural language processing (NLP) models (e.g., BERT or GPT-3®) are used to extract and organize the key points of the discussion from text data and sentiment data.

[0414] Meeting minutes generation method: Software is used to generate meeting minutes based on key points and sentiment data, and to display them to the user.

[0415] Method for suggesting the direction of discussion: Utilize software or algorithms with analytical capabilities to suggest the next topics to be discussed.

[0416] Feedback collection method: Use an interface and database to collect and store user feedback and correction requests.

[0417] Meeting minutes sharing method: It has software functionality to allow generated meeting minutes to be shared with other users or systems.

[0418] Specific examples of how the system works

[0419] 1. Login and meeting setup:

[0420] Users access the system from devices such as smartphones or computers and enter their username and password on the login page. The server verifies the entered authentication information against the database, and if authentication is successful, it creates a session and returns a session ID to the user. After logging in, the user selects either "Create a new meeting" or "Join an existing meeting," and the server generates or authenticates a meeting ID.

[0421] 2. Collection and transcription of audio data:

[0422] The server uses the ZOOM API to acquire audio data from meetings in real time. The acquired audio data is converted into text data using the Google Speech-to-Text API, and the text data is stored in a database by the server.

[0423] 3. Recognition of emotions:

[0424] The server sends the collected audio and text data to an emotion recognition API to analyze the user's emotions. The emotion recognition API returns the emotional state for each statement (e.g., joy, anger, sadness), and the server stores this in a database.

[0425] 4. Extracting the main points of the discussion:

[0426] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations. The extracted key points and sentiment data are then organized by the server and stored as data for generating meeting minutes.

[0427] 5. Creating and sharing meeting minutes:

[0428] The server generates meeting minutes summarizing the discussion based on key points and sentiment data. The generated minutes are converted into text files or HTML format and displayed on the user's device, as well as shared with other users through the sharing function.

[0429] 6. Proposed direction for the discussion:

[0430] The server analyzes the discussion content and sentiment data, and uses an NLP model to identify the next topic to discuss. Suggestions are notified to the user's terminal to support the progress of the discussion.

[0431] 7. Gathering feedback:

[0432] Users can input feedback on meeting minutes and proposals through the interface. The feedback provided is collected and stored by the server and used to improve the system.

[0433] Specific examples and prompt statements

[0434] For example, consider a scenario where store staff hold a meeting about how to display new products. Using this system, the following operations can be performed:

[0435] Staff members A and B log into the system from their respective devices, enter the meeting ID, and join the meeting.

[0436] The server collects audio data from the meeting and converts statements like "We will discuss how to display the new products" into text in real time.

[0437] At the same time, the server uses an emotion recognition API to analyze the emotional state of each statement, obtaining information such as, "Staff member A is excited when talking about the display method."

[0438] The system extracts key points such as "areas for improvement in display methods" and "customer reactions" from text data and sentiment data, and then automatically generates meeting minutes based on this information.

[0439] The generated meeting minutes can be shared with other staff members using the sharing function, allowing them to review them immediately.

[0440] The server suggests "Pricing for the new product" as the next topic to discuss and notifies the staff member's terminal.

[0441] Examples of specific prompt messages:

[0442] "Customer complaint about product quality."

[0443] "Next discussion topic: Improve customer complaint handling."

[0444] This makes it possible to efficiently extract key points of discussions, including user emotions, based on the form in which the invention is implemented, and to generate high-quality meeting minutes.

[0445] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0446] Step 1:

[0447] User login and meeting setup

[0448] Input: Users access the system from a device such as a smartphone or computer and enter their username and password.

[0449] Data processing / calculation: The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID.

[0450] Output: Upon successful authentication, the session ID will be displayed on the user's device.

[0451] Specific operation: The user interacts with the interface and selects either "Create a new meeting" or "Join an existing meeting." If "Create a new meeting" is selected, the server generates a meeting ID and notifies the user.

[0452] Step 2:

[0453] Audio data collection and sequential transcription

[0454] Input: The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[0455] Data processing / calculation: The acquired audio data is sent to a speech recognition API (Google Speech-to-Text API) and converted into text data.

[0456] Output: The converted text data is saved on the server.

[0457] Specific operation: The server updates text data in real time and displays it to the user on the interface.

[0458] Step 3:

[0459] Recognition of emotions

[0460] Input: The server sends collected audio and text data to the emotion recognition API.

[0461] Data Processing / Calculation: The emotion recognition API analyzes the emotional state (e.g., joy, anger, sadness, etc.) in each statement and returns it as emotion data.

[0462] Output: Emotional data is stored on the server.

[0463] Specific operation: The server displays an emotion indicator corresponding to each statement on the interface, allowing the user to visually confirm it.

[0464] Step 4:

[0465] Extracting the main points of the discussion

[0466] Input: The server sends stored text data and sentiment data to a natural language processing (NLP) model.

[0467] Data processing / calculation: The NLP model extracts important keywords and phrases and further analyzes emotional fluctuations.

[0468] Output: Extracted key points and sentiment data are stored on the server.

[0469] Specific operation: The server organizes the key points and displays them in real time on the interface as data for generating meeting minutes.

[0470] Step 5:

[0471] Creating and sharing meeting minutes

[0472] Input: The server summarizes the discussion based on key points and sentiment data.

[0473] Data processing / calculation: Automatically create meeting minutes using a meeting minutes generation algorithm.

[0474] Output: The generated meeting minutes are converted to a text file or HTML format.

[0475] Specific operation: Meeting minutes are displayed to the user on the interface and can be sent to other users via the sharing function.

[0476] Step 6:

[0477] Proposal for the direction of the discussion

[0478] Input: The server analyzes the content of the discussion and sentiment data.

[0479] Data processing / computation: Use an NLP model to identify the next topic to discuss.

[0480] Output: The proposed content is notified to the user's device.

[0481] Specific behavior: The user interface displays specific suggestions for the next topic to be discussed, such as "Pricing for the new product."

[0482] Step 7:

[0483] Gathering feedback

[0484] Input: The user provides feedback through the interface.

[0485] Data processing / calculation: Collect and store feedback information in a database.

[0486] Output: The collected feedback information will be used to improve the system in the future.

[0487] Specific operation: Based on user feedback, the system automatically adjusts areas for improvement and discussion topics for the next meeting.

[0488] 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.

[0489] 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.

[0490] 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.

[0491] [Second Embodiment]

[0492] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0493] 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.

[0494] 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).

[0495] 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.

[0496] 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.

[0497] 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).

[0498] 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.

[0499] 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.

[0500] 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.

[0501] 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.

[0502] 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.

[0503] 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".

[0504] This invention is a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggesting directions for future discussions. This system is realized through the cooperation of the user, server, and terminal.

[0505] System Configuration

[0506] The system consists of the following main components:

[0507] User interface means: A device (e.g., personal computer, tablet, smartphone) that allows a user to access and operate a system.

[0508] Audio data collection method: This refers to software or hardware (e.g., ZOOM API) that collects meeting audio data in real time.

[0509] Audio data conversion means: This refers to software (e.g., Google Speech-to-Text API) for converting collected audio data into text data.

[0510] Key Point Extraction Method: This is a natural language processing (NLP) model for extracting and organizing the key points of an argument from converted text data.

[0511] Meeting minutes generation method: This is software that generates meeting minutes based on key points and displays them to the user.

[0512] A means of suggesting the direction of discussion: This is an analytical function for suggesting the next topics to be discussed.

[0513] Feedback collection method: This is software for collecting and saving user feedback and requests for improvements.

[0514] System operation

[0515] 1. User login and meeting setup

[0516] The user accesses the system from their device and enters their username and password on the login page.

[0517] The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[0518] Once logged in, users can select either "Create a new meeting" or "Join an existing meeting" on the interface.

[0519] When creating a meeting, the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user can join by entering the meeting ID.

[0520] 2. Collection of audio data and sequential transcription

[0521] The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[0522] The acquired audio data is sent to a speech recognition API and converted into text data. The server then sequentially saves the converted text data to a database.

[0523] 3. Extracting the key points of the discussion

[0524] The server sends the stored text data to a natural language processing (NLP) model at regular intervals to extract important keywords and phrases.

[0525] The extracted key points are organized by the server and saved as data for meeting minutes.

[0526] 4. Creating and sharing meeting minutes

[0527] The server generates meeting minutes summarizing the discussion based on the key points.

[0528] The generated meeting minutes are converted into text files or HTML format and displayed on the user's device via Zoom's chat function or a dedicated interface.

[0529] 5. Proposal for the direction of the next discussion

[0530] The server analyzes the content of the discussion and uses an NLP model to identify the next topic to discuss.

[0531] The proposal will be notified to the user's device, supporting the progress of the discussion.

[0532] 6. Gathering Feedback

[0533] Users can provide feedback on meeting minutes and proposals from their devices.

[0534] The feedback provided will be collected and stored by the server and used to improve the system.

[0535] Specific example

[0536] For example, consider a scenario where users A, B, and C discuss the marketing strategy for a new product via Zoom. Using this system, the following operations can be performed:

[0537] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[0538] The server collects audio data from the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[0539] The system extracts key points such as "identifying the target market" and "competitor analysis" from text data, and then automatically generates meeting minutes based on this information.

[0540] The generated meeting minutes are displayed in the Zoom chat, allowing all users to view them in real time.

[0541] The server proposes "Pricing for the new product" as the next topic to discuss and notifies the terminal of this.

[0542] If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server saves this and uses it for future analyses.

[0543] In this way, the system of the present invention enables efficient and in-depth discussions by summarizing the key points of the discussion in real time and suggesting the next topics to be discussed.

[0544] The following describes the processing flow.

[0545] Step 1:

[0546] The user accesses the system's login page from their terminal. The user enters their username and password and completes the login process.

[0547] Step 2:

[0548] The server compares the user's authentication information with the database, and if correct, creates a session and returns a session ID to the user.

[0549] Step 3:

[0550] The user selects either "Create a new meeting" or "Join an existing meeting" on the terminal interface. If the user selects "Create a new meeting," the server generates a meeting ID and notifies the user of it.

[0551] Step 4:

[0552] When joining an existing meeting, the user enters the meeting ID, and the server authenticates that ID. Once the user joins the meeting, real-time audio data collection begins.

[0553] Step 5:

[0554] The server uses the ZOOM API to acquire audio data during the meeting in real time. The collected audio data is then sent to the server.

[0555] Step 6:

[0556] The server continuously sends audio data to a speech recognition API (e.g., Google Speech-to-Text API), converting the audio data into text data in real time. The converted text data is then stored in a database.

[0557] Step 7:

[0558] The server sends the stored text data to a natural language processing (NLP) model at regular intervals to extract important keywords and phrases.

[0559] Step 8:

[0560] The server organizes the extracted key points chronologically and saves them as data for meeting minutes.

[0561] Step 9:

[0562] The server generates meeting minutes in real time based on the stored key points data. The generated minutes are then instructed to be displayed on the device via Zoom's chat function or a dedicated interface.

[0563] Step 10:

[0564] The server analyzes the discussion and uses an NLP model to identify the next topic to discuss. The identified topic is generated as a suggestion and notified to the user's terminal.

[0565] Step 11:

[0566] Users input feedback on meeting minutes and proposals through the terminal's interface.

[0567] Step 12:

[0568] The server stores the collected feedback in a database. This stored feedback is then used as training data to improve the system's natural language processing model.

[0569] (Example 1)

[0570] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0571] In traditional meetings and discussions, creating meeting minutes was cumbersome, making it difficult to summarize key points of important discussions in real time. Furthermore, it was impossible to suggest future directions for discussions, and it was difficult to effectively collect participant feedback and use it for improvement. This resulted in decreased meeting efficiency and the risk of important discussions being overlooked.

[0572] 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.

[0573] In this invention, the server includes an interface means accessible to users, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for analyzing the content of the discussion and suggesting what should be discussed next, means for collecting and saving user feedback, means for using a communication service application programming interface for collecting audio data, means for using a speech recognition application programming interface to convert the audio data into text data, and means for using a natural language processing model for extracting main points. This makes it possible to automatically generate meeting minutes in real time, support the progress of meetings by suggesting what should be discussed next, and effectively collect feedback from participants to improve the system.

[0574] "Interface means" refers to devices and software that allow users to access and operate a system.

[0575] "Audio data" refers to the digital data of audio signals collected during meetings or discussions.

[0576] "Means of real-time collection" refers to equipment and software for instantly collecting audio data during a meeting.

[0577] "Text data" refers to character data generated by transcribing audio data.

[0578] "Methods for extracting and organizing the key points of an argument" refers to natural language processing models and algorithms that find important information from text data and organize it systematically.

[0579] "Means for generating and displaying meeting minutes" refers to software that automatically creates meeting minutes based on extracted key points and presents them to the user.

[0580] "Means of proposing a direction for discussion" refers to a function that analyzes the content of the current discussion and indicates the points that should be discussed next.

[0581] "Means for collecting and saving feedback" refers to a system for collecting user opinions and correction requests and saving them for later use.

[0582] "Application programming interface for communication services" refers to the means of utilizing functions through the programming interface provided by communication services.

[0583] A "Speech Recognition Application Programming Interface" refers to a programming interface that provides speech recognition functionality for converting speech data into text.

[0584] A "natural language processing model" refers to machine learning models and algorithms used to understand and analyze human language.

[0585] The system of this invention collects audio data in real time during meetings and discussions and transcribes it sequentially. It then extracts and organizes the key points of the discussion to automatically generate meeting minutes. Furthermore, it proposes directions for future discussions and collects and stores user feedback. This system is realized through the cooperation of the user, server, and terminal.

[0586] Key components of the system

[0587] Interface means: A device that allows a user to access and operate a system (e.g., a personal computer, tablet, or smartphone).

[0588] Audio data collection means: A communication service application programming interface (e.g., a communication service API) for collecting conference audio data in real time.

[0589] Audio data conversion means: A speech recognition application programming interface (e.g., speech recognition API) for converting collected audio data into text data.

[0590] Key point extraction method: A natural language processing model for extracting and organizing the key points of a discussion from text data.

[0591] Meeting minutes generation method: Software for generating and displaying meeting minutes based on key points.

[0592] A means of suggesting the direction of discussion: A function for analyzing the content of a discussion and suggesting what should be discussed next.

[0593] Feedback collection method: Software for collecting and storing user feedback and correction requests.

[0594] System operation

[0595] The user accesses the system from their terminal and enters their username and password on the login page. The server verifies the authentication information against the database, and if successful, creates a session and returns a session ID to the user. After logging in, the user selects either "Create a new meeting" or "Join an existing meeting." The server generates a meeting ID and notifies the user of it.

[0596] During the meeting, the server uses a communication service API to acquire audio data in real time. The acquired audio data is sent to a speech recognition API and converted into text data. The server then sequentially saves the converted text data to a database.

[0597] The server sends the stored text data to a natural language processing model at regular intervals to extract important keywords and phrases. The extracted key points are then organized by the server and stored as data for meeting minutes.

[0598] The server generates meeting minutes summarizing the discussion based on key points. The generated minutes are converted into text files or HTML format and displayed on the user's device through the communication service's chat function or a dedicated interface.

[0599] Furthermore, the server analyzes the content of the discussion and uses a natural language processing model to identify the next topic to discuss. This suggestion is then communicated to the user's terminal to support the progress of the discussion.

[0600] Users can provide feedback on meeting minutes and proposals from their devices. The feedback provided is collected and stored by the server and used to improve the system.

[0601] Specific example

[0602] For example, consider a scenario where users A, B, and C discuss the marketing strategy for a new product through a communication service. Using this system, the following operations can be performed:

[0603] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[0604] The server collects audio data from the meeting via a communication service API and converts statements such as "We will discuss the target market for the new product" into text in real time.

[0605] The system extracts key points such as "identifying the target market" and "competitor analysis" from text data, and then automatically generates meeting minutes based on this information.

[0606] The generated meeting minutes are displayed in the communication service's chat, allowing all users to view them in real time.

[0607] The server proposes "Pricing for the new product" as the next topic to discuss and notifies the terminal of this.

[0608] If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server saves this and uses it for future analyses.

[0609] Example of a prompt

[0610] "We've developed a system that collects meeting audio data in real time and transcribes it sequentially. It also has a function to identify and suggest the next topic of discussion. For example, if we're discussing the marketing strategy for a new product, and topics like the target market and competitor analysis come up, how can you tell me how to suggest the next topic to discuss?"

[0611] In this way, the system of the present invention enables efficient and in-depth discussions by summarizing the key points of the discussion in real time and suggesting the next topics to be discussed.

[0612] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0613] Step 1: User login and meeting setup

[0614] Specific operation: The user accesses the system from their terminal and enters their username and password on the login page.

[0615] Enter: Username and password.

[0616] Data processing / calculation: The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[0617] Output: Session ID.

[0618] After the user logs in, they can select either "Create a New Meeting" or "Join an Existing Meeting" on the interface. If a meeting is created, the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user enters the meeting ID.

[0619] Step 2: Audio data collection and sequential transcription

[0620] Specific operation: The server uses a communication service API to acquire audio data during the meeting in real time.

[0621] Input: Voice data obtained from a communication service API.

[0622] Data processing / calculation: The server sends the acquired audio data to a speech recognition API and converts it into text data. The server then sequentially saves the converted text data to a database.

[0623] Output: Text data.

[0624] Step 3: Extract the key points of the discussion

[0625] Specific operation: The server sends the stored text data to a natural language processing model at regular intervals, and extracts important keywords and phrases.

[0626] Input: Text data.

[0627] Data Processing / Calculation: A natural language processing model analyzes the text data and extracts important keywords and phrases. The server organizes the extracted key points and saves them as data for meeting minutes.

[0628] Output: Key data.

[0629] Step 4: Generate and share meeting minutes

[0630] Specific operation: The server generates meeting minutes summarizing the discussion based on key points.

[0631] Input: Key data.

[0632] Data Processing / Calculation: Automatically creates meeting minutes based on key data and converts them to text files or HTML format.

[0633] Output: Meeting minutes file (text file / HTML format).

[0634] The generated meeting minutes are displayed on the user's device through the communication service's chat function or a dedicated interface.

[0635] Step 5: Propose the direction of the next discussion

[0636] Specific operation: The server analyzes the content of the discussion and uses a natural language processing model to identify the next topic to discuss.

[0637] Input: Meeting minutes data.

[0638] Data processing / computation: Using a natural language processing model, identify the next topic to discuss. The server notifies the user's terminal of the identified topic as a suggestion.

[0639] Output: Suggestions for the next topic to discuss.

[0640] Step 6: Gathering Feedback

[0641] Specific functionality: Users can provide feedback on meeting minutes and proposals from their devices.

[0642] Input: User feedback.

[0643] Data processing / calculation: The server collects and stores user feedback, which is later used to improve the system.

[0644] Output: Saved feedback data.

[0645] In this way, a series of processes can be realized in which the server, terminal, and user cooperate with each other to summarize the key points of the discussion in real time and propose the next topics to be discussed.

[0646] (Application Example 1)

[0647] 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."

[0648] In modern meetings and discussions, creating meeting minutes, summarizing key points, and proposing directions for future discussions are crucial, yet extremely time-consuming. Therefore, there is a need for a system that collects audio data in real time, automatically generates meeting minutes, and efficiently proposes directions for future discussions. Furthermore, in environments requiring rapid decision-making, such as logistics centers, the implementation of an efficient meeting management system is urgently needed.

[0649] 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.

[0650] In this invention, the server includes an interface means accessible to the user, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for proposing the direction of future discussions, means for collecting and saving feedback, communication means for acquiring audio data during a meeting, means for transmitting the acquired audio data to a speech recognition platform and generating sequential transcript data, means for transmitting the converted text data to a natural language processing model and extracting important keywords and phrases, means for proposing topics to be discussed next based on the content of the discussion, and means for improving the system's performance using feedback. This enables real-time audio collection, automatic generation of meeting minutes on the spot, efficient and rapid progress of discussions and organization of main points, and proposals for topics to be discussed next.

[0651] "Interface means" refers to devices or software that users can access and operate.

[0652] "Means of collecting audio data in real time" refers to devices or software for instantly acquiring audio during meetings or discussions.

[0653] "Means for converting audio data into text data" refers to devices or software that convert collected audio data into textual information.

[0654] "Methods for extracting and organizing the key points of an argument" refer to techniques for finding important keywords and phrases from text data and organizing them systematically.

[0655] "Means for generating and displaying meeting minutes" refers to technology that records the content of a meeting based on extracted key points and provides it to the user.

[0656] "Methods for proposing the direction of future discussions" refers to techniques that suggest topics or themes to be discussed next, based on the current discussion.

[0657] "Means for collecting and storing feedback" refers to technologies that record user opinions and requests for later analysis and improvement.

[0658] "Communication means for acquiring audio data during a meeting" refers to technology for transmitting audio data from a meeting to a server.

[0659] A "speech recognition platform" refers to speech recognition technologies and APIs used to convert speech data into text data.

[0660] A "natural language processing model" is an artificial intelligence technology that analyzes the meaning and content of text data and extracts key points.

[0661] "A means of suggesting the next topic to be discussed" refers to a technology that analyzes the current discussion and suggests themes or topics to be addressed next to the user.

[0662] "Means for improving system performance" refer to technologies that analyze collected feedback and data to improve the overall efficiency and accuracy of the system.

[0663] This invention is a system that automatically generates meeting minutes at sites such as logistics centers, efficiently extracts the key points of the discussion, and suggests topics to be discussed next.

[0664] System Configuration

[0665] This system consists of the following main components:

[0666] User interface means: Devices and software (e.g., personal computers, mobile devices, smartphones) that allow users to access and operate a system.

[0667] Means of collecting audio data in real time: These are devices or software for instantly acquiring audio data from meetings (e.g., software or hardware that collects audio remotely using communication means).

[0668] Means of converting audio data into text data: These are devices or software that convert collected audio data into text information (e.g., Google Speech-to-Text API).

[0669] A means of extracting and organizing the key points of an argument: This refers to techniques for finding important keywords and phrases from text data and organizing them systematically (e.g., Google Natural Language API).

[0670] A means of generating and displaying meeting minutes: This technology records the content of a meeting based on extracted key points and provides it to the user.

[0671] A means of proposing the direction of future discussions: This is a technique that suggests topics or themes to be discussed next, based on the content of the current discussion.

[0672] A means of collecting and storing feedback: This is a technology for recording user opinions and requests and using them later for analysis and improvement.

[0673] System operation

[0674] 1. Starting the meeting and collecting audio data:

[0675] Users access the system through an interface and initiate a new meeting. During the meeting, audio data is collected in real time using communication methods.

[0676] 2. Converting audio to text data:

[0677] The collected audio data is sent to a speech recognition platform (e.g., Google Speech-to-Text API), where sequential transcription data is generated. The server stores the converted text data in a database.

[0678] 3. Extracting the key points of the discussion:

[0679] The server sends the stored text data to a natural language processing model (e.g., Google Natural Language API) at regular intervals to extract important keywords and phrases. The extracted key points are then organized by the server and stored as data for meeting minutes.

[0680] 4. Generating and displaying meeting minutes:

[0681] The server generates meeting minutes summarizing the discussion based on key points. The generated minutes are converted into text files or HTML format and displayed to the user through an interface.

[0682] 5. Proposed direction for the next discussion:

[0683] The server analyzes the discussion content and uses a natural language processing model to identify the next topic to discuss. The suggested topics are communicated to the user to support the progress of the discussion.

[0684] 6. Collecting and saving feedback:

[0685] Users can provide feedback on meeting minutes and proposals through the interface. This feedback is collected by the server and used to improve system performance.

[0686] Specific example

[0687] Consider a scenario where a manager at a logistics center uses this system to hold a meeting. The manager starts the meeting using their smartphone, and on-site staff participate. Audio data is collected in real time and converted into text data using the Google Speech-to-Text API. Key points are extracted using the Google Natural Language API, and meeting minutes are automatically generated and displayed. Furthermore, the system suggests topics to discuss next, allowing the manager to conduct the meeting efficiently. A feedback function collects user opinions, contributing to the improvement of the system's performance.

[0688] Example of a prompt

[0689] "We are developing an application to support logistics management meetings. We would like to implement the following functions using audio data collected during the meetings:

[0690] 1. Collect audio data in real time and convert it into text data.

[0691] 2. Extract important keywords and phrases from the converted text data and automatically generate meeting minutes.

[0692] 3. Based on the extracted data, propose the next topic to discuss.

[0693] The application will be implemented using speech recognition APIs and natural language processing APIs. We would appreciate your suggestions for this system.

[0694] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0695] Step 1:

[0696] The user initiates a new meeting using the interface. A meeting start request is sent to the server. The server generates a new meeting ID and returns it to the user. This completes the meeting preparation. The input is the user's request, and the output is the new meeting ID.

[0697] Step 2:

[0698] When the meeting begins, the terminal starts collecting audio data in real time. The collected audio data is sent to the server via a communication method. The server prepares to process the received audio data sequentially. The input is real-time audio data, and the output is the transmitted audio data.

[0699] Step 3:

[0700] The server sends audio data to a speech recognition platform (e.g., Google Speech-to-Text API) and converts it into text data. Once the audio data is converted to text data, the server stores it in a database. The input is audio data, and the output is text data.

[0701] Step 4:

[0702] The server sends the converted text data to a natural language processing model (e.g., Google Natural Language API) at regular intervals to extract important keywords and phrases. The extracted key points are organized by the server and saved for meeting minutes. The input is text data, and the output is important keywords and phrases.

[0703] Step 5:

[0704] The server automatically generates meeting minutes based on the extracted key points. The generated minutes are converted into text files or HTML format and displayed on the terminal via a user interface. Users can review the minutes in real time. Input consists of important keywords and phrases, while output is the meeting minutes data.

[0705] Step 6:

[0706] The server uses a natural language processing model to suggest the next topic to discuss, based on the generated meeting minutes data. The suggestions are then communicated to the user's terminal to support the discussion. The input is the meeting minutes data, and the output is the next topic to discuss.

[0707] Step 7:

[0708] After the meeting, users provide feedback using an interface. The server collects and stores this feedback. The collected feedback data is later used to improve the system. Input is user feedback, and output is data and suggestions for improvements regarding system performance.

[0709] 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.

[0710] This invention combines a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggests directions for future discussions, with an emotion engine that recognizes the user's emotions. By reflecting the user's emotions in real time, this system supports the progress of deeper and more appropriate discussions.

[0711] System Configuration

[0712] The system consists of the following main components:

[0713] User interface means: A device (e.g., personal computer, tablet, smartphone) that allows a user to access and operate a system.

[0714] Audio data collection method: This refers to software or hardware (e.g., ZOOM API) that collects meeting audio data in real time.

[0715] Audio data conversion means: This refers to software (e.g., Google Speech-to-Text API) for converting collected audio data into text data.

[0716] Emotion recognition means: This is an engine (e.g., emotion recognition API) that analyzes a user's emotions from voice data or text data.

[0717] Key Point Extraction Method: This is a natural language processing (NLP) model for extracting and organizing the key points of an argument from converted text data and sentiment data.

[0718] Meeting minutes generation method: This is software that generates meeting minutes based on key points and sentiment data, and displays them to the user.

[0719] A means of suggesting the direction of discussion: This is an analytical function for suggesting the next topics to be discussed.

[0720] Feedback collection method: This is software for collecting and saving user feedback and requests for improvements.

[0721] System operation

[0722] 1. User login and meeting setup

[0723] The user accesses the system from their device and enters their username and password on the login page.

[0724] The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[0725] Once logged in, users can select either "Create a new meeting" or "Join an existing meeting" on the interface.

[0726] When creating a meeting, the server generates a meeting ID and notifies the user of it. When joining an existing meeting, the user can join by entering the meeting ID.

[0727] 2. Collection of audio data and sequential transcription

[0728] The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[0729] The acquired audio data is sent to a speech recognition API and converted into text data. The server then stores the converted text data in a database.

[0730] 3. Recognition of emotions

[0731] The server sends the collected voice and text data to an emotion recognition API to analyze the user's emotions.

[0732] The emotion recognition API returns the emotional state (e.g., joy, anger, sadness) for each statement, and the server stores this information in a database.

[0733] 4. Extracting the key points of the discussion

[0734] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations.

[0735] The extracted key points and sentiment data are organized by the server and stored as data for meeting minutes.

[0736] 5. Creating and sharing meeting minutes

[0737] The server generates meeting minutes summarizing the discussion based on key points and sentiment data.

[0738] The generated meeting minutes are converted into text files or HTML format and displayed on the user's device via Zoom's chat function or a dedicated interface.

[0739] 6. Proposal for the direction of the next discussion

[0740] The server analyzes the content and sentiment data of the discussion and uses an NLP model to identify the next topic to discuss.

[0741] The proposal will be notified to the user's device, supporting the progress of the discussion.

[0742] 7. Gathering Feedback

[0743] Users can input meeting minutes and feedback on proposals through the terminal's interface.

[0744] The feedback provided will be collected and stored by the server and used to improve the system.

[0745] Specific example

[0746] For example, consider a scenario where users A, B, and C discuss the marketing strategy for a new product via Zoom. Using this system, the following operations can be performed:

[0747] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[0748] The server collects audio data from the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[0749] Simultaneously, the server uses an emotion recognition API to analyze the emotional state of each statement and obtains information such as, "User A is excited when talking about the target market."

[0750] Key points such as "identifying the target market" and "competitor analysis" are extracted from text data and sentiment data, and meeting minutes are automatically generated based on this information.

[0751] The generated meeting minutes are displayed in the Zoom chat, allowing all users to view them in real time.

[0752] The server proposes "Pricing for the new product" as the next topic to discuss and notifies the terminal of this.

[0753] If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server saves this and uses it for future analyses.

[0754] In this way, the system of the present invention enables efficient and in-depth discussions by summarizing the key points of the discussion and the users' feelings in real time and suggesting the next topics to discuss.

[0755] The following describes the processing flow.

[0756] Step 1:

[0757] The user accesses the system's login page from their terminal. The user enters their username and password and completes the login process.

[0758] Step 2:

[0759] The server compares the user's authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[0760] Step 3:

[0761] The user selects either "Create a new meeting" or "Join an existing meeting" on the terminal interface. If the user selects "Create a new meeting," the server generates a meeting ID and notifies the user of it.

[0762] Step 4:

[0763] When joining an existing meeting, the user enters the meeting ID, and the server authenticates that ID. Once the user joins the meeting, real-time audio data collection begins.

[0764] Step 5:

[0765] The server uses the ZOOM API to acquire audio data during the meeting in real time. The collected audio data is then sent to the server.

[0766] Step 6:

[0767] The server continuously sends audio data to a speech recognition API (e.g., Google Speech-to-Text API), converting the audio data into text data in real time. The converted text data is then stored in a database.

[0768] Step 7:

[0769] The server simultaneously sends audio and text data to an emotion recognition API to analyze the emotions contained in the user's statements. The emotion recognition API returns the emotional state for each statement (e.g., joy, anger, sadness), and the server stores this in a database.

[0770] Step 8:

[0771] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations.

[0772] Step 9:

[0773] The server organizes the extracted key points and sentiment data chronologically and saves them as data for meeting minutes.

[0774] Step 10:

[0775] The server generates meeting minutes in real time based on stored key points and sentiment data. The generated minutes are then instructed to be displayed on the device via Zoom's chat function or a dedicated interface. Sentiment information is also included in the meeting minutes.

[0776] Step 11:

[0777] The server analyzes the content and sentiment data of the discussion and uses an NLP model to identify the next topic to discuss. The identified topic is generated as a suggestion and notified to the user's terminal.

[0778] Step 12:

[0779] Users input feedback on meeting minutes and proposals through the terminal's interface.

[0780] Step 13:

[0781] The server stores the collected feedback in a database. This stored feedback is then used as training data to improve the system's natural language processing model.

[0782] (Example 2)

[0783] 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".

[0784] In today's business environment, effective meeting management is crucial. However, in many meetings, creating meeting minutes is time-consuming, making it difficult to summarize the key points of the discussion. Furthermore, it's challenging to understand participants' emotions and support the flow of the discussion. As a result, meaningful discussions fail, and meeting efficiency suffers. Additionally, there's the challenge of not being able to clearly define the direction the discussion should take next.

[0785] 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.

[0786] In this invention, the server includes an interface means accessible to users, means for collecting audio data in real time, means for converting the collected audio data into text data, means for recognizing emotions from the text data and audio data, means for extracting and organizing the main points of the discussion from the text data and emotion data, means for generating and displaying meeting minutes based on the main points and emotion data, means for proposing the direction of future discussions, and means for collecting and saving feedback. This makes it possible to grasp the main points of the discussion and the emotions of the participants in real time and propose the direction of the next discussion.

[0787] An "interface means" refers to a screen or device that allows a user to access and operate a system.

[0788] "Methods for real-time collection" refers to software and hardware systems for instantly acquiring audio data from meetings.

[0789] "Methods for converting audio data to text data" refers to software that uses speech recognition technology to convert audio data into text data.

[0790] "Means of recognizing emotions" refer to engines and algorithms that analyze voice and text data to identify a user's emotional state.

[0791] "Methods for extracting and organizing the key points of an argument" refer to natural language processing technologies that automatically identify important keywords and phrases and systematically summarize the points of contention.

[0792] "Methods for generating and displaying meeting minutes" refers to software that summarizes the content of a meeting based on extracted key points and sentiment data, and displays it in document format.

[0793] "Methods for suggesting direction" refers to algorithms that automatically determine and suggest the next topic to be discussed based on the content of the discussion and sentiment data.

[0794] A "means for collecting and saving feedback" refers to a system that records user opinions and requests for corrections in a database so that they can be used later.

[0795] This invention combines a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggests directions for future discussions, with an emotion engine that recognizes the user's emotions. By reflecting the user's emotions in real time, this system supports the progress of deeper and more appropriate discussions.

[0796] System Configuration

[0797] This system consists of the following main components:

[0798] 1. Interface means: A device (e.g., personal computer, tablet, smartphone) that allows the user to access and operate the system.

[0799] 2. Means of real-time collection: This refers to software or hardware (e.g., video conferencing APIs) that collect meeting audio data in real time.

[0800] 3. Means for converting audio data to text data: This refers to software (e.g., speech recognition API) for converting collected audio data into text data.

[0801] 4. Means of recognizing emotions: This is an engine that analyzes the user's emotions from voice data or text data (e.g., emotion recognition API).

[0802] 5. Means for extracting and organizing the key points of the discussion: This is a natural language processing (NLP) model for extracting and organizing the key points of the discussion from the converted text data and sentiment data.

[0803] 6. Means for generating and displaying meeting minutes: This is software for generating meeting minutes based on key points and sentiment data, and displaying them to the user.

[0804] 7. Means for proposing direction: This is an analytical function for proposing what should be discussed next.

[0805] 8. Means for collecting and saving feedback: This is software for collecting and saving user feedback and requests for improvements.

[0806] Specific examples of system operation

[0807] For example, consider a scenario where users A, B, and C discuss a new product's marketing strategy via video conference. Using this system, the following operations and actions can be performed:

[0808] 1. Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[0809] 2. The server collects the audio data of the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[0810] 3. Simultaneously, the server uses an emotion recognition API to analyze the emotional state of each statement and obtains information such as, "User A is excited when talking about the target market."

[0811] 4. Key points such as "identifying the target market" and "competitor analysis" are extracted from text data and sentiment data, and meeting minutes are automatically generated based on this.

[0812] 5. The generated meeting minutes are immediately displayed on the user's device via the video conference chat function or a dedicated interface.

[0813] 6. The server proposes "Pricing for the new product" as the next topic to be discussed and notifies the terminal of this.

[0814] 7. If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server will save this and use it for future analyses.

[0815] Example of a prompt

[0816] Examples of prompts to input into a generative AI model include the following:

[0817] Please prepare the minutes for the new product marketing strategy meeting. The topics discussed were as follows: The target market for the new product was discussed, including target market identification and competitive analysis. User emotions such as excitement, anticipation, and concern were observed. Pricing for the new product was proposed as the next topic to be discussed.

[0818] By providing specific context in this way, AI models can generate more appropriate meeting minutes.

[0819] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0820] Step 1: User login and meeting setup

[0821] Enter: Username and password

[0822] Specific actions: The user opens a browser on their device and accesses the system's login page. There, they enter their username and password and click the "Login" button.

[0823] Data processing: The terminal sends the input information to the server. The server compares the authentication information with the database, and if there is a match, authentication is considered successful. At this point, a session ID is generated.

[0824] Output: The server returns the generated session ID to the user's terminal, and the user becomes logged in.

[0825] For example, when a user selects "Create New Meeting," the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user enters the meeting ID to complete the joining process.

[0826] Step 2: Audio data collection and sequential transcription

[0827] Input: Meeting audio data

[0828] Specific operation: When a user starts a meeting, the server collects audio data in real time using the specified API (e.g., a video conferencing API).

[0829] Data processing: The collected audio data is sent to a speech recognition API and converted into text data.

[0830] Output: The converted text data is saved to the database by the server.

[0831] For example, if the audio data is "Discussing the target market for a new product," the speech recognition API will return this as text data.

[0832] Step 3: Recognizing Emotions

[0833] Input: Converted text data and audio data

[0834] Specific operation: The server sends the collected audio and text data to the specified emotion recognition API.

[0835] Data processing: The emotion recognition API analyzes the emotional state (e.g., joy, anger, sadness, etc.) in each statement and returns the result.

[0836] Output: The server saves the emotion data returned from the emotion recognition API to the database.

[0837] Example: If the emotion recognition API analyzes that User A is "excited" while talking about the target market, that information will be saved.

[0838] Step 4: Extract the key points of the discussion

[0839] Input: Saved text data and sentiment data

[0840] Specific operation: The server sends this data to a natural language processing (NLP) model and starts the extraction process.

[0841] Data processing: The NLP model identifies and extracts important keywords and phrases.

[0842] Output: The extracted key points data, along with sentiment data, is organized by the server and saved as data for meeting minutes.

[0843] Example: The NLP model extracts key points such as "identifying the target market" and "conducting competitive analysis."

[0844] Step 5: Generate and share meeting minutes

[0845] Input: Key data and sentiment data

[0846] Specific operation: Based on this data, the server generates meeting minutes summarizing the meeting content.

[0847] Data processing: The generated meeting minutes are converted into text files or HTML format.

[0848] Output: Meeting minutes are displayed on the user's device via the video conference chat function or a dedicated interface.

[0849] Example: All users will be able to view meeting minutes in real time.

[0850] Step 6: Propose the direction of the next discussion

[0851] Input: Discussion content and sentiment data

[0852] Specific operation: The server analyzes this data and uses an NLP model to identify the next topic to discuss.

[0853] Data processing: Determine the topic proposed by the NLP model.

[0854] Output: The proposed topic information will be notified to the user's terminal.

[0855] Example: The server suggests "Pricing for the new product" as the next discussion topic.

[0856] Step 7: Gathering Feedback

[0857] Input: User feedback information

[0858] Specific operation: Users input feedback on meeting minutes and proposals using the terminal's interface.

[0859] Data processing: The provided feedback information is collected and recorded by the server.

[0860] Output: Feedback information is stored in a database and used for future system improvements.

[0861] For example, if User B provides feedback such as "I would like to discuss the age range of the target market in more detail," that information will be saved.

[0862] (Application Example 2)

[0863] 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."

[0864] Traditional methods of recording meetings and discussions often involve manual transcription of audio data and extraction of key points, which is time-consuming and labor-intensive. Furthermore, it is difficult to consider the emotional state of participants during discussions, and there is a lack of means to support the proper progression of the discussion, resulting in inconsistent quality of meeting minutes and insufficient suggestions for the direction of future discussions.

[0865] 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.

[0866] In this invention, the server includes an interface means accessible to the user, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for proposing the direction of future discussions, means for collecting and saving feedback, emotion recognition means for recognizing and displaying the user's emotions in real time, means for identifying the content to be discussed next based on the main points and emotion data, and means for sharing the generated meeting minutes. This makes it possible to efficiently extract and organize the main points of the discussion, automatically generate high-quality meeting minutes, and propose the direction of the next discussion. Furthermore, by reflecting the user's emotional state in real time, it is possible to support the progress of discussions more deeply and appropriately.

[0867] "Interface means" refers to a device or software that allows a user to access and operate a system.

[0868] "Means for collecting audio data in real time" refers to a system component that acquires audio data from meetings and discussions and transmits it to a server in real time.

[0869] "Means for converting audio data to text data" refers to software or algorithms that transcribe collected audio data and convert it into text format.

[0870] "Methods for extracting and organizing the main points of a discussion from text data" refers to system components that analyze converted text data, extract important keywords and phrases, and organize the main points of the discussion.

[0871] "A means of generating and displaying meeting minutes based on key points" refers to a system component that automatically creates meeting minutes based on extracted key points and displays them in an easy-to-understand manner for the user.

[0872] "Means for proposing the direction of future discussions" refers to a system component that analyzes the content of discussions and sentiment data to identify and propose topics that should be discussed next.

[0873] "Means for collecting and storing feedback" refers to system components for collecting feedback information such as opinions and requests for corrections from users and storing it in a database.

[0874] An "emotion recognition tool" is an engine or software that analyzes and recognizes emotions in real time from a user's voice or text data.

[0875] "Means for sharing generated meeting minutes" refers to a system component that provides the functionality to share generated meeting minutes with other users or systems.

[0876] This invention combines a system that collects audio data from meetings and discussions in real time, transcribes it sequentially to extract and organize the key points of the discussion, and automatically generates meeting minutes with an emotion engine that recognizes the user's emotions in real time. This system is composed of the following various means.

[0877] Hardware and software to be used

[0878] User interface means: A smartphone, tablet, or personal computer is used as a device for the user to access and operate the system.

[0879] Audio data collection method: To collect meeting audio data in real time, for example, use the ZOOM API.

[0880] Audio data conversion method: The Google Speech-to-Text API is used to convert the collected audio data into text data.

[0881] Emotion recognition method: To analyze user emotions from voice and text data, an emotion recognition API (e.g., Microsoft Azure or AWS Rekognition) is used.

[0882] Key points of the discussion: Natural language processing (NLP) models (e.g., BERT or GPT-3) are used to extract and organize the key points of the discussion from text data and sentiment data.

[0883] Meeting minutes generation method: Software is used to generate meeting minutes based on key points and sentiment data, and to display them to the user.

[0884] Method for suggesting the direction of discussion: Utilize software or algorithms with analytical capabilities to suggest the next topics to be discussed.

[0885] Feedback collection method: Use an interface and database to collect and store user feedback and correction requests.

[0886] Meeting minutes sharing method: It has software functionality to allow generated meeting minutes to be shared with other users or systems.

[0887] Specific examples of how the system works

[0888] 1. Login and meeting setup:

[0889] Users access the system from devices such as smartphones or computers and enter their username and password on the login page. The server verifies the entered authentication information against the database, and if authentication is successful, it creates a session and returns a session ID to the user. After logging in, the user selects either "Create a new meeting" or "Join an existing meeting," and the server generates or authenticates a meeting ID.

[0890] 2. Collection and transcription of audio data:

[0891] The server uses the ZOOM API to acquire audio data from meetings in real time. The acquired audio data is converted into text data using the Google Speech-to-Text API, and the text data is stored in a database by the server.

[0892] 3. Recognition of emotions:

[0893] The server sends the collected audio and text data to an emotion recognition API to analyze the user's emotions. The emotion recognition API returns the emotional state for each statement (e.g., joy, anger, sadness), and the server stores this in a database.

[0894] 4. Extracting the main points of the discussion:

[0895] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations. The extracted key points and sentiment data are then organized by the server and stored as data for generating meeting minutes.

[0896] 5. Creating and sharing meeting minutes:

[0897] The server generates meeting minutes summarizing the discussion based on key points and sentiment data. The generated minutes are converted into text files or HTML format and displayed on the user's device, as well as shared with other users through the sharing function.

[0898] 6. Proposed direction for the discussion:

[0899] The server analyzes the discussion content and sentiment data, and uses an NLP model to identify the next topic to discuss. Suggestions are notified to the user's terminal to support the progress of the discussion.

[0900] 7. Gathering feedback:

[0901] Users can input feedback on meeting minutes and proposals through the interface. The feedback provided is collected and stored by the server and used to improve the system.

[0902] Specific examples and prompt statements

[0903] For example, consider a scenario where store staff hold a meeting about how to display new products. Using this system, the following operations can be performed:

[0904] Staff members A and B log into the system from their respective devices, enter the meeting ID, and join the meeting.

[0905] The server collects audio data from the meeting and converts statements like "We will discuss how to display the new products" into text in real time.

[0906] At the same time, the server uses an emotion recognition API to analyze the emotional state of each statement, obtaining information such as, "Staff member A is excited when talking about the display method."

[0907] The system extracts key points such as "areas for improvement in display methods" and "customer reactions" from text data and sentiment data, and then automatically generates meeting minutes based on this information.

[0908] The generated meeting minutes can be shared with other staff members using the sharing function, allowing them to review them immediately.

[0909] The server suggests "Pricing for the new product" as the next topic to discuss and notifies the staff member's terminal.

[0910] Examples of specific prompt messages:

[0911] "Customer complaint about product quality."

[0912] "Next discussion topic: Improve customer complaint handling."

[0913] This makes it possible to efficiently extract key points of discussions, including user emotions, based on the form in which the invention is implemented, and to generate high-quality meeting minutes.

[0914] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0915] Step 1:

[0916] User login and meeting setup

[0917] Input: Users access the system from a device such as a smartphone or computer and enter their username and password.

[0918] Data processing / calculation: The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID.

[0919] Output: Upon successful authentication, the session ID will be displayed on the user's device.

[0920] Specific operation: The user interacts with the interface and selects either "Create a new meeting" or "Join an existing meeting." If "Create a new meeting" is selected, the server generates a meeting ID and notifies the user.

[0921] Step 2:

[0922] Audio data collection and sequential transcription

[0923] Input: The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[0924] Data processing / calculation: The acquired audio data is sent to a speech recognition API (Google Speech-to-Text API) and converted into text data.

[0925] Output: The converted text data is saved on the server.

[0926] Specific operation: The server updates text data in real time and displays it to the user on the interface.

[0927] Step 3:

[0928] Recognition of emotions

[0929] Input: The server sends collected audio and text data to the emotion recognition API.

[0930] Data Processing / Calculation: The emotion recognition API analyzes the emotional state (e.g., joy, anger, sadness, etc.) in each statement and returns it as emotion data.

[0931] Output: Emotional data is stored on the server.

[0932] Specific operation: The server displays an emotion indicator corresponding to each statement on the interface, allowing the user to visually confirm it.

[0933] Step 4:

[0934] Extracting the main points of the discussion

[0935] Input: The server sends stored text data and sentiment data to a natural language processing (NLP) model.

[0936] Data processing / calculation: The NLP model extracts important keywords and phrases and further analyzes emotional fluctuations.

[0937] Output: Extracted key points and sentiment data are stored on the server.

[0938] Specific operation: The server organizes the key points and displays them in real time on the interface as data for generating meeting minutes.

[0939] Step 5:

[0940] Creating and sharing meeting minutes

[0941] Input: The server summarizes the discussion based on key points and sentiment data.

[0942] Data processing / calculation: Automatically create meeting minutes using a meeting minutes generation algorithm.

[0943] Output: The generated meeting minutes are converted to a text file or HTML format.

[0944] Specific operation: Meeting minutes are displayed to the user on the interface and can be sent to other users via the sharing function.

[0945] Step 6:

[0946] Proposal for the direction of the discussion

[0947] Input: The server analyzes the content of the discussion and sentiment data.

[0948] Data processing / computation: Use an NLP model to identify the next topic to discuss.

[0949] Output: The proposed content is notified to the user's device.

[0950] Specific behavior: The user interface displays specific suggestions for the next topic to be discussed, such as "Pricing for the new product."

[0951] Step 7:

[0952] Gathering feedback

[0953] Input: The user provides feedback through the interface.

[0954] Data processing / calculation: Collect and store feedback information in a database.

[0955] Output: The collected feedback information will be used to improve the system in the future.

[0956] Specific operation: Based on user feedback, the system automatically adjusts areas for improvement and discussion topics for the next meeting.

[0957] 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.

[0958] 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.

[0959] 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.

[0960] [Third Embodiment]

[0961] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0962] 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.

[0963] 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).

[0964] 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.

[0965] 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.

[0966] 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).

[0967] 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.

[0968] 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.

[0969] 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.

[0970] 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.

[0971] 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.

[0972] 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".

[0973] This invention is a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggesting directions for future discussions. This system is realized through the cooperation of the user, server, and terminal.

[0974] System Configuration

[0975] The system consists of the following main components:

[0976] User interface means: A device (e.g., personal computer, tablet, smartphone) that allows a user to access and operate a system.

[0977] Audio data collection method: This refers to software or hardware (e.g., ZOOM API) that collects meeting audio data in real time.

[0978] Audio data conversion means: This refers to software (e.g., Google Speech-to-Text API) for converting collected audio data into text data.

[0979] Key Point Extraction Method: This is a natural language processing (NLP) model for extracting and organizing the key points of an argument from converted text data.

[0980] Meeting minutes generation method: This is software that generates meeting minutes based on key points and displays them to the user.

[0981] A means of suggesting the direction of discussion: This is an analytical function for suggesting the next topics to be discussed.

[0982] Feedback collection method: This is software for collecting and saving user feedback and requests for improvements.

[0983] System operation

[0984] 1. User login and meeting setup

[0985] The user accesses the system from their device and enters their username and password on the login page.

[0986] The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[0987] Once logged in, users can select either "Create a new meeting" or "Join an existing meeting" on the interface.

[0988] When creating a meeting, the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user can join by entering the meeting ID.

[0989] 2. Collection of audio data and sequential transcription

[0990] The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[0991] The acquired audio data is sent to a speech recognition API and converted into text data. The server then sequentially saves the converted text data to a database.

[0992] 3. Extracting the key points of the discussion

[0993] The server sends the stored text data to a natural language processing (NLP) model at regular intervals to extract important keywords and phrases.

[0994] The extracted key points are organized by the server and saved as data for meeting minutes.

[0995] 4. Creating and sharing meeting minutes

[0996] The server generates meeting minutes summarizing the discussion based on the key points.

[0997] The generated meeting minutes are converted into text files or HTML format and displayed on the user's device via Zoom's chat function or a dedicated interface.

[0998] 5. Proposal for the direction of the next discussion

[0999] The server analyzes the content of the discussion and uses an NLP model to identify the next topic to discuss.

[1000] The proposal will be notified to the user's device, supporting the progress of the discussion.

[1001] 6. Gathering Feedback

[1002] Users can provide feedback on meeting minutes and proposals from their devices.

[1003] The feedback provided will be collected and stored by the server and used to improve the system.

[1004] Specific example

[1005] For example, consider a scenario where users A, B, and C discuss the marketing strategy for a new product via Zoom. Using this system, the following operations can be performed:

[1006] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[1007] The server collects audio data from the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[1008] The system extracts key points such as "identifying the target market" and "competitor analysis" from text data, and then automatically generates meeting minutes based on this information.

[1009] The generated meeting minutes are displayed in the Zoom chat, allowing all users to view them in real time.

[1010] The server proposes "Pricing for the new product" as the next topic to discuss and notifies the terminal of this.

[1011] If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server saves this and uses it for future analyses.

[1012] In this way, the system of the present invention enables efficient and in-depth discussions by summarizing the key points of the discussion in real time and suggesting the next topics to be discussed.

[1013] The following describes the processing flow.

[1014] Step 1:

[1015] The user accesses the system's login page from their terminal. The user enters their username and password and completes the login process.

[1016] Step 2:

[1017] The server compares the user's authentication information with the database, and if correct, creates a session and returns a session ID to the user.

[1018] Step 3:

[1019] The user selects either "Create a new meeting" or "Join an existing meeting" on the terminal interface. If the user selects "Create a new meeting," the server generates a meeting ID and notifies the user of it.

[1020] Step 4:

[1021] When joining an existing meeting, the user enters the meeting ID, and the server authenticates that ID. Once the user joins the meeting, real-time audio data collection begins.

[1022] Step 5:

[1023] The server uses the ZOOM API to acquire audio data during the meeting in real time. The collected audio data is then sent to the server.

[1024] Step 6:

[1025] The server continuously sends audio data to a speech recognition API (e.g., Google Speech-to-Text API), converting the audio data into text data in real time. The converted text data is then stored in a database.

[1026] Step 7:

[1027] The server sends the stored text data to a natural language processing (NLP) model at regular intervals to extract important keywords and phrases.

[1028] Step 8:

[1029] The server organizes the extracted key points chronologically and saves them as data for meeting minutes.

[1030] Step 9:

[1031] The server generates meeting minutes in real time based on the stored key points data. The generated minutes are then instructed to be displayed on the device via Zoom's chat function or a dedicated interface.

[1032] Step 10:

[1033] The server analyzes the discussion and uses an NLP model to identify the next topic to discuss. The identified topic is generated as a suggestion and notified to the user's terminal.

[1034] Step 11:

[1035] Users input feedback on meeting minutes and proposals through the terminal's interface.

[1036] Step 12:

[1037] The server stores the collected feedback in a database. This stored feedback is then used as training data to improve the system's natural language processing model.

[1038] (Example 1)

[1039] 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."

[1040] In traditional meetings and discussions, creating meeting minutes was cumbersome, making it difficult to summarize key points of important discussions in real time. Furthermore, it was impossible to suggest future directions for discussions, and it was difficult to effectively collect participant feedback and use it for improvement. This resulted in decreased meeting efficiency and the risk of important discussions being overlooked.

[1041] 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.

[1042] In this invention, the server includes an interface means accessible to users, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for analyzing the content of the discussion and suggesting what should be discussed next, means for collecting and saving user feedback, means for using a communication service application programming interface for collecting audio data, means for using a speech recognition application programming interface to convert the audio data into text data, and means for using a natural language processing model for extracting main points. This makes it possible to automatically generate meeting minutes in real time, support the progress of meetings by suggesting what should be discussed next, and effectively collect feedback from participants to improve the system.

[1043] "Interface means" refers to devices and software that allow users to access and operate a system.

[1044] "Audio data" refers to the digital data of audio signals collected during meetings or discussions.

[1045] "Means of real-time collection" refers to equipment and software for instantly collecting audio data during a meeting.

[1046] "Text data" refers to character data generated by transcribing audio data.

[1047] "Methods for extracting and organizing the key points of an argument" refers to natural language processing models and algorithms that find important information from text data and organize it systematically.

[1048] "Means for generating and displaying meeting minutes" refers to software that automatically creates meeting minutes based on extracted key points and presents them to the user.

[1049] "Means of proposing a direction for discussion" refers to a function that analyzes the content of the current discussion and indicates the points that should be discussed next.

[1050] "Means for collecting and saving feedback" refers to a system for collecting user opinions and correction requests and saving them for later use.

[1051] "Application programming interface for communication services" refers to the means of utilizing functions through the programming interface provided by communication services.

[1052] A "Speech Recognition Application Programming Interface" refers to a programming interface that provides speech recognition functionality for converting speech data into text.

[1053] A "natural language processing model" refers to machine learning models and algorithms used to understand and analyze human language.

[1054] The system of this invention collects audio data in real time during meetings and discussions and transcribes it sequentially. It then extracts and organizes the key points of the discussion to automatically generate meeting minutes. Furthermore, it proposes directions for future discussions and collects and stores user feedback. This system is realized through the cooperation of the user, server, and terminal.

[1055] Key components of the system

[1056] Interface means: A device that allows a user to access and operate a system (e.g., a personal computer, tablet, or smartphone).

[1057] Audio data collection means: A communication service application programming interface (e.g., a communication service API) for collecting conference audio data in real time.

[1058] Audio data conversion means: A speech recognition application programming interface (e.g., speech recognition API) for converting collected audio data into text data.

[1059] Key point extraction method: A natural language processing model for extracting and organizing the key points of a discussion from text data.

[1060] Meeting minutes generation method: Software for generating and displaying meeting minutes based on key points.

[1061] A means of suggesting the direction of discussion: A function for analyzing the content of a discussion and suggesting what should be discussed next.

[1062] Feedback collection method: Software for collecting and storing user feedback and correction requests.

[1063] System operation

[1064] The user accesses the system from their terminal and enters their username and password on the login page. The server verifies the authentication information against the database, and if successful, creates a session and returns a session ID to the user. After logging in, the user selects either "Create a new meeting" or "Join an existing meeting." The server generates a meeting ID and notifies the user of it.

[1065] During the meeting, the server uses a communication service API to acquire audio data in real time. The acquired audio data is sent to a speech recognition API and converted into text data. The server then sequentially saves the converted text data to a database.

[1066] The server sends the stored text data to a natural language processing model at regular intervals to extract important keywords and phrases. The extracted key points are then organized by the server and stored as data for meeting minutes.

[1067] The server generates meeting minutes summarizing the discussion based on key points. The generated minutes are converted into text files or HTML format and displayed on the user's device through the communication service's chat function or a dedicated interface.

[1068] Furthermore, the server analyzes the content of the discussion and uses a natural language processing model to identify the next topic to discuss. This suggestion is then communicated to the user's terminal to support the progress of the discussion.

[1069] Users can provide feedback on meeting minutes and proposals from their devices. The feedback provided is collected and stored by the server and used to improve the system.

[1070] Specific example

[1071] For example, consider a scenario where users A, B, and C discuss a new product's marketing strategy through a communication service. Using this system, the following operations can be performed:

[1072] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[1073] The server collects audio data from the meeting via a communication service API and converts statements such as "We will discuss the target market for the new product" into text in real time.

[1074] The system extracts key points such as "identifying the target market" and "competitor analysis" from text data, and then automatically generates meeting minutes based on this information.

[1075] The generated meeting minutes are displayed in the communication service's chat, allowing all users to view them in real time.

[1076] The server proposes "Pricing for the new product" as the next topic to discuss and notifies the terminal of this.

[1077] If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server saves this and uses it for future analyses.

[1078] Example of a prompt

[1079] "We've developed a system that collects meeting audio data in real time and transcribes it sequentially. It also has a function to identify and suggest the next topic of discussion. For example, if we're discussing the marketing strategy for a new product, and topics like the target market and competitor analysis come up, how can you tell me how to suggest the next topic to discuss?"

[1080] In this way, the system of the present invention enables efficient and in-depth discussions by summarizing the key points of the discussion in real time and suggesting the next topics to be discussed.

[1081] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1082] Step 1: User login and meeting setup

[1083] Specific operation: The user accesses the system from their terminal and enters their username and password on the login page.

[1084] Enter: Username and password.

[1085] Data processing / calculation: The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[1086] Output: Session ID.

[1087] After the user logs in, they can select either "Create a New Meeting" or "Join an Existing Meeting" on the interface. If a meeting is created, the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user enters the meeting ID.

[1088] Step 2: Audio data collection and sequential transcription

[1089] Specific operation: The server uses a communication service API to acquire audio data during the meeting in real time.

[1090] Input: Voice data obtained from a communication service API.

[1091] Data processing / calculation: The server sends the acquired audio data to a speech recognition API and converts it into text data. The server then sequentially saves the converted text data to a database.

[1092] Output: Text data.

[1093] Step 3: Extract the key points of the discussion

[1094] Specific operation: The server sends the stored text data to a natural language processing model at regular intervals, and extracts important keywords and phrases.

[1095] Input: Text data.

[1096] Data Processing / Calculation: A natural language processing model analyzes the text data and extracts important keywords and phrases. The server organizes the extracted key points and saves them as data for meeting minutes.

[1097] Output: Key data.

[1098] Step 4: Generate and share meeting minutes

[1099] Specific operation: The server generates meeting minutes summarizing the discussion based on key points.

[1100] Input: Key data.

[1101] Data Processing / Calculation: Automatically creates meeting minutes based on key data and converts them to text files or HTML format.

[1102] Output: Meeting minutes file (text file / HTML format).

[1103] The generated meeting minutes are displayed on the user's device through the communication service's chat function or a dedicated interface.

[1104] Step 5: Propose the direction of the next discussion

[1105] Specific operation: The server analyzes the content of the discussion and uses a natural language processing model to identify the next topic to discuss.

[1106] Input: Meeting minutes data.

[1107] Data processing / computation: A natural language processing model is used to identify the next topic to discuss. The server then notifies the user's terminal of the identified topic as a suggestion.

[1108] Output: Suggestions for the next topic to discuss.

[1109] Step 6: Gathering Feedback

[1110] Specific functionality: Users can provide feedback on meeting minutes and proposals from their devices.

[1111] Input: User feedback.

[1112] Data processing / calculation: The server collects and stores user feedback, which is later used to improve the system.

[1113] Output: Saved feedback data.

[1114] In this way, a series of processes can be realized in which the server, terminal, and user cooperate with each other to summarize the key points of the discussion in real time and propose the next topics to be discussed.

[1115] (Application Example 1)

[1116] 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."

[1117] In modern meetings and discussions, creating meeting minutes, summarizing key points, and proposing directions for future discussions are crucial, yet extremely time-consuming. Therefore, there is a need for a system that collects audio data in real time, automatically generates meeting minutes, and efficiently proposes directions for future discussions. Furthermore, in environments requiring rapid decision-making, such as logistics centers, the implementation of an efficient meeting management system is urgently needed.

[1118] 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.

[1119] In this invention, the server includes an interface means accessible to the user, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for proposing the direction of future discussions, means for collecting and saving feedback, communication means for acquiring audio data during a meeting, means for transmitting the acquired audio data to a speech recognition platform and generating sequential transcript data, means for transmitting the converted text data to a natural language processing model and extracting important keywords and phrases, means for proposing topics to be discussed next based on the content of the discussion, and means for improving the system's performance using feedback. This enables real-time audio collection, automatic generation of meeting minutes on the spot, efficient and rapid progress of discussions and organization of main points, and proposals for topics to be discussed next.

[1120] "Interface means" refers to devices or software that users can access and operate.

[1121] "Means of collecting audio data in real time" refers to devices or software for instantly acquiring audio during meetings or discussions.

[1122] "Means for converting audio data into text data" refers to devices or software that convert collected audio data into textual information.

[1123] "Methods for extracting and organizing the key points of an argument" refer to techniques for finding important keywords and phrases from text data and organizing them systematically.

[1124] "Means for generating and displaying meeting minutes" refers to technology that records the content of a meeting based on extracted key points and provides it to the user.

[1125] "Methods for proposing the direction of future discussions" refers to techniques that suggest topics or themes to be discussed next, based on the current discussion.

[1126] "Means for collecting and storing feedback" refers to technologies that record user opinions and requests for later analysis and improvement.

[1127] "Communication means for acquiring audio data during a meeting" refers to technology for transmitting audio data from a meeting to a server.

[1128] A "speech recognition platform" refers to speech recognition technologies and APIs used to convert speech data into text data.

[1129] A "natural language processing model" is an artificial intelligence technology that analyzes the meaning and content of text data and extracts key points.

[1130] "A means of suggesting the next topic to be discussed" refers to a technology that analyzes the current discussion and suggests themes or topics to be addressed next to the user.

[1131] "Means for improving system performance" refer to technologies that analyze collected feedback and data to improve the overall efficiency and accuracy of the system.

[1132] This invention is a system that automatically generates meeting minutes at sites such as logistics centers, efficiently extracts the key points of the discussion, and suggests topics to be discussed next.

[1133] System Configuration

[1134] This system consists of the following main components:

[1135] User interface means: Devices and software (e.g., personal computers, mobile devices, smartphones) that allow users to access and operate a system.

[1136] Means of collecting audio data in real time: These are devices or software for instantly acquiring audio data from meetings (e.g., software or hardware that collects audio remotely using communication means).

[1137] Means of converting audio data into text data: These are devices or software that convert collected audio data into text information (e.g., Google Speech-to-Text API).

[1138] A means of extracting and organizing the key points of an argument: This refers to techniques for finding important keywords and phrases from text data and organizing them systematically (e.g., Google Natural Language API).

[1139] A means of generating and displaying meeting minutes: This technology records the content of a meeting based on extracted key points and provides it to the user.

[1140] A means of proposing the direction of future discussions: This is a technique that suggests topics or themes to be discussed next, based on the content of the current discussion.

[1141] A means of collecting and storing feedback: This is a technology for recording user opinions and requests and using them later for analysis and improvement.

[1142] System operation

[1143] 1. Starting the meeting and collecting audio data:

[1144] Users access the system through an interface and initiate a new meeting. During the meeting, audio data is collected in real time using communication methods.

[1145] 2. Converting audio to text data:

[1146] The collected audio data is sent to a speech recognition platform (e.g., Google Speech-to-Text API), where sequential transcription data is generated. The server stores the converted text data in a database.

[1147] 3. Extracting the key points of the discussion:

[1148] The server sends the stored text data to a natural language processing model (e.g., Google Natural Language API) at regular intervals to extract important keywords and phrases. The extracted key points are then organized by the server and stored as data for meeting minutes.

[1149] 4. Generating and displaying meeting minutes:

[1150] The server generates meeting minutes summarizing the discussion based on key points. The generated minutes are converted into text files or HTML format and displayed to the user through an interface.

[1151] 5. Proposed direction for the next discussion:

[1152] The server analyzes the discussion content and uses a natural language processing model to identify the next topic to discuss. The suggested topics are communicated to the user to support the progress of the discussion.

[1153] 6. Collecting and saving feedback:

[1154] Users can provide feedback on meeting minutes and proposals through the interface. This feedback is collected by the server and used to improve system performance.

[1155] Specific example

[1156] Consider a scenario where a manager at a logistics center uses this system to hold a meeting. The manager starts the meeting using their smartphone, and on-site staff participate. Audio data is collected in real time and converted into text data using the Google Speech-to-Text API. Key points are extracted using the Google Natural Language API, and meeting minutes are automatically generated and displayed. Furthermore, the system suggests topics to discuss next, allowing the manager to conduct the meeting efficiently. A feedback function collects user opinions, contributing to the improvement of the system's performance.

[1157] Example of a prompt

[1158] "We are developing an application to support logistics management meetings. We would like to implement the following functions using audio data collected during the meetings:

[1159] 1. Collect audio data in real time and convert it into text data.

[1160] 2. Extract important keywords and phrases from the converted text data and automatically generate meeting minutes.

[1161] 3. Based on the extracted data, propose the next topic to discuss.

[1162] The application will be implemented using speech recognition APIs and natural language processing APIs. We would appreciate your suggestions for this system.

[1163] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1164] Step 1:

[1165] The user initiates a new meeting using the interface. A meeting start request is sent to the server. The server generates a new meeting ID and returns it to the user. This completes the meeting preparation. The input is the user's request, and the output is the new meeting ID.

[1166] Step 2:

[1167] When the meeting begins, the terminal starts collecting audio data in real time. The collected audio data is sent to the server via a communication method. The server prepares to process the received audio data sequentially. The input is real-time audio data, and the output is the transmitted audio data.

[1168] Step 3:

[1169] The server sends audio data to a speech recognition platform (e.g., Google Speech-to-Text API) and converts it into text data. Once the audio data is converted to text data, the server stores it in a database. The input is audio data, and the output is text data.

[1170] Step 4:

[1171] The server sends the converted text data to a natural language processing model (e.g., Google Natural Language API) at regular intervals to extract important keywords and phrases. The extracted key points are organized by the server and saved for meeting minutes. The input is text data, and the output is important keywords and phrases.

[1172] Step 5:

[1173] The server automatically generates meeting minutes based on the extracted key points. The generated minutes are converted into text files or HTML format and displayed on the terminal via a user interface. Users can review the minutes in real time. Input consists of important keywords and phrases, while output is the meeting minutes data.

[1174] Step 6:

[1175] The server uses a natural language processing model to suggest the next topic to discuss, based on the generated meeting minutes data. The suggestions are then communicated to the user's terminal to support the discussion. The input is the meeting minutes data, and the output is the next topic to discuss.

[1176] Step 7:

[1177] After the meeting, users provide feedback using an interface. The server collects and stores this feedback. The collected feedback data is later used to improve the system. Input is user feedback, and output is data and improvement suggestions related to system performance improvements.

[1178] 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.

[1179] This invention combines a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggests directions for future discussions, with an emotion engine that recognizes the user's emotions. By reflecting the user's emotions in real time, this system supports the progress of deeper and more appropriate discussions.

[1180] System Configuration

[1181] The system consists of the following main components:

[1182] User interface means: A device (e.g., personal computer, tablet, smartphone) that allows a user to access and operate a system.

[1183] Audio data collection method: This refers to software or hardware (e.g., ZOOM API) that collects meeting audio data in real time.

[1184] Audio data conversion means: This refers to software (e.g., Google Speech-to-Text API) for converting collected audio data into text data.

[1185] Emotion recognition means: This is an engine (e.g., emotion recognition API) that analyzes a user's emotions from voice data or text data.

[1186] Key Point Extraction Method: This is a natural language processing (NLP) model for extracting and organizing the key points of an argument from converted text data and sentiment data.

[1187] Meeting minutes generation method: This is software that generates meeting minutes based on key points and sentiment data, and displays them to the user.

[1188] A means of suggesting the direction of discussion: This is an analytical function for suggesting the next topics to be discussed.

[1189] Feedback collection method: This is software for collecting and saving user feedback and requests for improvements.

[1190] System operation

[1191] 1. User login and meeting setup

[1192] The user accesses the system from their device and enters their username and password on the login page.

[1193] The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[1194] Once logged in, users can select either "Create a new meeting" or "Join an existing meeting" on the interface.

[1195] When creating a meeting, the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user can join by entering the meeting ID.

[1196] 2. Collection of audio data and sequential transcription

[1197] The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[1198] The acquired audio data is sent to a speech recognition API and converted into text data. The server then stores the converted text data in a database.

[1199] 3. Recognition of emotions

[1200] The server sends the collected voice and text data to an emotion recognition API to analyze the user's emotions.

[1201] The emotion recognition API returns the emotional state (e.g., joy, anger, sadness) for each statement, and the server stores this information in a database.

[1202] 4. Extracting the key points of the discussion

[1203] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations.

[1204] The extracted key points and sentiment data are organized by the server and stored as data for meeting minutes.

[1205] 5. Creating and sharing meeting minutes

[1206] The server generates meeting minutes summarizing the discussion based on key points and sentiment data.

[1207] The generated meeting minutes are converted into text files or HTML format and displayed on the user's device via Zoom's chat function or a dedicated interface.

[1208] 6. Proposal for the direction of the next discussion

[1209] The server analyzes the content and sentiment data of the discussion and uses an NLP model to identify the next topic to discuss.

[1210] The proposal will be notified to the user's device, supporting the progress of the discussion.

[1211] 7. Gathering Feedback

[1212] Users can input meeting minutes and feedback on proposals through the terminal's interface.

[1213] The feedback provided will be collected and stored by the server and used to improve the system.

[1214] Specific example

[1215] For example, consider a scenario where users A, B, and C discuss the marketing strategy for a new product via Zoom. Using this system, the following operations can be performed:

[1216] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[1217] The server collects audio data from the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[1218] Simultaneously, the server uses an emotion recognition API to analyze the emotional state of each statement and obtains information such as, "User A is excited when talking about the target market."

[1219] Key points such as "identifying the target market" and "competitor analysis" are extracted from text data and sentiment data, and meeting minutes are automatically generated based on this information.

[1220] The generated meeting minutes are displayed in the Zoom chat, allowing all users to view them in real time.

[1221] The server proposes "Pricing for the new product" as the next topic to discuss and notifies the terminal of this.

[1222] If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server saves this and uses it for future analyses.

[1223] In this way, the system of the present invention enables efficient and in-depth discussions by summarizing the key points of the discussion and the users' feelings in real time and suggesting the next topics to discuss.

[1224] The following describes the processing flow.

[1225] Step 1:

[1226] The user accesses the system's login page from their terminal. The user enters their username and password and completes the login process.

[1227] Step 2:

[1228] The server compares the user's authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[1229] Step 3:

[1230] The user selects either "Create a new meeting" or "Join an existing meeting" on the terminal interface. If the user selects "Create a new meeting," the server generates a meeting ID and notifies the user of it.

[1231] Step 4:

[1232] When joining an existing meeting, the user enters the meeting ID, and the server authenticates that ID. Once the user joins the meeting, real-time audio data collection begins.

[1233] Step 5:

[1234] The server uses the ZOOM API to acquire audio data during the meeting in real time. The collected audio data is then sent to the server.

[1235] Step 6:

[1236] The server continuously sends audio data to a speech recognition API (e.g., Google Speech-to-Text API), converting the audio data into text data in real time. The converted text data is then stored in a database.

[1237] Step 7:

[1238] The server simultaneously sends audio and text data to an emotion recognition API to analyze the emotions contained in the user's statements. The emotion recognition API returns the emotional state for each statement (e.g., joy, anger, sadness), and the server stores this in a database.

[1239] Step 8:

[1240] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations.

[1241] Step 9:

[1242] The server organizes the extracted key points and sentiment data chronologically and saves them as data for meeting minutes.

[1243] Step 10:

[1244] The server generates meeting minutes in real time based on stored key points and sentiment data. The generated minutes are then instructed to be displayed on the device via Zoom's chat function or a dedicated interface. Sentiment information is also included in the meeting minutes.

[1245] Step 11:

[1246] The server analyzes the content and sentiment data of the discussion and uses an NLP model to identify the next topic to discuss. The identified topic is generated as a suggestion and notified to the user's terminal.

[1247] Step 12:

[1248] Users input feedback on meeting minutes and proposals through the terminal's interface.

[1249] Step 13:

[1250] The server stores the collected feedback in a database. This stored feedback is then used as training data to improve the system's natural language processing model.

[1251] (Example 2)

[1252] 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."

[1253] In today's business environment, effective meeting management is crucial. However, in many meetings, creating meeting minutes is time-consuming, making it difficult to summarize the key points of the discussion. Furthermore, it's challenging to understand participants' emotions and support the flow of the discussion. As a result, meaningful discussions fail, and meeting efficiency suffers. Additionally, there's the challenge of not being able to clearly define the direction the discussion should take next.

[1254] 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.

[1255] In this invention, the server includes an interface means accessible to users, means for collecting audio data in real time, means for converting the collected audio data into text data, means for recognizing emotions from the text data and audio data, means for extracting and organizing the main points of the discussion from the text data and emotion data, means for generating and displaying meeting minutes based on the main points and emotion data, means for proposing the direction of future discussions, and means for collecting and saving feedback. This makes it possible to grasp the main points of the discussion and the emotions of the participants in real time and propose the direction of the next discussion.

[1256] An "interface means" refers to a screen or device that allows a user to access and operate a system.

[1257] "Methods for real-time collection" refers to software and hardware systems for instantly acquiring audio data from meetings.

[1258] "Methods for converting audio data to text data" refers to software that uses speech recognition technology to convert audio data into text data.

[1259] "Means of recognizing emotions" refer to engines and algorithms that analyze voice and text data to identify a user's emotional state.

[1260] "Methods for extracting and organizing the key points of an argument" refer to natural language processing technologies that automatically identify important keywords and phrases and systematically summarize the points of contention.

[1261] "Methods for generating and displaying meeting minutes" refers to software that summarizes the content of a meeting based on extracted key points and sentiment data, and displays it in document format.

[1262] "Methods for suggesting direction" refers to algorithms that automatically determine and suggest the next topic to be discussed based on the content of the discussion and sentiment data.

[1263] A "means for collecting and saving feedback" refers to a system that records user opinions and requests for corrections in a database so that they can be used later.

[1264] This invention combines a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggests directions for future discussions, with an emotion engine that recognizes the user's emotions. By reflecting the user's emotions in real time, this system supports the progress of deeper and more appropriate discussions.

[1265] System Configuration

[1266] This system consists of the following main components:

[1267] 1. Interface means: A device (e.g., personal computer, tablet, smartphone) that allows the user to access and operate the system.

[1268] 2. Means of real-time collection: This refers to software or hardware (e.g., video conferencing APIs) that collect meeting audio data in real time.

[1269] 3. Means for converting audio data to text data: This refers to software (e.g., speech recognition API) for converting collected audio data into text data.

[1270] 4. Means of recognizing emotions: This is an engine that analyzes the user's emotions from voice data or text data (e.g., emotion recognition API).

[1271] 5. Means for extracting and organizing the key points of the discussion: This is a natural language processing (NLP) model for extracting and organizing the key points of the discussion from the converted text data and sentiment data.

[1272] 6. Means for generating and displaying meeting minutes: This is software for generating meeting minutes based on key points and sentiment data, and displaying them to the user.

[1273] 7. Means for proposing direction: This is an analytical function for proposing what should be discussed next.

[1274] 8. Means for collecting and saving feedback: This is software for collecting and saving user feedback and requests for improvements.

[1275] Specific examples of system operation

[1276] For example, consider a scenario where users A, B, and C discuss a new product's marketing strategy via video conference. Using this system, the following operations and actions can be performed:

[1277] 1. Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[1278] 2. The server collects the audio data of the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[1279] 3. Simultaneously, the server uses an emotion recognition API to analyze the emotional state of each statement and obtains information such as, "User A is excited when talking about the target market."

[1280] 4. Key points such as "identifying the target market" and "competitor analysis" are extracted from text data and sentiment data, and meeting minutes are automatically generated based on this.

[1281] 5. The generated meeting minutes are immediately displayed on the user's device via the video conference chat function or a dedicated interface.

[1282] 6. The server proposes "Pricing for the new product" as the next topic to be discussed and notifies the terminal of this.

[1283] 7. If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server will save this and use it for future analyses.

[1284] Example of a prompt

[1285] Examples of prompts to input into a generative AI model include the following:

[1286] Please prepare the minutes for the new product marketing strategy meeting. The topics discussed were as follows: The target market for the new product was discussed, including target market identification and competitive analysis. User emotions such as excitement, anticipation, and concern were observed. Pricing for the new product was proposed as the next topic to be discussed.

[1287] By providing specific context in this way, AI models can generate more appropriate meeting minutes.

[1288] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1289] Step 1: User login and meeting setup

[1290] Enter: Username and password

[1291] Specific actions: The user opens a browser on their device and accesses the system's login page. There, they enter their username and password and click the "Login" button.

[1292] Data processing: The terminal sends the input information to the server. The server compares the authentication information with the database, and if there is a match, authentication is considered successful. At this point, a session ID is generated.

[1293] Output: The server returns the generated session ID to the user's terminal, and the user becomes logged in.

[1294] For example, when a user selects "Create New Meeting," the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user enters the meeting ID to complete the joining process.

[1295] Step 2: Audio data collection and sequential transcription

[1296] Input: Meeting audio data

[1297] Specific operation: When a user starts a meeting, the server collects audio data in real time using the specified API (e.g., a video conferencing API).

[1298] Data processing: The collected audio data is sent to a speech recognition API and converted into text data.

[1299] Output: The converted text data is saved to the database by the server.

[1300] For example, if the audio data is "Discussing the target market for a new product," the speech recognition API will return this as text data.

[1301] Step 3: Recognizing Emotions

[1302] Input: Converted text data and audio data

[1303] Specific operation: The server sends the collected audio and text data to the specified emotion recognition API.

[1304] Data processing: The emotion recognition API analyzes the emotional state (e.g., joy, anger, sadness, etc.) in each statement and returns the result.

[1305] Output: The server saves the emotion data returned from the emotion recognition API to the database.

[1306] Example: If the emotion recognition API analyzes that User A is "excited" while talking about the target market, that information will be saved.

[1307] Step 4: Extract the key points of the discussion

[1308] Input: Saved text data and sentiment data

[1309] Specific operation: The server sends this data to a natural language processing (NLP) model and starts the extraction process.

[1310] Data processing: The NLP model identifies and extracts important keywords and phrases.

[1311] Output: The extracted key points data, along with sentiment data, is organized by the server and saved as data for meeting minutes.

[1312] Example: The NLP model extracts key points such as "identifying the target market" and "conducting competitive analysis."

[1313] Step 5: Generate and share meeting minutes

[1314] Input: Key point data and sentiment data

[1315] Specific operation: Based on this data, the server generates meeting minutes summarizing the meeting content.

[1316] Data processing: The generated meeting minutes are converted into text files or HTML format.

[1317] Output: Meeting minutes are displayed on the user's device via the video conference chat function or a dedicated interface.

[1318] Example: All users will be able to view meeting minutes in real time.

[1319] Step 6: Propose the direction of the next discussion

[1320] Input: Discussion content and sentiment data

[1321] Specific operation: The server analyzes this data and uses an NLP model to identify the next topic to discuss.

[1322] Data processing: Determine the topic proposed by the NLP model.

[1323] Output: The proposed topic information will be notified to the user's terminal.

[1324] Example: The server suggests "Pricing for the new product" as the next discussion topic.

[1325] Step 7: Gathering Feedback

[1326] Input: User feedback information

[1327] Specific operation: Users input feedback on meeting minutes and proposals using the terminal's interface.

[1328] Data processing: The provided feedback information is collected and recorded by the server.

[1329] Output: Feedback information is stored in a database and used for future system improvements.

[1330] For example, if User B provides feedback such as "I would like to discuss the age range of the target market in more detail," that information will be saved.

[1331] (Application Example 2)

[1332] 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."

[1333] Traditional methods of recording meetings and discussions often involve manual transcription of audio data and extraction of key points, which is time-consuming and labor-intensive. Furthermore, it is difficult to consider the emotional state of participants during discussions, and there is a lack of means to support the proper progression of the discussion, resulting in inconsistent quality of meeting minutes and insufficient suggestions for the direction of future discussions.

[1334] 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.

[1335] In this invention, the server includes an interface means accessible to the user, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for proposing the direction of future discussions, means for collecting and saving feedback, emotion recognition means for recognizing and displaying the user's emotions in real time, means for identifying the content to be discussed next based on the main points and emotion data, and means for sharing the generated meeting minutes. This makes it possible to efficiently extract and organize the main points of the discussion, automatically generate high-quality meeting minutes, and propose the direction of the next discussion. Furthermore, by reflecting the user's emotional state in real time, it is possible to support the progress of discussions more deeply and appropriately.

[1336] "Interface means" refers to a device or software that allows a user to access and operate a system.

[1337] "Means for collecting audio data in real time" refers to a system component that acquires audio data from meetings and discussions and transmits it to a server in real time.

[1338] "Means for converting audio data to text data" refers to software or algorithms that transcribe collected audio data and convert it into text format.

[1339] "Methods for extracting and organizing the main points of a discussion from text data" refers to system components that analyze converted text data, extract important keywords and phrases, and organize the main points of the discussion.

[1340] "A means of generating and displaying meeting minutes based on key points" refers to a system component that automatically creates meeting minutes based on extracted key points and displays them in an easy-to-understand manner for the user.

[1341] "Means for proposing the direction of future discussions" refers to a system component that analyzes the content of discussions and sentiment data to identify and propose topics that should be discussed next.

[1342] "Means for collecting and storing feedback" refers to system components for collecting feedback information such as opinions and requests for corrections from users and storing it in a database.

[1343] An "emotion recognition tool" is an engine or software that analyzes and recognizes emotions in real time from a user's voice or text data.

[1344] "Means for sharing generated meeting minutes" refers to a system component that provides the functionality to share generated meeting minutes with other users or systems.

[1345] This invention combines a system that collects audio data from meetings and discussions in real time, transcribes it sequentially to extract and organize the key points of the discussion, and automatically generates meeting minutes with an emotion engine that recognizes the user's emotions in real time. This system is composed of the following various means.

[1346] Hardware and software to be used

[1347] User interface means: A smartphone, tablet, or personal computer is used as a device for the user to access and operate the system.

[1348] Audio data collection method: To collect meeting audio data in real time, for example, use the ZOOM API.

[1349] Audio data conversion method: The Google Speech-to-Text API is used to convert the collected audio data into text data.

[1350] Emotion recognition method: To analyze user emotions from voice and text data, an emotion recognition API (e.g., Microsoft Azure or AWS Rekognition) is used.

[1351] Key points of the discussion: Natural language processing (NLP) models (e.g., BERT or GPT-3) are used to extract and organize the key points of the discussion from text data and sentiment data.

[1352] Meeting minutes generation method: Software is used to generate meeting minutes based on key points and sentiment data, and to display them to the user.

[1353] Method for suggesting the direction of discussion: Utilize software or algorithms with analytical capabilities to suggest the next topics to be discussed.

[1354] Feedback collection method: Use an interface and database to collect and store user feedback and correction requests.

[1355] Meeting minutes sharing method: It has software functionality to allow generated meeting minutes to be shared with other users or systems.

[1356] Specific examples of how the system works

[1357] 1. Login and meeting setup:

[1358] Users access the system from devices such as smartphones or computers and enter their username and password on the login page. The server verifies the entered authentication information against the database, and if authentication is successful, it creates a session and returns a session ID to the user. After logging in, the user selects either "Create a new meeting" or "Join an existing meeting," and the server generates or authenticates a meeting ID.

[1359] 2. Collection and transcription of audio data:

[1360] The server uses the ZOOM API to acquire audio data from meetings in real time. The acquired audio data is converted into text data using the Google Speech-to-Text API, and the text data is stored in a database by the server.

[1361] 3. Recognition of emotions:

[1362] The server sends the collected audio and text data to an emotion recognition API to analyze the user's emotions. The emotion recognition API returns the emotional state for each statement (e.g., joy, anger, sadness), and the server stores this in a database.

[1363] 4. Extracting the main points of the discussion:

[1364] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations. The extracted key points and sentiment data are then organized by the server and stored as data for generating meeting minutes.

[1365] 5. Creating and sharing meeting minutes:

[1366] The server generates meeting minutes summarizing the discussion based on key points and sentiment data. The generated minutes are converted into text files or HTML format and displayed on the user's device, as well as shared with other users through the sharing function.

[1367] 6. Proposed direction for the discussion:

[1368] The server analyzes the discussion content and sentiment data, and uses an NLP model to identify the next topic to discuss. Suggestions are notified to the user's terminal to support the progress of the discussion.

[1369] 7. Gathering feedback:

[1370] Users can input feedback on meeting minutes and proposals through the interface. The feedback provided is collected and stored by the server and used to improve the system.

[1371] Specific examples and prompt statements

[1372] For example, consider a scenario where store staff hold a meeting about how to display new products. Using this system, the following operations can be performed:

[1373] Staff members A and B log into the system from their respective devices, enter the meeting ID, and join the meeting.

[1374] The server collects audio data from the meeting and converts statements like "We will discuss how to display the new products" into text in real time.

[1375] At the same time, the server uses an emotion recognition API to analyze the emotional state of each statement, obtaining information such as, "Staff member A is excited when talking about the display method."

[1376] The system extracts key points such as "areas for improvement in display methods" and "customer reactions" from text data and sentiment data, and then automatically generates meeting minutes based on this information.

[1377] The generated meeting minutes can be shared with other staff members using the sharing function, allowing them to review them immediately.

[1378] The server suggests "Pricing for the new product" as the next topic to discuss and notifies the staff member's terminal.

[1379] Examples of specific prompt messages:

[1380] "Customer complaint about product quality."

[1381] "Next discussion topic: Improve customer complaint handling."

[1382] This makes it possible to efficiently extract key points of discussions, including user emotions, based on the form in which the invention is implemented, and to generate high-quality meeting minutes.

[1383] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1384] Step 1:

[1385] User login and meeting setup

[1386] Input: Users access the system from a device such as a smartphone or computer and enter their username and password.

[1387] Data processing / calculation: The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID.

[1388] Output: Upon successful authentication, the session ID will be displayed on the user's device.

[1389] Specific operation: The user interacts with the interface and selects either "Create a new meeting" or "Join an existing meeting." If "Create a new meeting" is selected, the server generates a meeting ID and notifies the user.

[1390] Step 2:

[1391] Audio data collection and sequential transcription

[1392] Input: The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[1393] Data processing / calculation: The acquired audio data is sent to a speech recognition API (Google Speech-to-Text API) and converted into text data.

[1394] Output: The converted text data is saved on the server.

[1395] Specific operation: The server updates text data in real time and displays it to the user on the interface.

[1396] Step 3:

[1397] Recognition of emotions

[1398] Input: The server sends collected audio and text data to the emotion recognition API.

[1399] Data Processing / Calculation: The emotion recognition API analyzes the emotional state (e.g., joy, anger, sadness, etc.) in each statement and returns it as emotion data.

[1400] Output: Emotional data is stored on the server.

[1401] Specific operation: The server displays an emotion indicator corresponding to each statement on the interface, allowing the user to visually confirm it.

[1402] Step 4:

[1403] Extracting the main points of the discussion

[1404] Input: The server sends stored text data and sentiment data to a natural language processing (NLP) model.

[1405] Data processing / calculation: The NLP model extracts important keywords and phrases and further analyzes emotional fluctuations.

[1406] Output: Extracted key points and sentiment data are stored on the server.

[1407] Specific operation: The server organizes the key points and displays them in real time on the interface as data for generating meeting minutes.

[1408] Step 5:

[1409] Creating and sharing meeting minutes

[1410] Input: The server summarizes the discussion based on key points and sentiment data.

[1411] Data processing / calculation: Automatically create meeting minutes using a meeting minutes generation algorithm.

[1412] Output: The generated meeting minutes are converted to a text file or HTML format.

[1413] Specific operation: Meeting minutes are displayed to the user on the interface and can be sent to other users via the sharing function.

[1414] Step 6:

[1415] Proposal for the direction of the discussion

[1416] Input: The server analyzes the content of the discussion and sentiment data.

[1417] Data processing / computation: Use an NLP model to identify the next topic to discuss.

[1418] Output: The proposed content is notified to the user's device.

[1419] Specific behavior: The user interface displays specific suggestions for the next topic to be discussed, such as "Pricing for the new product."

[1420] Step 7:

[1421] Gathering feedback

[1422] Input: The user provides feedback through the interface.

[1423] Data processing / calculation: Collect and store feedback information in a database.

[1424] Output: The collected feedback information will be used to improve the system in the future.

[1425] Specific operation: Based on user feedback, the system automatically adjusts areas for improvement and discussion topics for the next meeting.

[1426] 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.

[1427] 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.

[1428] 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.

[1429] [Fourth Embodiment]

[1430] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1431] 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.

[1432] 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).

[1433] 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.

[1434] 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.

[1435] 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).

[1436] 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.

[1437] 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.

[1438] 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.

[1439] 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.

[1440] 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.

[1441] 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.

[1442] 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".

[1443] This invention is a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggesting directions for future discussions. This system is realized through the cooperation of the user, server, and terminal.

[1444] System Configuration

[1445] The system consists of the following main components:

[1446] User interface means: A device (e.g., personal computer, tablet, smartphone) that allows a user to access and operate a system.

[1447] Audio data collection method: This refers to software or hardware (e.g., ZOOM API) that collects meeting audio data in real time.

[1448] Audio data conversion means: This refers to software (e.g., Google Speech-to-Text API) for converting collected audio data into text data.

[1449] Key Point Extraction Method: This is a natural language processing (NLP) model for extracting and organizing the key points of an argument from converted text data.

[1450] Meeting minutes generation method: This is software that generates meeting minutes based on key points and displays them to the user.

[1451] A means of suggesting the direction of discussion: This is an analytical function for suggesting the next topics to be discussed.

[1452] Feedback collection method: This is software for collecting and saving user feedback and requests for improvements.

[1453] System operation

[1454] 1. User login and meeting setup

[1455] The user accesses the system from their device and enters their username and password on the login page.

[1456] The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[1457] Once logged in, users can select either "Create a new meeting" or "Join an existing meeting" on the interface.

[1458] When creating a meeting, the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user can join by entering the meeting ID.

[1459] 2. Collection of audio data and sequential transcription

[1460] The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[1461] The acquired audio data is sent to a speech recognition API and converted into text data. The server then sequentially saves the converted text data to a database.

[1462] 3. Extracting the key points of the discussion

[1463] The server sends the stored text data to a natural language processing (NLP) model at regular intervals to extract important keywords and phrases.

[1464] The extracted key points are organized by the server and saved as data for meeting minutes.

[1465] 4. Creating and sharing meeting minutes

[1466] The server generates meeting minutes summarizing the discussion based on the key points.

[1467] The generated meeting minutes are converted into text files or HTML format and displayed on the user's device via Zoom's chat function or a dedicated interface.

[1468] 5. Proposal for the direction of the next discussion

[1469] The server analyzes the content of the discussion and uses an NLP model to identify the next topic to discuss.

[1470] The proposal will be notified to the user's device, supporting the progress of the discussion.

[1471] 6. Gathering Feedback

[1472] Users can provide feedback on meeting minutes and proposals from their devices.

[1473] The feedback provided will be collected and stored by the server and used to improve the system.

[1474] Specific example

[1475] For example, consider a scenario where users A, B, and C discuss the marketing strategy for a new product via Zoom. Using this system, the following operations can be performed:

[1476] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[1477] The server collects audio data from the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[1478] The system extracts key points such as "identifying the target market" and "competitor analysis" from text data, and then automatically generates meeting minutes based on this information.

[1479] The generated meeting minutes are displayed in the Zoom chat, allowing all users to view them in real time.

[1480] The server proposes "Pricing for the new product" as the next topic to discuss and notifies the terminal of this.

[1481] If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server saves this and uses it for future analyses.

[1482] In this way, the system of the present invention enables efficient and in-depth discussions by summarizing the key points of the discussion in real time and suggesting the next topics to be discussed.

[1483] The following describes the processing flow.

[1484] Step 1:

[1485] The user accesses the system's login page from their terminal. The user enters their username and password and completes the login process.

[1486] Step 2:

[1487] The server compares the user's authentication information with the database, and if correct, creates a session and returns a session ID to the user.

[1488] Step 3:

[1489] The user selects either "Create a new meeting" or "Join an existing meeting" on the terminal interface. If the user selects "Create a new meeting," the server generates a meeting ID and notifies the user of it.

[1490] Step 4:

[1491] When joining an existing meeting, the user enters the meeting ID, and the server authenticates that ID. Once the user joins the meeting, real-time audio data collection begins.

[1492] Step 5:

[1493] The server uses the ZOOM API to acquire audio data during the meeting in real time. The collected audio data is then sent to the server.

[1494] Step 6:

[1495] The server continuously sends audio data to a speech recognition API (e.g., Google Speech-to-Text API), converting the audio data into text data in real time. The converted text data is then stored in a database.

[1496] Step 7:

[1497] The server sends the stored text data to a natural language processing (NLP) model at regular intervals to extract important keywords and phrases.

[1498] Step 8:

[1499] The server organizes the extracted key points chronologically and saves them as data for meeting minutes.

[1500] Step 9:

[1501] The server generates meeting minutes in real time based on the stored key points data. The generated minutes are then instructed to be displayed on the device via Zoom's chat function or a dedicated interface.

[1502] Step 10:

[1503] The server analyzes the discussion and uses an NLP model to identify the next topic to discuss. The identified topic is generated as a suggestion and notified to the user's terminal.

[1504] Step 11:

[1505] Users input feedback on meeting minutes and proposals through the terminal's interface.

[1506] Step 12:

[1507] The server stores the collected feedback in a database. This stored feedback is then used as training data to improve the system's natural language processing model.

[1508] (Example 1)

[1509] 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".

[1510] In traditional meetings and discussions, creating meeting minutes was cumbersome, making it difficult to summarize key points of important discussions in real time. Furthermore, it was impossible to suggest future directions for discussions, and it was difficult to effectively collect participant feedback and use it for improvement. This resulted in decreased meeting efficiency and the risk of important discussions being overlooked.

[1511] 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.

[1512] In this invention, the server includes an interface means accessible to users, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for analyzing the content of the discussion and suggesting what should be discussed next, means for collecting and saving user feedback, means for using a communication service application programming interface for collecting audio data, means for using a speech recognition application programming interface to convert the audio data into text data, and means for using a natural language processing model for extracting main points. This makes it possible to automatically generate meeting minutes in real time, support the progress of meetings by suggesting what should be discussed next, and effectively collect feedback from participants to improve the system.

[1513] "Interface means" refers to devices and software that allow users to access and operate a system.

[1514] "Audio data" refers to the digital data of audio signals collected during meetings or discussions.

[1515] "Means of real-time collection" refers to equipment and software for instantly collecting audio data during a meeting.

[1516] "Text data" refers to character data generated by transcribing audio data.

[1517] "Methods for extracting and organizing the key points of an argument" refers to natural language processing models and algorithms that find important information from text data and organize it systematically.

[1518] "Means for generating and displaying meeting minutes" refers to software that automatically creates meeting minutes based on extracted key points and presents them to the user.

[1519] "Means of proposing a direction for discussion" refers to a function that analyzes the content of the current discussion and indicates the points that should be discussed next.

[1520] "Means for collecting and saving feedback" refers to a system for collecting user opinions and correction requests and saving them for later use.

[1521] "Application programming interface for communication services" refers to the means of utilizing functions through the programming interface provided by communication services.

[1522] A "Speech Recognition Application Programming Interface" refers to a programming interface that provides speech recognition functionality for converting speech data into text.

[1523] A "natural language processing model" refers to machine learning models and algorithms used to understand and analyze human language.

[1524] The system of this invention collects audio data in real time during meetings and discussions and transcribes it sequentially. It then extracts and organizes the key points of the discussion to automatically generate meeting minutes. Furthermore, it proposes directions for future discussions and collects and stores user feedback. This system is realized through the cooperation of the user, server, and terminal.

[1525] Key components of the system

[1526] Interface means: A device that allows a user to access and operate a system (e.g., a personal computer, tablet, or smartphone).

[1527] Audio data collection means: A communication service application programming interface (e.g., a communication service API) for collecting conference audio data in real time.

[1528] Audio data conversion means: A speech recognition application programming interface (e.g., speech recognition API) for converting collected audio data into text data.

[1529] Key point extraction method: A natural language processing model for extracting and organizing the key points of a discussion from text data.

[1530] Meeting minutes generation method: Software for generating and displaying meeting minutes based on key points.

[1531] A means of suggesting the direction of discussion: A function for analyzing the content of a discussion and suggesting what should be discussed next.

[1532] Feedback collection method: Software for collecting and storing user feedback and correction requests.

[1533] System operation

[1534] The user accesses the system from their terminal and enters their username and password on the login page. The server verifies the authentication information against the database, and if successful, creates a session and returns a session ID to the user. After logging in, the user selects either "Create a new meeting" or "Join an existing meeting." The server generates a meeting ID and notifies the user of it.

[1535] During the meeting, the server uses a communication service API to acquire audio data in real time. The acquired audio data is sent to a speech recognition API and converted into text data. The server then sequentially saves the converted text data to a database.

[1536] The server sends the stored text data to a natural language processing model at regular intervals to extract important keywords and phrases. The extracted key points are then organized by the server and stored as data for meeting minutes.

[1537] The server generates meeting minutes summarizing the discussion based on key points. The generated minutes are converted into text files or HTML format and displayed on the user's device through the communication service's chat function or a dedicated interface.

[1538] Furthermore, the server analyzes the content of the discussion and uses a natural language processing model to identify the next topic to discuss. This suggestion is then communicated to the user's terminal to support the progress of the discussion.

[1539] Users can provide feedback on meeting minutes and proposals from their devices. The feedback provided is collected and stored by the server and used to improve the system.

[1540] Specific example

[1541] For example, consider a scenario where users A, B, and C discuss a new product's marketing strategy through a communication service. Using this system, the following operations can be performed:

[1542] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[1543] The server collects audio data from the meeting via a communication service API and converts statements such as "We will discuss the target market for the new product" into text in real time.

[1544] The system extracts key points such as "identifying the target market" and "competitor analysis" from text data, and then automatically generates meeting minutes based on this information.

[1545] The generated meeting minutes are displayed in the communication service's chat, allowing all users to view them in real time.

[1546] The server proposes "Pricing for the new product" as the next topic to discuss and notifies the terminal of this.

[1547] If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server saves this and uses it for future analyses.

[1548] Example of a prompt

[1549] "We've developed a system that collects meeting audio data in real time and transcribes it sequentially. It also has a function to identify and suggest the next topic of discussion. For example, if we're discussing the marketing strategy for a new product, and topics like the target market and competitor analysis come up, how can you tell me how to suggest the next topic to discuss?"

[1550] In this way, the system of the present invention enables efficient and in-depth discussions by summarizing the key points of the discussion in real time and suggesting the next topics to be discussed.

[1551] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1552] Step 1: User login and meeting setup

[1553] Specific operation: The user accesses the system from their terminal and enters their username and password on the login page.

[1554] Enter: Username and password.

[1555] Data processing / calculation: The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[1556] Output: Session ID.

[1557] After the user logs in, they can select either "Create a New Meeting" or "Join an Existing Meeting" on the interface. If a meeting is created, the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user enters the meeting ID.

[1558] Step 2: Audio data collection and sequential transcription

[1559] Specific operation: The server uses a communication service API to acquire audio data during the meeting in real time.

[1560] Input: Voice data obtained from a communication service API.

[1561] Data processing / calculation: The server sends the acquired audio data to a speech recognition API and converts it into text data. The server then sequentially saves the converted text data to a database.

[1562] Output: Text data.

[1563] Step 3: Extract the key points of the discussion

[1564] Specific operation: The server sends the stored text data to a natural language processing model at regular intervals, and extracts important keywords and phrases.

[1565] Input: Text data.

[1566] Data Processing / Calculation: A natural language processing model analyzes the text data and extracts important keywords and phrases. The server organizes the extracted key points and saves them as data for meeting minutes.

[1567] Output: Key data.

[1568] Step 4: Generate and share meeting minutes

[1569] Specific operation: The server generates meeting minutes summarizing the discussion based on key points.

[1570] Input: Key data.

[1571] Data Processing / Calculation: Automatically creates meeting minutes based on key data and converts them to text files or HTML format.

[1572] Output: Meeting minutes file (text file / HTML format).

[1573] The generated meeting minutes are displayed on the user's device through the communication service's chat function or a dedicated interface.

[1574] Step 5: Propose the direction of the next discussion

[1575] Specific operation: The server analyzes the content of the discussion and uses a natural language processing model to identify the next topic to discuss.

[1576] Input: Meeting minutes data.

[1577] Data processing / computation: A natural language processing model is used to identify the next topic to discuss. The server then notifies the user's terminal of the identified topic as a suggestion.

[1578] Output: Suggestions for the next topic to discuss.

[1579] Step 6: Gathering Feedback

[1580] Specific functionality: Users can provide feedback on meeting minutes and proposals from their devices.

[1581] Input: User feedback.

[1582] Data processing / calculation: The server collects and stores user feedback, which is later used to improve the system.

[1583] Output: Saved feedback data.

[1584] In this way, a series of processes can be realized in which the server, terminal, and user cooperate with each other to summarize the key points of the discussion in real time and propose the next topics to be discussed.

[1585] (Application Example 1)

[1586] 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".

[1587] In modern meetings and discussions, creating meeting minutes, summarizing key points, and proposing directions for future discussions are crucial, yet extremely time-consuming. Therefore, there is a need for a system that collects audio data in real time, automatically generates meeting minutes, and efficiently proposes directions for future discussions. Furthermore, in environments requiring rapid decision-making, such as logistics centers, the implementation of an efficient meeting management system is urgently needed.

[1588] 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.

[1589] In this invention, the server includes an interface means accessible to the user, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for proposing the direction of future discussions, means for collecting and saving feedback, communication means for acquiring audio data during a meeting, means for transmitting the acquired audio data to a speech recognition platform and generating sequential transcript data, means for transmitting the converted text data to a natural language processing model and extracting important keywords and phrases, means for proposing topics to be discussed next based on the content of the discussion, and means for improving the system's performance using feedback. This enables real-time audio collection, automatic generation of meeting minutes on the spot, efficient and rapid progress of discussions and organization of main points, and proposals for topics to be discussed next.

[1590] "Interface means" refers to devices or software that users can access and operate.

[1591] "Means of collecting audio data in real time" refers to devices or software for instantly acquiring audio during meetings or discussions.

[1592] "Means for converting audio data into text data" refers to devices or software that convert collected audio data into textual information.

[1593] "Methods for extracting and organizing the key points of an argument" refer to techniques for finding important keywords and phrases from text data and organizing them systematically.

[1594] "Means for generating and displaying meeting minutes" refers to technology that records the content of a meeting based on extracted key points and provides it to the user.

[1595] "Methods for proposing the direction of future discussions" refers to techniques that suggest topics or themes to be discussed next, based on the current discussion.

[1596] "Means for collecting and storing feedback" refers to technologies that record user opinions and requests for later analysis and improvement.

[1597] "Communication means for acquiring audio data during a meeting" refers to technology for transmitting audio data from a meeting to a server.

[1598] A "speech recognition platform" refers to speech recognition technologies and APIs used to convert speech data into text data.

[1599] A "natural language processing model" is an artificial intelligence technology that analyzes the meaning and content of text data and extracts key points.

[1600] "A means of suggesting the next topic to be discussed" refers to a technology that analyzes the current discussion and suggests themes or topics to be addressed next to the user.

[1601] "Means for improving system performance" refer to technologies that analyze collected feedback and data to improve the overall efficiency and accuracy of the system.

[1602] This invention is a system that automatically generates meeting minutes at sites such as logistics centers, efficiently extracts the key points of the discussion, and suggests topics to be discussed next.

[1603] System Configuration

[1604] This system consists of the following main components:

[1605] User interface means: Devices and software (e.g., personal computers, mobile devices, smartphones) that allow users to access and operate a system.

[1606] Means of collecting audio data in real time: These are devices or software for instantly acquiring audio data from meetings (e.g., software or hardware that collects audio remotely using communication means).

[1607] Means of converting audio data into text data: These are devices or software that convert collected audio data into text information (e.g., Google Speech-to-Text API).

[1608] A means of extracting and organizing the key points of an argument: This refers to techniques for finding important keywords and phrases from text data and organizing them systematically (e.g., Google Natural Language API).

[1609] A means of generating and displaying meeting minutes: This technology records the content of a meeting based on extracted key points and provides it to the user.

[1610] A means of proposing the direction of future discussions: This is a technique that suggests topics or themes to be discussed next, based on the content of the current discussion.

[1611] A means of collecting and storing feedback: This is a technology for recording user opinions and requests and using them later for analysis and improvement.

[1612] System operation

[1613] 1. Starting the meeting and collecting audio data:

[1614] Users access the system through an interface and initiate a new meeting. During the meeting, audio data is collected in real time using communication methods.

[1615] 2. Converting audio to text data:

[1616] The collected audio data is sent to a speech recognition platform (e.g., Google Speech-to-Text API), where sequential transcription data is generated. The server stores the converted text data in a database.

[1617] 3. Extracting the key points of the discussion:

[1618] The server sends the stored text data to a natural language processing model (e.g., Google Natural Language API) at regular intervals to extract important keywords and phrases. The extracted key points are then organized by the server and stored as data for meeting minutes.

[1619] 4. Generating and displaying meeting minutes:

[1620] The server generates meeting minutes summarizing the discussion based on key points. The generated minutes are converted into text files or HTML format and displayed to the user through an interface.

[1621] 5. Proposed direction for the next discussion:

[1622] The server analyzes the discussion content and uses a natural language processing model to identify the next topic to discuss. The suggested topics are communicated to the user to support the progress of the discussion.

[1623] 6. Collecting and saving feedback:

[1624] Users can provide feedback on meeting minutes and proposals through the interface. This feedback is collected by the server and used to improve system performance.

[1625] Specific example

[1626] Consider a scenario where a manager at a logistics center uses this system to hold a meeting. The manager starts the meeting using their smartphone, and on-site staff participate. Audio data is collected in real time and converted into text data using the Google Speech-to-Text API. Key points are extracted using the Google Natural Language API, and meeting minutes are automatically generated and displayed. Furthermore, the system suggests topics to discuss next, allowing the manager to conduct the meeting efficiently. A feedback function collects user opinions, contributing to the improvement of the system's performance.

[1627] Example of a prompt

[1628] "We are developing an application to support logistics management meetings. We would like to implement the following functions using audio data collected during the meetings:

[1629] 1. Collect audio data in real time and convert it into text data.

[1630] 2. Extract important keywords and phrases from the converted text data and automatically generate meeting minutes.

[1631] 3. Based on the extracted data, propose the next topic to discuss.

[1632] The application will be implemented using speech recognition APIs and natural language processing APIs. We would appreciate your suggestions for this system.

[1633] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1634] Step 1:

[1635] The user initiates a new meeting using the interface. A meeting start request is sent to the server. The server generates a new meeting ID and returns it to the user. This completes the meeting preparation. The input is the user's request, and the output is the new meeting ID.

[1636] Step 2:

[1637] When the meeting begins, the terminal starts collecting audio data in real time. The collected audio data is sent to the server via a communication method. The server prepares to process the received audio data sequentially. The input is real-time audio data, and the output is the transmitted audio data.

[1638] Step 3:

[1639] The server sends audio data to a speech recognition platform (e.g., Google Speech-to-Text API) and converts it into text data. Once the audio data is converted to text data, the server stores it in a database. The input is audio data, and the output is text data.

[1640] Step 4:

[1641] The server sends the converted text data to a natural language processing model (e.g., Google Natural Language API) at regular intervals to extract important keywords and phrases. The extracted key points are organized by the server and saved for meeting minutes. The input is text data, and the output is important keywords and phrases.

[1642] Step 5:

[1643] The server automatically generates meeting minutes based on the extracted key points. The generated minutes are converted into text files or HTML format and displayed on the terminal via a user interface. Users can review the minutes in real time. Input consists of important keywords and phrases, while output is the meeting minutes data.

[1644] Step 6:

[1645] The server uses a natural language processing model to suggest the next topic to discuss, based on the generated meeting minutes data. The suggestions are then communicated to the user's terminal to support the discussion. The input is the meeting minutes data, and the output is the next topic to discuss.

[1646] Step 7:

[1647] After the meeting, users provide feedback using an interface. The server collects and stores this feedback. The collected feedback data is later used to improve the system. Input is user feedback, and output is data and improvement suggestions related to system performance improvements.

[1648] 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.

[1649] This invention combines a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggests directions for future discussions, with an emotion engine that recognizes the user's emotions. By reflecting the user's emotions in real time, this system supports the progress of deeper and more appropriate discussions.

[1650] System Configuration

[1651] The system consists of the following main components:

[1652] User interface means: A device (e.g., personal computer, tablet, smartphone) that allows a user to access and operate a system.

[1653] Audio data collection method: This refers to software or hardware (e.g., ZOOM API) that collects meeting audio data in real time.

[1654] Audio data conversion means: This refers to software (e.g., Google Speech-to-Text API) for converting collected audio data into text data.

[1655] Emotion recognition means: This is an engine (e.g., emotion recognition API) that analyzes a user's emotions from voice data or text data.

[1656] Key Point Extraction Method: This is a natural language processing (NLP) model for extracting and organizing the key points of an argument from converted text data and sentiment data.

[1657] Meeting minutes generation method: This is software that generates meeting minutes based on key points and sentiment data, and displays them to the user.

[1658] A means of suggesting the direction of discussion: This is an analytical function for suggesting the next topics to be discussed.

[1659] Feedback collection method: This is software for collecting and saving user feedback and requests for improvements.

[1660] System operation

[1661] 1. User login and meeting setup

[1662] The user accesses the system from their device and enters their username and password on the login page.

[1663] The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[1664] Once logged in, users can select either "Create a new meeting" or "Join an existing meeting" on the interface.

[1665] When creating a meeting, the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user can join by entering the meeting ID.

[1666] 2. Collection of audio data and sequential transcription

[1667] The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[1668] The acquired audio data is sent to a speech recognition API and converted into text data. The server then stores the converted text data in a database.

[1669] 3. Recognition of emotions

[1670] The server sends the collected voice and text data to an emotion recognition API to analyze the user's emotions.

[1671] The emotion recognition API returns the emotional state (e.g., joy, anger, sadness) for each statement, and the server stores this information in a database.

[1672] 4. Extracting the key points of the discussion

[1673] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations.

[1674] The extracted key points and sentiment data are organized by the server and stored as data for meeting minutes.

[1675] 5. Creating and sharing meeting minutes

[1676] The server generates meeting minutes summarizing the discussion based on key points and sentiment data.

[1677] The generated meeting minutes are converted into text files or HTML format and displayed on the user's device via Zoom's chat function or a dedicated interface.

[1678] 6. Proposal for the direction of the next discussion

[1679] The server analyzes the content and sentiment data of the discussion and uses an NLP model to identify the next topic to discuss.

[1680] The proposal will be notified to the user's device, supporting the progress of the discussion.

[1681] 7. Gathering Feedback

[1682] Users can input meeting minutes and feedback on proposals through the terminal's interface.

[1683] The feedback provided will be collected and stored by the server and used to improve the system.

[1684] Specific example

[1685] For example, consider a scenario where users A, B, and C discuss the marketing strategy for a new product via Zoom. Using this system, the following operations can be performed:

[1686] Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[1687] The server collects audio data from the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[1688] Simultaneously, the server uses an emotion recognition API to analyze the emotional state of each statement and obtains information such as, "User A is excited when talking about the target market."

[1689] Key points such as "identifying the target market" and "competitor analysis" are extracted from text data and sentiment data, and meeting minutes are automatically generated based on this information.

[1690] The generated meeting minutes are displayed in the Zoom chat, allowing all users to view them in real time.

[1691] The server proposes "Pricing for the new product" as the next topic to discuss and notifies the terminal of this.

[1692] If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server saves this and uses it for future analyses.

[1693] In this way, the system of the present invention enables efficient and in-depth discussions by summarizing the key points of the discussion and the users' feelings in real time and suggesting the next topics to discuss.

[1694] The following describes the processing flow.

[1695] Step 1:

[1696] The user accesses the system's login page from their terminal. The user enters their username and password and completes the login process.

[1697] Step 2:

[1698] The server compares the user's authentication information with the database, and if authentication is successful, it creates a session and returns a session ID to the user.

[1699] Step 3:

[1700] The user selects either "Create a new meeting" or "Join an existing meeting" on the terminal interface. If the user selects "Create a new meeting," the server generates a meeting ID and notifies the user of it.

[1701] Step 4:

[1702] When joining an existing meeting, the user enters the meeting ID, and the server authenticates that ID. Once the user joins the meeting, real-time audio data collection begins.

[1703] Step 5:

[1704] The server uses the ZOOM API to acquire audio data during the meeting in real time. The collected audio data is then sent to the server.

[1705] Step 6:

[1706] The server continuously sends audio data to a speech recognition API (e.g., Google Speech-to-Text API), converting the audio data into text data in real time. The converted text data is then stored in a database.

[1707] Step 7:

[1708] The server simultaneously sends audio and text data to an emotion recognition API to analyze the emotions contained in the user's statements. The emotion recognition API returns the emotional state for each statement (e.g., joy, anger, sadness), and the server stores this in a database.

[1709] Step 8:

[1710] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations.

[1711] Step 9:

[1712] The server organizes the extracted key points and sentiment data chronologically and saves them as data for meeting minutes.

[1713] Step 10:

[1714] The server generates meeting minutes in real time based on stored key points and sentiment data. The generated minutes are then instructed to be displayed on the device via Zoom's chat function or a dedicated interface. Sentiment information is also included in the meeting minutes.

[1715] Step 11:

[1716] The server analyzes the content and sentiment data of the discussion and uses an NLP model to identify the next topic to discuss. The identified topic is generated as a suggestion and notified to the user's terminal.

[1717] Step 12:

[1718] Users input feedback on meeting minutes and proposals through the terminal's interface.

[1719] Step 13:

[1720] The server stores the collected feedback in a database. This stored feedback is then used as training data to improve the system's natural language processing model.

[1721] (Example 2)

[1722] 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".

[1723] In today's business environment, effective meeting management is crucial. However, in many meetings, creating meeting minutes is time-consuming, making it difficult to summarize the key points of the discussion. Furthermore, it's challenging to understand participants' emotions and support the flow of the discussion. As a result, meaningful discussions fail, and meeting efficiency suffers. Additionally, there's the challenge of not being able to clearly define the direction the discussion should take next.

[1724] 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.

[1725] In this invention, the server includes an interface means accessible to users, means for collecting audio data in real time, means for converting the collected audio data into text data, means for recognizing emotions from the text data and audio data, means for extracting and organizing the main points of the discussion from the text data and emotion data, means for generating and displaying meeting minutes based on the main points and emotion data, means for proposing the direction of future discussions, and means for collecting and saving feedback. This makes it possible to grasp the main points of the discussion and the emotions of the participants in real time and propose the direction of the next discussion.

[1726] An "interface means" refers to a screen or device that allows a user to access and operate a system.

[1727] "Methods for real-time collection" refers to software and hardware systems for instantly acquiring audio data from meetings.

[1728] "Methods for converting audio data to text data" refers to software that uses speech recognition technology to convert audio data into text data.

[1729] "Means of recognizing emotions" refer to engines and algorithms that analyze voice and text data to identify a user's emotional state.

[1730] "Methods for extracting and organizing the key points of an argument" refer to natural language processing technologies that automatically identify important keywords and phrases and systematically summarize the points of contention.

[1731] "Methods for generating and displaying meeting minutes" refers to software that summarizes the content of a meeting based on extracted key points and sentiment data, and displays it in document format.

[1732] "Methods for suggesting direction" refers to algorithms that automatically determine and suggest the next topic to be discussed based on the content of the discussion and sentiment data.

[1733] A "means for collecting and saving feedback" refers to a system that records user opinions and requests for corrections in a database so that they can be used later.

[1734] This invention combines a system that automatically generates meeting minutes by collecting audio data from meetings and discussions in real time, transcribing it sequentially, extracting, organizing, and displaying the key points of the discussion, and further suggests directions for future discussions, with an emotion engine that recognizes the user's emotions. By reflecting the user's emotions in real time, this system supports the progress of deeper and more appropriate discussions.

[1735] System Configuration

[1736] This system consists of the following main components:

[1737] 1. Interface means: A device (e.g., personal computer, tablet, smartphone) that allows the user to access and operate the system.

[1738] 2. Means of real-time collection: This refers to software or hardware (e.g., video conferencing APIs) that collect meeting audio data in real time.

[1739] 3. Means for converting audio data to text data: This refers to software (e.g., speech recognition API) for converting collected audio data into text data.

[1740] 4. Means of recognizing emotions: This is an engine that analyzes the user's emotions from voice data or text data (e.g., emotion recognition API).

[1741] 5. Means for extracting and organizing the key points of the discussion: This is a natural language processing (NLP) model for extracting and organizing the key points of the discussion from the converted text data and sentiment data.

[1742] 6. Means for generating and displaying meeting minutes: This is software for generating meeting minutes based on key points and sentiment data, and displaying them to the user.

[1743] 7. Means for proposing direction: This is an analytical function for proposing what should be discussed next.

[1744] 8. Means for collecting and saving feedback: This is software for collecting and saving user feedback and requests for improvements.

[1745] Specific examples of system operation

[1746] For example, consider a scenario where users A, B, and C discuss a new product's marketing strategy via video conference. Using this system, the following operations and actions can be performed:

[1747] 1. Users A, B, and C log in to the system from their respective devices, enter the meeting ID, and join the meeting.

[1748] 2. The server collects the audio data of the meeting and converts statements such as "We will discuss the target market for the new product" into text in real time.

[1749] 3. Simultaneously, the server uses an emotion recognition API to analyze the emotional state of each statement and obtains information such as, "User A is excited when talking about the target market."

[1750] 4. Key points such as "identifying the target market" and "competitor analysis" are extracted from text data and sentiment data, and meeting minutes are automatically generated based on this.

[1751] 5. The generated meeting minutes are immediately displayed on the user's device via the video conference chat function or a dedicated interface.

[1752] 6. The server proposes "Pricing for the new product" as the next topic to be discussed and notifies the terminal of this.

[1753] 7. If User B enters "I would like to discuss the age group of the target market in more detail" as feedback, the server will save this and use it for future analyses.

[1754] Example of a prompt

[1755] Examples of prompts to input into a generative AI model include the following:

[1756] Please prepare the minutes for the new product marketing strategy meeting. The topics discussed were as follows: The target market for the new product was discussed, including target market identification and competitive analysis. User emotions such as excitement, anticipation, and concern were observed. Pricing for the new product was proposed as the next topic to be discussed.

[1757] By providing specific context in this way, AI models can generate more appropriate meeting minutes.

[1758] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1759] Step 1: User login and meeting setup

[1760] Enter: Username and password

[1761] Specific actions: The user opens a browser on their device and accesses the system's login page. There, they enter their username and password and click the "Login" button.

[1762] Data processing: The terminal sends the input information to the server. The server compares the authentication information with the database, and if there is a match, authentication is considered successful. At this point, a session ID is generated.

[1763] Output: The server returns the generated session ID to the user's terminal, and the user becomes logged in.

[1764] For example, when a user selects "Create New Meeting," the server generates a meeting ID and notifies the user of it. To join an existing meeting, the user enters the meeting ID to complete the joining process.

[1765] Step 2: Audio data collection and sequential transcription

[1766] Input: Meeting audio data

[1767] Specific operation: When a user starts a meeting, the server collects audio data in real time using the specified API (e.g., a video conferencing API).

[1768] Data processing: The collected audio data is sent to a speech recognition API and converted into text data.

[1769] Output: The converted text data is saved to the database by the server.

[1770] For example, if the audio data is "Discussing the target market for a new product," the speech recognition API will return this as text data.

[1771] Step 3: Recognizing Emotions

[1772] Input: Converted text data and audio data

[1773] Specific operation: The server sends the collected audio and text data to the specified emotion recognition API.

[1774] Data processing: The emotion recognition API analyzes the emotional state (e.g., joy, anger, sadness, etc.) in each statement and returns the result.

[1775] Output: The server saves the emotion data returned from the emotion recognition API to the database.

[1776] Example: If the emotion recognition API analyzes that User A is "excited" while talking about the target market, that information will be saved.

[1777] Step 4: Extract the key points of the discussion

[1778] Input: Saved text data and sentiment data

[1779] Specific operation: The server sends this data to a natural language processing (NLP) model and starts the extraction process.

[1780] Data processing: The NLP model identifies and extracts important keywords and phrases.

[1781] Output: The extracted key points data, along with sentiment data, is organized by the server and saved as data for meeting minutes.

[1782] Example: The NLP model extracts key points such as "identifying the target market" and "conducting competitive analysis."

[1783] Step 5: Generate and share meeting minutes

[1784] Input: Key point data and sentiment data

[1785] Specific operation: Based on this data, the server generates meeting minutes summarizing the meeting content.

[1786] Data processing: The generated meeting minutes are converted into text files or HTML format.

[1787] Output: Meeting minutes are displayed on the user's device via the video conference chat function or a dedicated interface.

[1788] Example: All users will be able to view meeting minutes in real time.

[1789] Step 6: Propose the direction of the next discussion

[1790] Input: Discussion content and sentiment data

[1791] Specific operation: The server analyzes this data and uses an NLP model to identify the next topic to discuss.

[1792] Data processing: Determine the topic proposed by the NLP model.

[1793] Output: The proposed topic information will be notified to the user's terminal.

[1794] Example: The server suggests "Pricing for the new product" as the next discussion topic.

[1795] Step 7: Gathering Feedback

[1796] Input: User feedback information

[1797] Specific operation: Users input feedback on meeting minutes and proposals using the terminal's interface.

[1798] Data processing: The provided feedback information is collected and recorded by the server.

[1799] Output: Feedback information is stored in a database and used for future system improvements.

[1800] For example, if User B provides feedback such as "I would like to discuss the age range of the target market in more detail," that information will be saved.

[1801] (Application Example 2)

[1802] 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".

[1803] Traditional methods of recording meetings and discussions often involve manual transcription of audio data and extraction of key points, which is time-consuming and labor-intensive. Furthermore, it is difficult to consider the emotional state of participants during discussions, and there is a lack of means to support the proper progression of the discussion, resulting in inconsistent quality of meeting minutes and insufficient suggestions for the direction of future discussions.

[1804] 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.

[1805] In this invention, the server includes an interface means accessible to the user, means for collecting audio data in real time, means for converting the collected audio data into text data, means for extracting and organizing the main points of the discussion from the text data, means for generating and displaying meeting minutes based on the main points, means for proposing the direction of future discussions, means for collecting and saving feedback, emotion recognition means for recognizing and displaying the user's emotions in real time, means for identifying the content to be discussed next based on the main points and emotion data, and means for sharing the generated meeting minutes. This makes it possible to efficiently extract and organize the main points of the discussion, automatically generate high-quality meeting minutes, and propose the direction of the next discussion. Furthermore, by reflecting the user's emotional state in real time, it is possible to support the progress of discussions more deeply and appropriately.

[1806] "Interface means" refers to a device or software that allows a user to access and operate a system.

[1807] "Means for collecting audio data in real time" refers to a system component that acquires audio data from meetings and discussions and transmits it to a server in real time.

[1808] "Means for converting audio data to text data" refers to software or algorithms that transcribe collected audio data and convert it into text format.

[1809] "Methods for extracting and organizing the main points of a discussion from text data" refers to system components that analyze converted text data, extract important keywords and phrases, and organize the main points of the discussion.

[1810] "A means of generating and displaying meeting minutes based on key points" refers to a system component that automatically creates meeting minutes based on extracted key points and displays them in an easy-to-understand manner for the user.

[1811] "Means for proposing the direction of future discussions" refers to a system component that analyzes the content of discussions and sentiment data to identify and propose topics that should be discussed next.

[1812] "Means for collecting and storing feedback" refers to system components for collecting feedback information such as opinions and requests for corrections from users and storing it in a database.

[1813] An "emotion recognition tool" is an engine or software that analyzes and recognizes emotions in real time from a user's voice or text data.

[1814] "Means for sharing generated meeting minutes" refers to a system component that provides the functionality to share generated meeting minutes with other users or systems.

[1815] This invention combines a system that collects audio data from meetings and discussions in real time, transcribes it sequentially to extract and organize the key points of the discussion, and automatically generates meeting minutes with an emotion engine that recognizes the user's emotions in real time. This system is composed of the following various means.

[1816] Hardware and software to be used

[1817] User interface means: A smartphone, tablet, or personal computer is used as a device for the user to access and operate the system.

[1818] Audio data collection method: To collect meeting audio data in real time, for example, use the ZOOM API.

[1819] Audio data conversion method: The Google Speech-to-Text API is used to convert the collected audio data into text data.

[1820] Emotion Recognition Method: To analyze user emotions from voice and text data, we use emotion recognition APIs (e.g., Microsoft Azure or AWS Rekognition).

[1821] Key points extraction method for discussions: Natural language processing (NLP) models (e.g., BERT or GPT-3) are used to extract and organize the key points of discussions from text data and sentiment data.

[1822] Meeting minutes generation method: Software is used to generate meeting minutes based on key points and sentiment data, and to display them to the user.

[1823] Methods for suggesting the direction of discussion: Utilize software or algorithms with analytical capabilities to suggest the next topics to be discussed.

[1824] Feedback collection method: Use an interface and database to collect and store user feedback and correction requests.

[1825] Meeting minutes sharing method: It has software functionality to allow generated meeting minutes to be shared with other users or systems.

[1826] Specific examples of how the system works

[1827] 1. Login and meeting setup:

[1828] Users access the system from devices such as smartphones or computers and enter their username and password on the login page. The server verifies the entered authentication information against the database, and if authentication is successful, it creates a session and returns a session ID to the user. After logging in, the user selects either "Create a new meeting" or "Join an existing meeting," and the server generates or authenticates a meeting ID.

[1829] 2. Collection and transcription of audio data:

[1830] The server uses the ZOOM API to acquire audio data in real time during meetings. The acquired audio data is converted into text data using the Google Speech-to-Text API, and the text data is stored in a database by the server.

[1831] 3. Recognition of emotions:

[1832] The server sends the collected audio and text data to an emotion recognition API to analyze the user's emotions. The emotion recognition API returns the emotional state for each statement (e.g., joy, anger, sadness), and the server stores this in a database.

[1833] 4. Extracting the main points of the discussion:

[1834] The server sends the stored text and sentiment data to a natural language processing (NLP) model to extract important keywords, phrases, and sentiment fluctuations. The extracted key points and sentiment data are then organized by the server and stored as data for generating meeting minutes.

[1835] 5. Creating and sharing meeting minutes:

[1836] The server generates meeting minutes summarizing the discussion based on key points and sentiment data. The generated minutes are converted into text files or HTML format and displayed on the user's device, as well as shared with other users through the sharing function.

[1837] 6. Proposed direction for the discussion:

[1838] The server analyzes the discussion content and sentiment data, and uses an NLP model to identify the next topic to discuss. Suggestions are notified to the user's terminal to support the progress of the discussion.

[1839] 7. Gathering feedback:

[1840] Users can input feedback on meeting minutes and proposals through the interface. The feedback provided is collected and stored by the server and used to improve the system.

[1841] Specific examples and prompt statements

[1842] For example, consider a scenario where store staff hold a meeting about how to display new products. Using this system, the following operations can be performed:

[1843] Staff members A and B log into the system from their respective devices, enter the meeting ID, and join the meeting.

[1844] The server collects audio data from the meeting and converts statements like "We will discuss how to display the new products" into text in real time.

[1845] At the same time, the server uses an emotion recognition API to analyze the emotional state of each statement, obtaining information such as, "Staff member A is excited when talking about the display method."

[1846] The system extracts key points such as "areas for improvement in display methods" and "customer reactions" from text data and sentiment data, and then automatically generates meeting minutes based on this information.

[1847] The generated meeting minutes can be shared with other staff members using the sharing function, allowing them to review them immediately.

[1848] The server suggests "Pricing for the new product" as the next topic to discuss and notifies the staff member's terminal.

[1849] Examples of specific prompt messages:

[1850] "Customer complaint about product quality."

[1851] "Next discussion topic: Improve customer complaint handling."

[1852] This makes it possible to efficiently extract key points of discussions, including user emotions, based on the form in which the invention is implemented, and to generate high-quality meeting minutes.

[1853] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1854] Step 1:

[1855] User login and meeting setup

[1856] Input: Users access the system from a device such as a smartphone or computer and enter their username and password.

[1857] Data processing / calculation: The server compares the entered authentication information with the database, and if authentication is successful, it creates a session and returns a session ID.

[1858] Output: Upon successful authentication, the session ID will be displayed on the user's device.

[1859] Specific operation: The user interacts with the interface and selects either "Create a new meeting" or "Join an existing meeting." If "Create a new meeting" is selected, the server generates a meeting ID and notifies the user.

[1860] Step 2:

[1861] Audio data collection and sequential transcription

[1862] Input: The server uses the ZOOM API to retrieve audio data from the meeting in real time.

[1863] Data processing / calculation: The acquired audio data is sent to a speech recognition API (Google Speech-to-Text API) and converted into text data.

[1864] Output: The converted text data is saved on the server.

[1865] Specific operation: The server updates text data in real time and displays it to the user on the interface.

[1866] Step 3:

[1867] Recognition of emotions

[1868] Input: The server sends collected audio and text data to the emotion recognition API.

[1869] Data Processing / Calculation: The emotion recognition API analyzes the emotional state (e.g., joy, anger, sadness, etc.) in each statement and returns it as emotion data.

[1870] Output: Emotional data is stored on the server.

[1871] Specific operation: The server displays an emotion indicator corresponding to each statement on the interface, allowing the user to visually confirm it.

[1872] Step 4:

[1873] Extracting the main points of the discussion

[1874] Input: The server sends stored text data and sentiment data to a natural language processing (NLP) model.

[1875] Data processing / calculation: The NLP model extracts important keywords and phrases and further analyzes emotional fluctuations.

[1876] Output: Extracted key points and sentiment data are stored on the server.

[1877] Specific operation: The server organizes the key points and displays them in real time on the interface as data for generating meeting minutes.

[1878] Step 5:

[1879] Creating and sharing meeting minutes

[1880] Input: The server summarizes the discussion based on key points and sentiment data.

[1881] Data processing / calculation: Automatically create meeting minutes using a meeting minutes generation algorithm.

[1882] Output: The generated meeting minutes are converted to a text file or HTML format.

[1883] Specific operation: Meeting minutes are displayed to the user on the interface and can be sent to other users via the sharing function.

[1884] Step 6:

[1885] Proposal for the direction of the discussion

[1886] Input: The server analyzes the content of the discussion and sentiment data.

[1887] Data processing / computation: Use an NLP model to identify the next topic to discuss.

[1888] Output: The proposed content is notified to the user's device.

[1889] Specific behavior: The user interface displays specific suggestions for the next topic to be discussed, such as "Pricing for the new product."

[1890] Step 7:

[1891] Gathering feedback

[1892] Input: The user provides feedback through the interface.

[1893] Data processing / calculation: Collect and store feedback information in a database.

[1894] Output: The collected feedback information will be used to improve the system in the future.

[1895] Specific operation: Based on user feedback, the system automatically adjusts areas for improvement and discussion topics for the next meeting.

[1896] 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.

[1897] 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.

[1898] 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.

[1899] 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.

[1900] 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.

[1901] 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.

[1902] 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.

[1903] 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 ...

Claims

1. An interface means that users can access, A means of collecting audio data in real time, A means of converting collected audio data into text data, A method for extracting and organizing the key points of a discussion from text data, A means of generating and displaying meeting minutes based on key points, A means of proposing the direction of future discussions, Means for collecting and storing feedback, A system that includes this.

2. The system according to claim 1, which sends collected audio data to a speech recognition API and generates sequential transcription data.

3. The system according to claim 1, which transmits text data to a natural language processing model and extracts key points.

Images