System

Abstract

A system is provided.SOLUTION: A system, comprising: an automation conference record generation unit; a participant speech volume tracking unit; a speech promotion unit; a deviation detection and discussion guide unit; and a key point visualization unit.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] In today's online meetings, manually creating meeting minutes is time-consuming and labor-intensive, and there are frequent issues with recording speech content and missing important points. Furthermore, the lack of diversity of opinions due to certain participants' participation can lead to a decrease in meeting efficiency as discussions stray from the planned topic. Furthermore, there is also the issue of it being difficult to visually grasp the key points of the meeting content. Technological solutions are needed to solve these issues and ensure efficient and productive online meetings. [Means for solving the problem]

[0005] The present invention provides an online conference system that includes an automated conference record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, and a key point visualization means, and solves the following problems by using these means:

[0006] 1. The automated meeting recording system uses speech recognition technology to transcribe what is said during meetings in real time and automatically records the results as minutes, eliminating the need to manually create minutes and preventing important comments and decisions from being overlooked.

[0007] 2. Participant speech volume tracking measures are used to track the number of times each participant speaks in real time. Participants who speak less are identified and automatically notified to speak, ensuring diversity of opinion and providing an environment where everyone can actively participate.

[0008] 3. Digression detection and discussion guidance uses natural language processing technology to determine in real time whether the discussion topic has strayed and suggests returning to the main topic, thereby maintaining meeting efficiency.

[0009] 4. Key Points Visualization allows meeting key points to be displayed on the whiteboard in real time, allowing participants to visually grasp the key points, making it easier for everyone to follow the flow of the discussion and improving meeting productivity.

[0010] By the above means, the present invention solves many of the problems associated with online meetings and enables meetings to be run efficiently and productively.

[0011] The "automated meeting record generator" is a function that transcribes audio during a meeting in real time and automatically records what is said and what decisions are made.

[0012] The "participant speech volume tracking means" is a function that tracks the number of speeches and speech duration of each participant in real time and accumulates the data.

[0013] The "speech promotion means" is a function that identifies participants who make few comments and automatically generates and sends notifications and questions to those participants to encourage them to speak.

[0014] The "digression detection and discussion guidance means" is a function that uses natural language processing technology to detect whether the topic of discussion has strayed from the main topic and makes suggestions to bring the discussion back on track.

[0015] "Key Points Visualization" is a function that displays important statements and decisions made during a meeting on a whiteboard or screen in real time, allowing them to be shared visually. [Brief explanation of the drawings]

[0016] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

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

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

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

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

[0021] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0022] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0023] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0024] [First embodiment]

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

[0026] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0027] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0028] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0029] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0030] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0031] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

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

[0033] 2, in the data processing device 12, a specific process 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" according to the technology of the present 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 process 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.

[0034] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0035] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0036] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0037] Initialization

[0038] When a meeting starts, the server creates an instance of the AI ​​facilitator. Next, it loads basic information about participants (such as their names, positions, and past speech data) from the database. It also initializes a minutes template and prepares the data structures necessary for the meeting to proceed.

[0039] Automatic meeting minutes creation

[0040] The device uses speech recognition technology to transcribe each statement in real time during the meeting. The transcribed content and speaker identification information are sent to the server as the meeting progresses. The server receives the data and automatically records it as meeting minutes.

[0041] Encouraging participants to speak up

[0042] The server tracks each participant's speech volume (number of times and duration) in real time. It identifies participants who speak less and sends them notifications encouraging them to speak. Notifications are automatically generated according to the progress of the conference and are delivered to users.

[0043] Derailment Detection and Discussion Guide

[0044] The server uses natural language processing technology to determine in real time whether a discussion has deviated from the planned topic. If a deviation from the topic is detected, the server generates suggestions to bring the discussion back on topic and notifies the user, thereby reducing the risk of the discussion dragging on unnecessarily or drifting off topic.

[0045] Visualization of key points

[0046] The device extracts important points and decisions from what is being said during the meeting and displays them on a whiteboard in real time. This whiteboard is constantly updated during the meeting, making it easier for users to visually follow the important points. This allows meetings to proceed efficiently without losing track of the flow of discussion or important points.

[0047] Specific examples

[0048] For example, let's say a company is holding a project progress meeting. When the meeting begins, the server launches the AI ​​facilitator and loads the data of the participants, "Tanaka," "Yamada," and "Sato." As the meeting progresses, if Tanaka says, "This project is going well," his remarks are transcribed in real time by the device. The transcription data and Tanaka's identification information are sent to the server and recorded in the minutes.

[0049] Next, tracking reveals that Yamada has barely spoken anything. The server sends Yamada a notification saying, "You haven't spoken much. Do you have any input?" This gives Yamada a chance to express his opinion.

[0050] If, at some point, Sato starts talking about recent news, the server will determine that the remark is unrelated to the project's progress and send the user a suggestion: "The discussion is veering off-topic. Let's return to the main agenda."

[0051] Furthermore, the device automatically displays important points in the meeting, such as "Project progress is going smoothly," on the whiteboard, allowing users to visually confirm important matters during the meeting in real time.

[0052] As described above, this system provides various functions to ensure that online meetings proceed efficiently and productively, contributing to improving the quality of meetings.

[0053] The processing flow will be explained below.

[0054] Step 1: Initialization

[0055] The server creates an instance of the AI ​​facilitator at the start of the meeting, which launches a system to assist in the progress of the online meeting.

[0056] Step 2: Load participant information

[0057] The server loads basic information of participants (such as name, position, and past comment data) from the database, which enables identification of each participant and support based on past data.

[0058] Step 3: Initialize the meeting minutes template

[0059] The server initializes a meeting template, which provides the data structures necessary for the minutes format and flow of the meeting.

[0060] Step 4: Start real-time transcription

[0061] As the meeting progresses, the device uses speech recognition technology to transcribe each utterance in real time, and the speech is automatically converted into text.

[0062] Step 5: Sending speech data

[0063] The device sends the transcribed speech and the speaker's identity to the server, where the speech is ready to be recorded.

[0064] Step 6: Update the minutes

[0065] The server automatically updates the meeting minutes based on the received speech data, ensuring that each speech is recorded appropriately.

[0066] Step 7: Track your speech volume

[0067] The server tracks each participant's speech volume (number of times and duration) in real time, allowing the activity level of participants to be monitored.

[0068] Step 8: Identify quiet participants

[0069] The server identifies participants who are less active: participants whose number of comments falls below a certain threshold.

[0070] Step 9: Send a prompt

[0071] The server generates a notification for quiet participants to speak up and sends it to the user, with a message such as "You haven't spoken much. Do you have any input?"

[0072] Step 10: Derailment detection

[0073] The server uses natural language processing technology to determine in real time whether the current comment is off-topic.

[0074] Step 11: Sending a discussion guide notification

[0075] If the server detects a digression, it generates a suggestion to bring the discussion back on topic and sends it to the user: "The discussion is veering off-topic. Let's return to the main agenda."

[0076] Step 12: Extracting key points

[0077] The device extracts important points and decisions from what is said during the meeting.

[0078] Step 13: Display on the Whiteboard

[0079] The device displays the extracted important points on a whiteboard in real time, allowing users to visually confirm the important points.

[0080] Step 14: End the meeting and export the minutes

[0081] At the end of the meeting, the server exports the accumulated minutes in an editable format, which users can review and modify.

[0082] Through these steps, this system enables online meetings to proceed efficiently and productively.

[0083] Example 1

[0084] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0085] Existing online meeting systems face the challenge of efficiently managing a wide range of tasks, including meeting progress, minutes creation, participant prompting, digression detection and discussion guidance, and visualization of key points. This can lead to poor meeting quality, redundant discussions, and participants being reluctant to speak up. Furthermore, creating meeting minutes manually is time-consuming and labor-intensive, and lacks accuracy.

[0086] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0087] In this invention, the server includes a meeting start detection means, an AI facilitator instance generation means, a participant information database loading means, a minutes template initialization means, a speech transcription means using speech recognition, a speech data transmission means, a speech data recording means, a participant speech volume tracking means, a speech promotion notification generation means, a speech promotion notification transmission means, a discussion digression detection means, a discussion guide notification generation means, a discussion guide notification transmission means, an important point extraction means, and a whiteboard display means. This makes it possible to manage the progress of the meeting, create minutes, encourage participants to speak, guide the discussion, and visualize important points in real time.

[0088] The "conference start detection means" is a means for recognizing the start of a conference and for the system to automatically start the necessary initial processing.

[0089] An "AI facilitator instance generation means" is a means for generating an instance of an AI model to smoothly proceed with a meeting.

[0090] The "participant information database loading means" is a means for loading basic information of conference participants from a database.

[0091] The "minutes template initialization means" is a means for initializing a template for creating minutes of a meeting.

[0092] A "means for transcribing speech using speech recognition" is a means for using speech recognition technology to transcribe speech during a meeting in real time.

[0093] The "utterance data transmission means" is a means for transmitting the transcribed utterance data to the server.

[0094] The "utterance data recording means" is a means for recording received utterance data in a database.

[0095] The "participant speech volume tracking means" is a means for tracking the speech volume of each participant during a conference in real time.

[0096] The "means for generating a speech prompting notification" is a means for generating a notification that prompts a participant who makes few comments to speak.

[0097] The "speech prompting notification sending means" is a means for sending the generated speech prompting notification to the relevant participant.

[0098] The "discussion deviation detection means" is a means for detecting whether the discussion in the meeting has deviated from the planned topic.

[0099] The "discussion guide notification generating means" is a means for generating a suggestion to bring the discussion back to the main topic when the discussion goes off track.

[0100] The "discussion guide notification sending means" is a means for sending the generated discussion guide notification to the participants.

[0101] The "means for extracting important points" is a means for extracting important points and decisions from the content of statements made during a meeting.

[0102] The "whiteboard display means" is a means for displaying the extracted important points on a whiteboard in real time.

[0103] The present invention is a system for conducting online meetings efficiently and productively. This system automates various processes from the start to the end of a meeting and combines multiple means to improve the quality of the meeting.

[0104] Initialization

[0105] When the server detects the start of a meeting, it creates an instance of an AI facilitator. This uses an AI framework such as TensorFlow or PyTorch. Next, it loads participant information from a database (e.g., MySQL or PostgreSQL). This includes participant names, positions, and past speech data. Next, the server initializes a minutes template. This template uses a predefined file in Markdown format or similar.

[0106] Automatic meeting minutes creation

[0107] During a meeting, the device uses speech recognition technology such as the Google Speech-to-Text API or IBM Watson Speech to Text to transcribe each utterance in real time. The transcribed utterances and the speaker's identity are then sent to the server via an HTTP POST request. The server receives this data and automatically records it in a database such as Elasticsearch.

[0108] Encouraging participants to speak up

[0109] The server uses MongoDB or similar to track the amount of speech each participant makes (number of times they make a comment and how long they speak). For participants who make little comments, the server uses the Slack API or Microsoft Teams API to generate and send notifications encouraging them to speak. These notifications are generated automatically and sent to users.

[0110] Derailment Detection and Discussion Guide

[0111] The server uses natural language processing techniques, such as Hugging Face's Transformers library, to determine in real time whether the discussion is deviating from the planned topic. If a deviation is detected, the server compares it with pre-defined agenda information, generates suggestions to bring the discussion back on topic, and communicates them to the user. This notification is sent automatically at the appropriate time.

[0112] Visualization of key points

[0113] The device uses natural language processing libraries such as SpaCy and NLTK to extract key points and decisions from what is being said during the meeting. The extracted key points are displayed in real time on a whiteboard service such as Miro Board or Microsoft Whiteboard. This display is constantly updated as the meeting progresses, making it easy for users to visually follow the key points.

[0114] Specific examples

[0115] For example, when a company holds a project status meeting, the server detects the start of the meeting through the Google Calendar API, creates an instance of an AI facilitator using TensorFlow, and then loads the data of the participants, "Tanaka," "Yamada," and "Sato," from a MySQL database.

[0116] As the meeting progresses, if Tanaka says, "This project is going well," the device will transcribe his speech in real time using the Google Speech-to-Text API. The transcription data and Tanaka's identification information will be sent to the server via an HTTP POST request and recorded in Elasticsearch.

[0117] Once MongoDB data tracks that Yamada has barely spoken, the server uses the Slack API to send a notification to Yamada saying, "You haven't spoken much. Do you have any input?"

[0118] If Sato starts talking about recent news midway through, the server will use Hugging Face's Transformers library to determine that his remarks are unrelated to the project's progress and send the user a suggestion: "The discussion is veering off-topic. Let's return to the main agenda."

[0119] The device automatically displays the key point of the meeting, such as "Project progress is going well," on the Miro board, allowing users to visually check important matters during the meeting in real time through the whiteboard.

[0120] Prompt Sentence Examples

[0121] Below is an example prompt for requesting minutes of a project status meeting:

[0122] A project progress meeting was held. The participants were Tanaka, Yamada, and Sato. Tanaka stated, "This project is progressing smoothly." Based on the minutes, please create and propose the following action items:

[0123] As described above, the present invention provides various functions for conducting online conferences efficiently and productively, thereby contributing to improving the quality of conferences.

[0124] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0125] Processing flow

[0126] Step 1:

[0127] The server detects when a meeting starts. To do this, it uses the Google Calendar API to monitor scheduled events. The input is the calendar information obtained from the API, and the output is a meeting start event.

[0128] Step 2:

[0129] The server creates an instance of the AI ​​facilitator. First, it loads the AI ​​model using the TensorFlow or PyTorch library and creates an instance. The input is the pre-trained AI model, and the output is the generated instance of the AI ​​facilitator.

[0130] Step 3:

[0131] The server loads participant information from a database. It issues SQL queries to a database such as MySQL or PostgreSQL to retrieve participant names, positions, and past comment data. The input is the database connection information and query, and the output is the retrieved participant information.

[0132] Step 4:

[0133] The server initializes the minutes template. It reads a predefined Markdown template file and prepares the initial state. The input is the path to the template file, and the output is the initialized minutes template.

[0134] Step 5:

[0135] The device transcribes speech during meetings in real time. It uses the Google Speech-to-Text API to convert audio data captured from the microphone into text data. The input is audio data, and the output is transcribed text data.

[0136] Step 6:

[0137] The device sends the transcribed speech data and the speaker's identification information to the server using an HTTP POST request. The input is the speech data and the identification information, and the output is the success status of the data transmission to the server.

[0138] Step 7:

[0139] The server records the received utterance data in Elasticsearch, analyzes the data, and adds it to the appropriate index. The input is the received utterance data, and the output is the index information stored in Elasticsearch.

[0140] Step 8:

[0141] The server tracks the amount of speech of each participant in real time. It records the speech data in MongoDB and executes aggregation queries to calculate the number of speeches and speech duration. The input is the received speech data, and the output is the tracked speech volume.

[0142] Step 9:

[0143] The server generates notifications to encourage participants who speak less to speak. The notification content is automatically generated using the Slack API or Microsoft Teams API. The input is the tracked speech volume data, and the output is the generated notification message.

[0144] Step 10:

[0145] The server sends the generated notification to the appropriate user using the Slack API or Microsoft Teams API. The input is the generated notification message, and the output is the success status of the notification sending.

[0146] Step 11:

[0147] The server detects deviations in the discussion. It uses Hugging Face's Transformers library to compare the content of comments with the agenda and determine whether or not a deviation has occurred. The input is the comment data and agenda information, and the output is the deviation detection result.

[0148] Step 12:

[0149] When a digression is detected, the server generates a suggestion to bring the discussion back to the main topic. The suggestion content is generated using a generative AI model. The input is the detected digression result, and the output is the generated suggestion message.

[0150] Step 13:

[0151] The server notifies the user of the proposal message. It sends the notification using the chat application's API. The input is the generated proposal message, and the output is the success status of sending the notification.

[0152] Step 14:

[0153] The device extracts important points from what is said during the meeting. It uses SpaCy and NLTK to analyze the speech and extract important points and decisions. The input is the speech data, and the output is the extracted important points.

[0154] Step 15:

[0155] The device displays the extracted key points on a whiteboard. It updates the display in real time using the APIs of Miro Board and Microsoft Whiteboard. The input is the extracted key points, and the output is the display result on the whiteboard.

[0156] Through the above processing steps, the system manages the progress of the meeting, creates minutes, encourages participants to speak up, directs the discussion, and visualizes important points.

[0157] (Application example 1)

[0158] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0159] Modern meeting systems lack features such as taking minutes, encouraging participants to speak up, and detecting derailed discussions. Some participants often don't speak much, or the discussion often goes off track, wasting time. Furthermore, participants can lose track of the flow of the discussion because they cannot visually track important points and decisions. These issues directly affect production efficiency, particularly in factory progress meetings, so there is a high need for a system that can quickly and efficiently conduct meetings.

[0160] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0161] In this invention, the server includes an automated meeting record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, a means for transcribing speeches made during a meeting in real time using speech recognition technology and recording the results, a means for sending notifications to participants who speak less to encourage them to speak, a means for determining in real time whether the discussion has deviated from the planned topic using natural language processing technology and generating suggestions, and a means for displaying key points and decisions on a whiteboard in real time, thereby enabling improvements in the efficiency and productivity of meetings.

[0162] The "automated meeting record generator" is a function that automatically recognizes statements made during a meeting and creates minutes in real time.

[0163] The "participant speech volume tracking means" is a function for measuring and recording the number of speeches and speech duration of each conference participant in real time.

[0164] The "speech promotion means" is a function for sending notifications to participants who do not speak much, encouraging them to speak.

[0165] "Derailment detection and discussion guidance means" is a function that uses natural language processing technology to determine whether a discussion has deviated from the planned topic and suggests an appropriate direction for the discussion.

[0166] The "important point visualization means" is a function that displays important points and decisions made during a meeting on a display device in real time, allowing participants to visually grasp important information.

[0167] "Speech recognition technology" is a technology for converting speech into text, and in this invention is a technology used to transcribe statements made during a meeting in real time.

[0168] "Real-time transcription means" is a function that uses voice recognition technology to instantly convert meeting speech into text and record it.

[0169] "Means for encouraging participants who speak infrequently to speak by sending notifications" is a function for tracking the volume of comments in real time and sending notifications to participants who speak infrequently to encourage them to speak.

[0170] "Natural language processing technology" is a technology for understanding and analyzing human language, and in this invention, it is a technology used to detect digressions in discussions and provide appropriate discussion guidance.

[0171] "A means of displaying important points and decisions on a whiteboard in real time" is a function that instantly reflects important statements and decisions made during a meeting on a display device such as a whiteboard, allowing all participants to visually confirm them.

[0172] The system embodying this invention is designed to streamline factory progress meetings and is composed of a server, terminals, and users with various functions. Specifically, it is composed of an automated meeting record generation means, a means for tracking the amount of speech by participants, a means for encouraging speech, a means for detecting deviations and guiding discussions, and a means for visualizing important points.

[0173] When a meeting starts, the server creates an instance of the AI ​​facilitator and loads basic information about participants (such as their names, positions, and past speech data) from a database. This database uses a system such as SQLite. It also initializes a minutes template and prepares the data structure necessary for the meeting to proceed.

[0174] The device uses speech recognition technology to transcribe speech in real time during a meeting. The transcribed speech and speaker identification information are sent to a server, which records the data as meeting minutes in real time. The speech recognition technology used here includes the speech_recognition library.

[0175] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them a notification to encourage them to speak. This notification is sent to an application on a smartphone or tablet.

[0176] The digression detection and discussion guidance mechanism uses natural language processing (NLP) technology to determine in real time whether a discussion has deviated from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user. The NLP technology used here includes the spacy library.

[0177] The important points visualization feature extracts important points and decisions from what is said during the meeting and displays them in real time on a whiteboard or display, allowing users to visually confirm important points during the meeting.

[0178] Examples:

[0179] Consider a situation where a meeting is being held in a factory to discuss the progress of a manufacturing process. When the meeting begins, the server launches the AI ​​facilitator and loads the data of the participants, "Participant A," "Participant B," and "Participant C." The device transcribes Participant A's statement, "Production line 1 is progressing as planned," in real time and sends it to the server. The server records this statement in the minutes in real time.

[0180] Next, if tracking reveals that participant B is not speaking much, the server sends a notification to participant B's device saying, "Please speak more." Furthermore, if participant C goes off on a tangent about a topic unrelated to the discussion, the server generates and notifies the participant B with a suggestion saying, "The conversation is going off topic. Let's return to the original topic."

[0181] Important comments are immediately displayed on displays in the factory, allowing all participants to share the latest information and conduct meetings efficiently.

[0182] Example prompts for generative AI models:

[0183] "Please propose an application that utilizes an AI facilitator with the functions of automatic minutes creation, prompting, digression detection, and visualization of key points in status meetings on a manufacturing line. Please also include a specific usage scenario and example program."

[0184] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0185] Step 1:

[0186] Execution of initialization process

[0187] The server creates an instance of the AI ​​facilitator when the meeting starts. It also loads basic information about participants (such as name, position, and past speech data) from a database. This information is stored using SQLite or similar. It also initializes a minutes template and prepares the data structures necessary for the meeting to proceed.

[0188] Input: Conference start signal, database (basic information of participants)

[0189] Output: An instance of the AI ​​facilitator, an initialized data structure, and a meeting minutes template.

[0190] Specific operation: The server uses a program such as Java or Python to access the conference management server and load participant data.

[0191] Step 2:

[0192] Real-time audio transcription

[0193] During the meeting, the device uses speech recognition technology (e.g., the speech_recognition library) to transcribe each utterance in real time, and then sends the transcribed utterances to the server.

[0194] Input: Audio data (speech during a meeting)

[0195] Output: Transcription data (text format), speaker identification information

[0196] Specific operation: The device receives voice input from the microphone, converts it into text data, and sends it to the server.

[0197] Step 3:

[0198] Real-time meeting minutes updates

[0199] The server updates the minutes in real time based on the received transcription data, thereby recording every word spoken during the meeting.

[0200] Input: Transcription data, speaker identification information

[0201] Output: Updated minutes

[0202] Specific operation: The server updates and stores the minutes data using a list or tree structure.

[0203] Step 4:

[0204] Participant speech volume tracking

[0205] The server tracks each participant's speech volume (number of times and duration) in real time and identifies participants who are not speaking much, allowing the server to provide appropriate feedback.

[0206] Input: Speech data, speaker identification information

[0207] Output: Tracking data (speech volume record)

[0208] Specific operation: The server measures the volume of messages using timestamp analysis and statistical functions.

[0209] Step 5:

[0210] Sending a prompt to speak

[0211] For participants who do not speak much, the server generates a notification encouraging them to speak and sends it to their smartphone or tablet.

[0212] Input: Tracking data, target participant identification information

[0213] Output: Speech prompt notification

[0214] Specific operation: The server uses a notification service (e.g., Firebase Cloud Messaging) to send a notification to the target participant's device.

[0215] Step 6:

[0216] Derailment detection and guide suggestions

[0217] The server uses natural language processing techniques (such as the Spacy library) to determine in real time whether the discussion is deviating from the planned topic, and if a digression is detected, it generates suggestions to bring the discussion back on topic and notifies the user.

[0218] Input: Transcription data, topic information

[0219] Output: Derailment notice, discussion guide suggestion

[0220] Specific operation: The server uses natural language processing technology to perform text analysis and analyze the direction of the discussion.

[0221] Step 7:

[0222] Visualizing key points

[0223] The device automatically extracts important points and decisions during the meeting and displays them on a whiteboard in real time, making it easier for participants to visually follow important matters during the meeting.

[0224] Input: Speech data, important points data

[0225] Output: Display (whiteboard)

[0226] Specific operation: The device uses the display output function to display important points on the whiteboard.

[0227] Step 8:

[0228] Data processing after the meeting

[0229] After the meeting ends, the server performs a final check and saves the minutes data, preparing it for distribution to participants.

[0230] Input: Final minutes data

[0231] Output: minutes for distribution

[0232] Specific operation: The server uses a database system to store the minutes data and export it in an appropriate format.

[0233] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0234] The present invention relates to a system and method that provides multiple functions for efficiently and productively conducting online meetings, in particular, a combination of an automated meeting record generator, a participant speech volume tracking unit, a speech promotion unit, a digression detection and discussion guide unit, a key point visualization unit, and an emotion engine that recognizes user emotions.

[0235] Initialization

[0236] The server creates an instance of the AI ​​facilitator at the start of a meeting. At this stage, it loads basic information about each user participating in the meeting (such as name, position, and past speech data) from a database. At the same time, the emotion engine is initialized, and emotional data is ready to be collected in real time through facial expression recognition and voice tone analysis.

[0237] Automatic meeting minutes creation

[0238] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, and the transcribed content and speaker identification information are sent to a server, which receives the data and automatically updates the minutes.

[0239] Encouraging participants to speak up

[0240] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them notifications encouraging them to speak. This promotes active participation by everyone in the meeting.

[0241] Derailment Detection and Discussion Guide

[0242] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user.

[0243] Visualization of key points

[0244] The device extracts important points and decisions from what is being said during the meeting and displays them on the whiteboard in real time. By visually displaying important points, users can easily follow the flow of the discussion, making meetings more efficient.

[0245] Introducing the Emotion Engine

[0246] The server uses an emotion engine to analyze the user's emotional state in real time. Emotions are identified from the user's facial expressions, tone of voice, and content of comments, and the emotional data is fed back to the system. For example, if a user is feeling stressed, this information is notified to other participants, encouraging them to take the discussion further and adjust its content accordingly.

[0247] Specific examples

[0248] Suppose a company is holding a project progress meeting. At the start of the meeting, the server initializes the AI ​​facilitator and emotion engine and loads the information of the participants (Tanaka, Yamada, Sato). As the meeting progresses, if Tanaka says, "This project is going well," this statement is transcribed in real time and sent to the server. The server records this statement in the minutes and simultaneously tracks the amount of speech Tanaka made.

[0249] When it becomes clear that Yamada has barely spoken, the server sends Yamada a notification saying, "You haven't spoken much. Do you have any input?" If the discussion veers away from project progress and starts talking about recent news, the server sends a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda."

[0250] The device displays the important point, "The project is progressing smoothly," on the whiteboard. On the other hand, if the emotion engine detects that Tanaka is feeling very stressed, the server sends a notification to other participants saying, "Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[0251] In this way, the present invention significantly improves the efficiency and productivity of online meetings. The introduction of an emotion engine allows smooth communication while taking into consideration the emotional state of users during a meeting.

[0252] The processing flow will be explained below.

[0253] Step 1: Initialize the system

[0254] The server instantiates the AI ​​facilitator and emotion engine at the start of a meeting. This initialization ensures that the entire system is ready to assist in running the meeting.

[0255] Step 2: Load participant information

[0256] The server loads basic information of participants (such as name, title, and past comment data) from the database. This information is used for subsequent comment tracking and sentiment analysis.

[0257] Step 3: Start speech recognition

[0258] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, instantly converting what is said during the meeting into text data.

[0259] Step 4: Sending speech data

[0260] The device sends the transcribed speech and the speaker's identity to the server, where it is ready to be recorded.

[0261] Step 5: Update the minutes

[0262] The server automatically updates the meeting minutes based on the received speech data, allowing the content of speech to be recorded in real time.

[0263] Step 6: Track your speech volume

[0264] The server tracks each participant's speech volume (number of times and duration) in real time, and the tracking data is used to encourage subsequent speech.

[0265] Step 7: Identify quiet participants

[0266] The server identifies participants who are quiet based on the tracking data. Participants who are below a certain threshold are identified.

[0267] Step 8: Send a prompt

[0268] The server generates a notification for participants who have not spoken much and sends it to the user to encourage them to speak up. Example of notification: "You haven't spoken much. Do you have any input?"

[0269] Step 9: Derailment detection with natural language processing

[0270] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the topic, and if there are signs of derailment, subsequent processing will be based on that information.

[0271] Step 10: Sending a discussion guide notification

[0272] If the server detects a digression, it generates a suggestion to bring the discussion back on topic and sends it to the user. Example notification: "The discussion is veering off-topic. Let's return to the main agenda."

[0273] Step 11: Extract and display key points

[0274] The device extracts important points and decisions from what is being said during the meeting and displays them on the whiteboard in real time in a format that is easy to follow visually.

[0275] Step 12: Collect and analyze emotion data

[0276] The server uses an emotion engine to collect emotional data from participants' facial expressions, tone of voice, and content of remarks, etc. Based on the analyzed emotional data, feedback is provided to the progress of the meeting in real time.

[0277] Step 13: Communicating Emotional State

[0278] When the server detects an important emotional state, such as a specific user being stressed, it notifies other participants of that information. Example of notification: "Mr. Tanaka seems to be feeling stressed. Please be considerate in the progress of the discussion."

[0279] Step 14: End the meeting and export data

[0280] At the end of the meeting, the server exports the accumulated minutes and emotion data in an editable format, which users can download and review as needed.

[0281] Through the above steps, this system incorporating emotion analysis not only enables online meetings to proceed efficiently and productively, but also ensures smooth communication that takes into consideration the emotions of participants.

[0282] Example 2

[0283] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0284] In online meetings, it is extremely difficult to automatically create minutes, track the amount of speech, prevent discussions from going off track, and visualize important points in real time. Another challenge is accurately understanding participants' emotional states and reflecting them in the progress of the discussion. By solving these issues, it is necessary to realize efficient and productive online meetings and reduce user stress.

[0285] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0286] In this invention, the server includes an automated meeting record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, and a sentiment analysis means, which make it possible to transcribe speeches made during a meeting in real time, track the speech volume of each participant, prevent digressions and bring the discussion back to the main topic, visualize key points, and analyze sentiment data to reflect the progress of the meeting.

[0287] "Automated meeting record generation means" is a technology that automatically records what is said in online meetings and generates minutes of the meeting.

[0288] The "participant speech volume tracking means" is a technology for measuring and recording the number of speeches and speech duration of each participant during a conference in real time.

[0289] The "speech promotion means" is a function for sending a notification to participants who do not speak much to encourage them to speak.

[0290] The "digression detection and discussion guidance means" is a technology for determining whether a discussion during a meeting has deviated from the planned topic and generating appropriate suggestions to bring the discussion back on topic.

[0291] "Key point visualization means" is a technology for extracting important points and decisions from what is said during a meeting and displaying them visually.

[0292] The "emotion analysis means" is a technology for analyzing the user's emotional state in real time using facial expression recognition technology and voice tone analysis technology.

[0293] "Speech recognition technology" is a technology for converting voice data into text data.

[0294] "Natural language processing technology" is a technology for analyzing text data, understanding its meaning, and processing it.

[0295] "Whiteboard software" is a digital tool for visually sharing information during meetings.

[0296] "Real time" is a time concept that refers to processing and display occurring almost simultaneously.

[0297] This invention relates to a system and method for efficiently and productively conducting online meetings, which combines an automated meeting record generation unit, a participant speech volume tracking unit, a speech promotion unit, a digression detection and discussion guide unit, a key point visualization unit, and a sentiment analysis unit. The operation of each unit will be described in detail below.

[0298] Initialization

[0299] The server instantiates an AI facilitator at the start of a meeting. This instantiation is done using a Python program that loads each participant's name, title, and past speech data from a database (e.g., MySQL). It also initializes an emotion engine using the Microsoft Azure Face API and Amazon Rekognition to recognize the user's facial expressions and analyze their speech tone.

[0300] Automatic meeting minutes creation

[0301] The device uses speech recognition technology to transcribe what is said during the meeting in real time using the Google Cloud Speech-to-Text API and IBM Watson Speech to Text. The transcribed content and speaker identification information are sent to a server, which then automatically updates the meeting minutes based on this data.

[0302] Encouraging participants to speak up

[0303] The server tracks each participant's speech volume (number of times and duration) in real time. This allows it to identify participants who are not speaking much and send them a notification to encourage them to speak up. For example, it uses the Slack API to send a message like, "You haven't spoken much. Do you have any input?"

[0304] Derailment Detection and Discussion Guide

[0305] The server uses natural language processing (NLP) technology to analyze the content of the meeting in real time. Specifically, it uses OpenAI GPT-3 and Google BERT to determine whether the discussion is deviating from the planned topic. If a deviation is detected, it generates a suggestion to the user, saying, "The discussion is veering off-topic. Let's return to the main agenda."

[0306] Visualization of key points

[0307] The device uses NLP technology to extract key points and decisions from what is being said, and the extracted content is displayed in real time on whiteboard software such as Miro or Microsoft Whiteboard.

[0308] Introducing sentiment analysis

[0309] The server uses emotion analysis technology to analyze the user's emotional state in real time. It uses Amazon Rekognition and Microsoft Azure Face API to analyze the user's facial expressions and voice tone, and identifies their emotional state based on that data. For example, if a user is feeling stressed, the server notifies other participants, saying, "Mr. Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[0310] Specific examples

[0311] Suppose a company is holding a project progress meeting. At the start of the meeting, the server initializes the AI ​​facilitator and sentiment analysis engine, and loads the information of the participants (Tanaka, Yamada, Sato) from the database. As the meeting progresses, if Tanaka says, "This project is going well," his remarks are transcribed in real time and sent to the server. The server updates the minutes based on this data, while simultaneously tracking the amount of speech Tanaka made.

[0312] If John has barely spoken, the server sends him a notification saying, "You haven't spoken much. Do you have any input?" If the discussion veers away from project progress and starts talking about recent news, the server sends him a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda."

[0313] The device displays the key point on the whiteboard: "The project is progressing smoothly." However, if the emotion analysis engine detects that Tanaka is feeling extremely stressed, it sends a notification to other participants saying, "Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[0314] Prompt Sentence Examples

[0315] "Describe a system that generates real-time online meeting minutes. Detail the specific hardware and software used and how the data is processed. Also, mention the role of sentiment analysis."

[0316] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0317] Step 1:

[0318] The server creates an instance of the AI ​​facilitator at the start of a meeting. As input, it retrieves the list of meeting participants and past speech data from a database (e.g., MySQL). The output is the initialization of the AI ​​facilitator and the initialization of the sentiment analysis engine. Specifically, it runs a Python program to load the names, titles, and past speech data of participants. It also initializes the sentiment analysis engine using the Microsoft Azure Face API and Amazon Rekognition.

[0319] Step 2:

[0320] The device uses speech recognition technology to transcribe what is said during a meeting in real time. The input is audio data collected during the meeting via a microphone. The output is transcribed text data, which is sent to the server via WebSocket or HTTP request. Specifically, it uses the Google Cloud Speech-to-Text API or IBM Watson Speech to Text to generate text data from the input audio data.

[0321] Step 3:

[0322] The server receives the submitted transcription data and automatically updates the minutes with speaker information. It receives the transcription data and speaker identification information as input. The output is the updated minutes data, which is shared with all participants in real time. Specifically, it runs a Python program that adds new utterances to the database and updates the current minutes.

[0323] Step 4:

[0324] The server tracks each participant's speech volume (number of times spoken and duration) in real time. It uses transcription data and time information as input. The output is speech volume data for each participant, which is displayed on a dashboard. Specifically, it runs a tracking algorithm, calculates the speech volume for each participant, and stores it in a database.

[0325] Step 5:

[0326] When a participant who has not spoken much is identified, the server sends a notification to that participant's device encouraging them to speak. Speech volume data is used as input. The output is a notification message encouraging them to speak. Specifically, the server uses the Slack API to send the message "You haven't spoken much. Do you have any input?"

[0327] Step 6:

[0328] The server uses natural language processing (NLP) technology to analyze the content of the meeting in real time. The input is the transcription data. The output is a judgment result on whether the discussion has gone off track. Specifically, it uses OpenAI GPT-3 and Google BERT to analyze the input data and determine whether there has been any deviation.

[0329] Step 7:

[0330] When it detects that the discussion is veering off-topic, the server generates a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda." and notifies the user. The deviation detection result is used as input. The output is a suggestion message to return the discussion. The specific operation is to generate a suggestion message and send it to the user via Slack or email.

[0331] Step 8:

[0332] The device extracts key points and decisions from what is being said. It uses transcription data as input. The output is key points that are displayed in the whiteboard software. Specifically, it uses NLP technology to extract key information and display it on Miro or Microsoft Whiteboard.

[0333] Step 9:

[0334] The server uses emotion analysis technology to analyze the user's emotional state in real time. Facial expression data and voice tone data are collected as input. The output is analyzed emotional data. Specifically, emotions are analyzed using Amazon Rekognition and Microsoft Azure Face API, and the results are reflected in the system.

[0335] Step 10:

[0336] Based on the analyzed emotional data, the server sends appropriate notifications to other participants according to the user's emotions. Emotional data is used as input. The output is a notification message based on the emotional state. A specific operation would be to generate a notification such as "Tanaka seems to be feeling stressed. Please be considerate in the progress of the discussion" and send it to other participants.

[0337] (Application example 2)

[0338] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0339] In modern factories, many meetings are held for production management and business improvement, but it is common for not all participants to speak up actively, and discussions to go off track. It is also difficult to grasp the emotional state of participants, making stress management and improving the quality of communication challenging. Therefore, a system that can ensure efficient and productive meeting progress while also taking into consideration the emotions of participants is needed.

[0340] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes an automated conference record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, and an emotion recognition engine. This allows for real-time transcription of comments made during a conference, promotes active participation by all participants, prevents the discussion from going off track, and visually displays key points, enabling the conference to proceed efficiently. Furthermore, by understanding participants' emotions in real time and taking into consideration any stress or dissatisfaction, it is possible to create a healthier communication environment.

[0341] An "automated meeting record generator" is a system that automatically records what is said in a meeting, transcribes it, and generates minutes in real time.

[0342] The "participant speech volume tracking means" is a system that records and manages the number of speeches and speech duration of each participant during a conference in real time.

[0343] The "speech promotion means" is a system that automatically sends a notification to participants who make few comments to encourage them to speak.

[0344] The "digression detection and discussion guidance means" is a system that detects when a meeting discussion deviates from the planned topic and provides instructions or suggestions to bring the discussion back to the original topic.

[0345] The "important point visualization means" is a system that extracts important statements and decisions made during a meeting and displays them visually.

[0346] The "emotion recognition engine" is a system that analyzes participants' emotions in real time based on their speech, facial expressions, tone of voice, etc., and provides feedback on the results.

[0347] The present invention provides a system for supporting efficient and productive meetings in factories, which is composed of a server, terminals, and users for data processing and interaction.

[0348] Initialization

[0349] The server creates an instance of the AI ​​facilitator when a meeting starts. Basic information about each user participating in the meeting (such as name, position, and past speech data) is loaded from the database. At the same time, the emotion recognition engine is initialized, and emotional data is collected in real time by recognizing the user's facial expressions and analyzing their voice tone.

[0350] Automatic meeting minutes creation

[0351] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, and the transcribed content and speaker identification information are sent to a server, which receives the data and automatically updates the minutes.

[0352] Encouraging participants to speak up

[0353] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them notifications encouraging them to speak. This promotes active participation by everyone in the meeting.

[0354] Derailment Detection and Discussion Guide

[0355] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user.

[0356] Visualization of key points

[0357] The device extracts important points and decisions from what is being said during the meeting and displays them in real time. By visually displaying important points, users can easily follow the flow of the discussion, making meetings more efficient.

[0358] Introduction of emotion recognition engineers

[0359] The server uses an emotion recognition engine to analyze the user's emotional state in real time. Emotions are identified from the user's facial expressions, tone of voice, and content of remarks, and the emotional data is fed back to the system. For example, if a user is feeling stressed, this information is notified to other participants, encouraging them to take the discussion further and adjust its content accordingly.

[0360] Specific examples

[0361] Suppose a project progress meeting is being held at a factory. At the start of the meeting, the server initializes the AI ​​facilitator and emotion recognition engine, and loads information about the participants (User A, User B, and User C). As the meeting progresses, if User A says, "This project is going well," the device transcribes this statement in real time and sends it to the server. The server records this statement in the minutes and simultaneously tracks the amount of speech that User A made.

[0362] If it becomes clear that User B has barely spoken, the server sends User B a notification asking, "Can I have your opinion?" If the discussion strays from the project's progress and begins talking about recent news, the server sends a suggestion saying, "The discussion is drifting. Let's get back to the original topic." The device then displays the key point: "Project progress is going well."

[0363] On the other hand, if the emotion recognition engine detects that User A is feeling extremely stressed, the server will send a notification to other participants saying, "User A is feeling stressed. Please be considerate in the discussion." This will allow the meeting to proceed efficiently and productively, and by taking into consideration the emotions of the participants, smooth communication will be achieved.

[0364] Prompt Sentence Examples

[0365] Please implement a function that allows the AI ​​facilitator to track the amount of speech participants make in online meetings and send notifications to encourage participants who are not speaking often. Also, please add a function that uses an emotion recognition engine to analyze the emotional state of participants and notify other participants if they are feeling stressed.

[0366] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0367] Step 1:

[0368] At the start of a meeting, the server instantiates an AI facilitator and loads basic participant information from a database. Input includes participant name, role, and past speech data. Based on this data, the server creates individual participant profiles as output and provides them to the AI ​​facilitator.

[0369] Step 2:

[0370] The server initializes the emotion recognition engine and prepares to collect the user's emotion data in real time. The inputs are facial expression recognition data, voice tone data, and speech content data, and the data is analyzed to output the user's emotional state.

[0371] Step 3:

[0372] The device uses speech recognition technology to transcribe speech in real time during a meeting. The input is the user's speech audio data, and the output is the transcribed text data. This allows the speech content to be recorded as text.

[0373] Step 4:

[0374] The device sends the transcribed speech and the speaker's identification information to the server. The input is the speech transcription data and the speaker ID, and the output is a packet containing these data. The server receives this data and automatically updates the minutes.

[0375] Step 5:

[0376] The server tracks the amount of speech of each participant in real time. The input is speech transcription data and speaker ID, and accumulates the number of speeches and speech duration in a database. The output is speech volume data for each participant.

[0377] Step 6:

[0378] The server sends a notification to participants who speak less to encourage them to speak. The input is the speech volume data of each participant, and the output is a notification message to participants who speak less.

[0379] Step 7:

[0380] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. The input is speech transcription data, and the output is a flag indicating whether the topic is deviating. If a derailment is detected, the server generates instructions to return to the original topic and notifies the user via their device.

[0381] Step 8:

[0382] The device extracts important points and decisions from what is said during the meeting and displays them on the screen. The input is the transcription data, and the output is a visual representation of the extracted important points. By visually displaying the important points, users can easily follow the flow of the discussion.

[0383] Step 9:

[0384] The server uses an emotion recognition engine to analyze the user's emotional state. The inputs are the user's facial expression data, voice tone data, and speech content data, and the output is emotional state data. For example, if the user is in a stressful state, that information is notified to other participants.

[0385] Step 10:

[0386] At the end of the meeting, the server saves all meeting minutes, emotion data, and speech volume data, and outputs them as a report if necessary. The input is all the data collected during the meeting, and the output is the final meeting report. This leaves a record to review later.

[0387] 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 a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the 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.

[0388] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0389] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0390] [Second embodiment]

[0391] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0392] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0393] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0394] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0395] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0396] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0397] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0398] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0399] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[0400] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0401] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0402] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. 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."

[0403] Initialization

[0404] When a meeting starts, the server creates an instance of the AI ​​facilitator. Next, it loads basic information about participants (such as their names, positions, and past speech data) from the database. It also initializes a minutes template and prepares the data structures necessary for the meeting to proceed.

[0405] Automatic meeting minutes creation

[0406] The device uses speech recognition technology to transcribe each statement in real time during the meeting. The transcribed content and speaker identification information are sent to the server as the meeting progresses. The server receives the data and automatically records it as meeting minutes.

[0407] Encouraging participants to speak up

[0408] The server tracks each participant's speech volume (number of times and duration) in real time. It identifies participants who speak less and sends them notifications encouraging them to speak. Notifications are automatically generated according to the progress of the conference and are delivered to users.

[0409] Derailment Detection and Discussion Guide

[0410] The server uses natural language processing technology to determine in real time whether a discussion has deviated from the planned topic. If a deviation from the topic is detected, the server generates suggestions to bring the discussion back on topic and notifies the user, thereby reducing the risk of the discussion dragging on unnecessarily or drifting off topic.

[0411] Visualization of key points

[0412] The device extracts important points and decisions from what is being said during the meeting and displays them on a whiteboard in real time. This whiteboard is constantly updated during the meeting, making it easier for users to visually follow the important points. This allows meetings to proceed efficiently without losing track of the flow of discussion or important points.

[0413] Specific examples

[0414] For example, let's say a company is holding a project progress meeting. When the meeting begins, the server launches the AI ​​facilitator and loads the data of the participants, "Tanaka," "Yamada," and "Sato." As the meeting progresses, if Tanaka says, "This project is going well," his remarks are transcribed in real time by the device. The transcription data and Tanaka's identification information are sent to the server and recorded in the minutes.

[0415] Next, tracking reveals that Yamada has barely spoken anything. The server sends Yamada a notification saying, "You haven't spoken much. Do you have any input?" This gives Yamada a chance to express his opinion.

[0416] If, at some point, Sato starts talking about recent news, the server will determine that the remark is unrelated to the project's progress and send the user a suggestion: "The discussion is veering off-topic. Let's return to the main agenda."

[0417] Furthermore, the device automatically displays important points in the meeting, such as "Project progress is going smoothly," on the whiteboard, allowing users to visually confirm important matters during the meeting in real time.

[0418] As described above, this system provides various functions to ensure that online meetings proceed efficiently and productively, contributing to improving the quality of meetings.

[0419] The processing flow will be explained below.

[0420] Step 1: Initialization

[0421] The server creates an instance of the AI ​​facilitator at the start of the meeting, which launches a system to assist in the progress of the online meeting.

[0422] Step 2: Load participant information

[0423] The server loads basic information of participants (such as name, position, and past comment data) from the database, which enables identification of each participant and support based on past data.

[0424] Step 3: Initialize the meeting minutes template

[0425] The server initializes a meeting template, which provides the data structures necessary for the minutes format and flow of the meeting.

[0426] Step 4: Start real-time transcription

[0427] As the meeting progresses, the device uses speech recognition technology to transcribe each utterance in real time, and the speech is automatically converted into text.

[0428] Step 5: Sending speech data

[0429] The device sends the transcribed speech and the speaker's identity to the server, where the speech is ready to be recorded.

[0430] Step 6: Update the minutes

[0431] The server automatically updates the meeting minutes based on the received speech data, ensuring that each speech is recorded appropriately.

[0432] Step 7: Track your speech volume

[0433] The server tracks each participant's speech volume (number of times and duration) in real time, allowing the activity level of participants to be monitored.

[0434] Step 8: Identify quiet participants

[0435] The server identifies participants who are less active: participants whose number of comments falls below a certain threshold.

[0436] Step 9: Send a prompt

[0437] The server generates a notification for quiet participants to speak up and sends it to the user, with a message such as "You haven't spoken much. Do you have any input?"

[0438] Step 10: Derailment detection

[0439] The server uses natural language processing technology to determine in real time whether the current comment is off-topic.

[0440] Step 11: Sending a discussion guide notification

[0441] If the server detects a digression, it generates a suggestion to bring the discussion back on topic and sends it to the user: "The discussion is veering off-topic. Let's return to the main agenda."

[0442] Step 12: Extracting key points

[0443] The device extracts important points and decisions from what is said during the meeting.

[0444] Step 13: Display on the Whiteboard

[0445] The device displays the extracted important points on a whiteboard in real time, allowing users to visually confirm the important points.

[0446] Step 14: End the meeting and export the minutes

[0447] At the end of the meeting, the server exports the accumulated minutes in an editable format, which users can review and modify.

[0448] Through these steps, this system enables online meetings to proceed efficiently and productively.

[0449] Example 1

[0450] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0451] Existing online meeting systems face the challenge of efficiently managing a wide range of tasks, including meeting progress, minutes creation, participant prompting, digression detection and discussion guidance, and visualization of key points. This can lead to poor meeting quality, redundant discussions, and participants being reluctant to speak up. Furthermore, creating meeting minutes manually is time-consuming and labor-intensive, and lacks accuracy.

[0452] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0453] In this invention, the server includes a meeting start detection means, an AI facilitator instance generation means, a participant information database loading means, a minutes template initialization means, a speech transcription means using speech recognition, a speech data transmission means, a speech data recording means, a participant speech volume tracking means, a speech promotion notification generation means, a speech promotion notification transmission means, a discussion digression detection means, a discussion guide notification generation means, a discussion guide notification transmission means, an important point extraction means, and a whiteboard display means. This makes it possible to manage the progress of the meeting, create minutes, encourage participants to speak, guide the discussion, and visualize important points in real time.

[0454] The "conference start detection means" is a means for recognizing the start of a conference and for the system to automatically start the necessary initial processing.

[0455] An "AI facilitator instance generation means" is a means for generating an instance of an AI model to smoothly proceed with a meeting.

[0456] The "participant information database loading means" is a means for loading basic information of conference participants from a database.

[0457] The "minutes template initialization means" is a means for initializing a template for creating minutes of a meeting.

[0458] A "means for transcribing speech using speech recognition" is a means for using speech recognition technology to transcribe speech during a meeting in real time.

[0459] The "utterance data transmission means" is a means for transmitting the transcribed utterance data to the server.

[0460] The "utterance data recording means" is a means for recording received utterance data in a database.

[0461] The "participant speech volume tracking means" is a means for tracking the speech volume of each participant during a conference in real time.

[0462] The "means for generating a speech prompting notification" is a means for generating a notification that prompts a participant who makes few comments to speak.

[0463] The "speech prompting notification sending means" is a means for sending the generated speech prompting notification to the relevant participant.

[0464] The "discussion deviation detection means" is a means for detecting whether the discussion in the meeting has deviated from the planned topic.

[0465] The "discussion guide notification generating means" is a means for generating a suggestion to bring the discussion back to the main topic when the discussion goes off track.

[0466] The "discussion guide notification sending means" is a means for sending the generated discussion guide notification to the participants.

[0467] The "means for extracting important points" is a means for extracting important points and decisions from the content of statements made during a meeting.

[0468] The "whiteboard display means" is a means for displaying the extracted important points on a whiteboard in real time.

[0469] The present invention is a system for conducting online meetings efficiently and productively. This system automates various processes from the start to the end of a meeting and combines multiple means to improve the quality of the meeting.

[0470] Initialization

[0471] When the server detects the start of a meeting, it creates an instance of an AI facilitator. This uses an AI framework such as TensorFlow or PyTorch. Next, it loads participant information from a database (e.g., MySQL or PostgreSQL). This includes participant names, positions, and past speech data. Next, the server initializes a minutes template. This template uses a predefined file in Markdown format or similar.

[0472] Automatic meeting minutes creation

[0473] During a meeting, the device uses speech recognition technology such as the Google Speech-to-Text API or IBM Watson Speech to Text to transcribe each utterance in real time. The transcribed utterances and the speaker's identity are then sent to the server via an HTTP POST request. The server receives this data and automatically records it in a database such as Elasticsearch.

[0474] Encouraging participants to speak up

[0475] The server uses MongoDB or similar to track the amount of speech each participant makes (number of times they make a comment and how long they speak). For participants who make little comments, the server uses the Slack API or Microsoft Teams API to generate and send notifications encouraging them to speak. These notifications are generated automatically and sent to users.

[0476] Derailment Detection and Discussion Guide

[0477] The server uses natural language processing techniques, such as Hugging Face's Transformers library, to determine in real time whether the discussion is deviating from the planned topic. If a deviation is detected, the server compares it with pre-defined agenda information, generates suggestions to bring the discussion back on topic, and communicates them to the user. This notification is sent automatically at the appropriate time.

[0478] Visualization of key points

[0479] The device uses natural language processing libraries such as SpaCy and NLTK to extract key points and decisions from what is being said during the meeting. The extracted key points are displayed in real time on a whiteboard service such as Miro Board or Microsoft Whiteboard. This display is constantly updated as the meeting progresses, making it easy for users to visually follow the key points.

[0480] Specific examples

[0481] For example, when a company holds a project status meeting, the server detects the start of the meeting through the Google Calendar API, creates an instance of an AI facilitator using TensorFlow, and then loads the data of the participants, "Tanaka," "Yamada," and "Sato," from a MySQL database.

[0482] As the meeting progresses, if Tanaka says, "This project is going well," the device will transcribe his speech in real time using the Google Speech-to-Text API. The transcription data and Tanaka's identification information will be sent to the server via an HTTP POST request and recorded in Elasticsearch.

[0483] Once MongoDB data tracks that Yamada has barely spoken, the server uses the Slack API to send a notification to Yamada saying, "You haven't spoken much. Do you have any input?"

[0484] If Sato starts talking about recent news midway through, the server will use Hugging Face's Transformers library to determine that his remarks are unrelated to the project's progress and send the user a suggestion: "The discussion is veering off-topic. Let's return to the main agenda."

[0485] The device automatically displays the key point of the meeting, such as "Project progress is going well," on the Miro board, allowing users to visually check important matters during the meeting in real time through the whiteboard.

[0486] Prompt Sentence Examples

[0487] Below is an example prompt for requesting minutes of a project status meeting:

[0488] A project progress meeting was held. The participants were Tanaka, Yamada, and Sato. Tanaka stated, "This project is progressing smoothly." Based on the minutes, please create and propose the following action items:

[0489] As described above, the present invention provides various functions for conducting online conferences efficiently and productively, thereby contributing to improving the quality of conferences.

[0490] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0491] Processing flow

[0492] Step 1:

[0493] The server detects when a meeting starts. To do this, it uses the Google Calendar API to monitor scheduled events. The input is the calendar information obtained from the API, and the output is a meeting start event.

[0494] Step 2:

[0495] The server creates an instance of the AI ​​facilitator. First, it loads the AI ​​model using the TensorFlow or PyTorch library and creates an instance. The input is the pre-trained AI model, and the output is the generated instance of the AI ​​facilitator.

[0496] Step 3:

[0497] The server loads participant information from a database. It issues SQL queries to a database such as MySQL or PostgreSQL to retrieve participant names, positions, and past comment data. The input is the database connection information and query, and the output is the retrieved participant information.

[0498] Step 4:

[0499] The server initializes the minutes template. It reads a predefined Markdown template file and prepares the initial state. The input is the path to the template file, and the output is the initialized minutes template.

[0500] Step 5:

[0501] The device transcribes speech during meetings in real time. It uses the Google Speech-to-Text API to convert audio data captured from the microphone into text data. The input is audio data, and the output is transcribed text data.

[0502] Step 6:

[0503] The device sends the transcribed speech data and the speaker's identification information to the server using an HTTP POST request. The input is the speech data and the identification information, and the output is the success status of the data transmission to the server.

[0504] Step 7:

[0505] The server records the received utterance data in Elasticsearch, analyzes the data, and adds it to the appropriate index. The input is the received utterance data, and the output is the index information stored in Elasticsearch.

[0506] Step 8:

[0507] The server tracks the amount of speech of each participant in real time. It records the speech data in MongoDB and executes aggregation queries to calculate the number of speeches and speech duration. The input is the received speech data, and the output is the tracked speech volume.

[0508] Step 9:

[0509] The server generates notifications to encourage participants who speak less to speak. The notification content is automatically generated using the Slack API or Microsoft Teams API. The input is the tracked speech volume data, and the output is the generated notification message.

[0510] Step 10:

[0511] The server sends the generated notification to the appropriate user using the Slack API or Microsoft Teams API. The input is the generated notification message, and the output is the success status of the notification sending.

[0512] Step 11:

[0513] The server detects deviations in the discussion. It uses Hugging Face's Transformers library to compare the content of comments with the agenda and determine whether or not a deviation has occurred. The input is the comment data and agenda information, and the output is the deviation detection result.

[0514] Step 12:

[0515] When a digression is detected, the server generates a suggestion to bring the discussion back to the main topic. The suggestion content is generated using a generative AI model. The input is the detected digression result, and the output is the generated suggestion message.

[0516] Step 13:

[0517] The server notifies the user of the proposal message. It sends the notification using the chat application's API. The input is the generated proposal message, and the output is the success status of sending the notification.

[0518] Step 14:

[0519] The device extracts important points from what is said during the meeting. It uses SpaCy and NLTK to analyze the speech and extract important points and decisions. The input is the speech data, and the output is the extracted important points.

[0520] Step 15:

[0521] The device displays the extracted key points on a whiteboard. It updates the display in real time using the APIs of Miro Board and Microsoft Whiteboard. The input is the extracted key points, and the output is the display result on the whiteboard.

[0522] Through the above processing steps, the system manages the progress of the meeting, creates minutes, encourages participants to speak up, directs the discussion, and visualizes important points.

[0523] (Application example 1)

[0524] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0525] Modern meeting systems lack features such as taking minutes, encouraging participants to speak up, and detecting derailed discussions. Some participants often don't speak much, or the discussion often goes off track, wasting time. Furthermore, participants can lose track of the flow of the discussion because they cannot visually track important points and decisions. These issues directly affect production efficiency, particularly in factory progress meetings, so there is a high need for a system that can quickly and efficiently conduct meetings.

[0526] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0527] In this invention, the server includes an automated meeting record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, a means for transcribing speeches made during a meeting in real time using speech recognition technology and recording the results, a means for sending notifications to participants who speak less to encourage them to speak, a means for determining in real time whether the discussion has deviated from the planned topic using natural language processing technology and generating suggestions, and a means for displaying key points and decisions on a whiteboard in real time, thereby enabling improvements in the efficiency and productivity of meetings.

[0528] The "automated meeting record generator" is a function that automatically recognizes statements made during a meeting and creates minutes in real time.

[0529] The "participant speech volume tracking means" is a function for measuring and recording the number of speeches and speech duration of each conference participant in real time.

[0530] The "speech promotion means" is a function for sending notifications to participants who do not speak much, encouraging them to speak.

[0531] "Derailment detection and discussion guidance means" is a function that uses natural language processing technology to determine whether a discussion has deviated from the planned topic and suggests an appropriate direction for the discussion.

[0532] The "important point visualization means" is a function that displays important points and decisions made during a meeting on a display device in real time, allowing participants to visually grasp important information.

[0533] "Speech recognition technology" is a technology for converting speech into text, and in this invention is a technology used to transcribe statements made during a meeting in real time.

[0534] "Real-time transcription means" is a function that uses voice recognition technology to instantly convert meeting speech into text and record it.

[0535] "Means for encouraging participants who speak infrequently to speak by sending notifications" is a function for tracking the volume of comments in real time and sending notifications to participants who speak infrequently to encourage them to speak.

[0536] "Natural language processing technology" is a technology for understanding and analyzing human language, and in this invention, it is a technology used to detect digressions in discussions and provide appropriate discussion guidance.

[0537] "A means of displaying important points and decisions on a whiteboard in real time" is a function that instantly reflects important statements and decisions made during a meeting on a display device such as a whiteboard, allowing all participants to visually confirm them.

[0538] The system embodying this invention is designed to streamline factory progress meetings and is composed of a server, terminals, and users with various functions. Specifically, it is composed of an automated meeting record generation means, a means for tracking the amount of speech by participants, a means for encouraging speech, a means for detecting deviations and guiding discussions, and a means for visualizing important points.

[0539] When a meeting starts, the server creates an instance of the AI ​​facilitator and loads basic information about participants (such as their names, positions, and past speech data) from a database. This database uses a system such as SQLite. It also initializes a minutes template and prepares the data structure necessary for the meeting to proceed.

[0540] The device uses speech recognition technology to transcribe speech in real time during a meeting. The transcribed speech and speaker identification information are sent to a server, which records the data as meeting minutes in real time. The speech recognition technology used here includes the speech_recognition library.

[0541] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them a notification to encourage them to speak. This notification is sent to an application on a smartphone or tablet.

[0542] The digression detection and discussion guidance mechanism uses natural language processing (NLP) technology to determine in real time whether a discussion has deviated from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user. The NLP technology used here includes the spacy library.

[0543] The important points visualization feature extracts important points and decisions from what is said during the meeting and displays them in real time on a whiteboard or display, allowing users to visually confirm important points during the meeting.

[0544] Examples:

[0545] Consider a situation where a meeting is being held in a factory to discuss the progress of a manufacturing process. When the meeting begins, the server launches the AI ​​facilitator and loads the data of the participants, "Participant A," "Participant B," and "Participant C." The device transcribes Participant A's statement, "Production line 1 is progressing as planned," in real time and sends it to the server. The server records this statement in the minutes in real time.

[0546] Next, if tracking reveals that participant B is not speaking much, the server sends a notification to participant B's device saying, "Please speak more." Furthermore, if participant C goes off on a tangent about a topic unrelated to the discussion, the server generates and notifies the participant B with a suggestion saying, "The conversation is going off topic. Let's return to the original topic."

[0547] Important comments are immediately displayed on displays in the factory, allowing all participants to share the latest information and conduct meetings efficiently.

[0548] Example prompts for generative AI models:

[0549] "Please propose an application that utilizes an AI facilitator with the functions of automatic minutes creation, prompting, digression detection, and visualization of key points in status meetings on a manufacturing line. Please also include a specific usage scenario and example program."

[0550] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0551] Step 1:

[0552] Execution of initialization process

[0553] The server creates an instance of the AI ​​facilitator when the meeting starts. It also loads basic information about participants (such as name, position, and past speech data) from a database. This information is stored using SQLite or similar. It also initializes a minutes template and prepares the data structures necessary for the meeting to proceed.

[0554] Input: Conference start signal, database (basic information of participants)

[0555] Output: An instance of the AI ​​facilitator, an initialized data structure, and a meeting minutes template.

[0556] Specific operation: The server uses a program such as Java or Python to access the conference management server and load participant data.

[0557] Step 2:

[0558] Real-time audio transcription

[0559] During the meeting, the device uses speech recognition technology (e.g., the speech_recognition library) to transcribe each utterance in real time, and then sends the transcribed utterances to the server.

[0560] Input: Audio data (speech during a meeting)

[0561] Output: Transcription data (text format), speaker identification information

[0562] Specific operation: The device receives voice input from the microphone, converts it into text data, and sends it to the server.

[0563] Step 3:

[0564] Real-time meeting minutes updates

[0565] The server updates the minutes in real time based on the received transcription data, thereby recording every word spoken during the meeting.

[0566] Input: Transcription data, speaker identification information

[0567] Output: Updated minutes

[0568] Specific operation: The server updates and stores the minutes data using a list or tree structure.

[0569] Step 4:

[0570] Participant speech volume tracking

[0571] The server tracks each participant's speech volume (number of times and duration) in real time and identifies participants who are not speaking much, allowing the server to provide appropriate feedback.

[0572] Input: Speech data, speaker identification information

[0573] Output: Tracking data (speech volume record)

[0574] Specific operation: The server measures the volume of messages using timestamp analysis and statistical functions.

[0575] Step 5:

[0576] Sending a prompt to speak

[0577] For participants who do not speak much, the server generates a notification encouraging them to speak and sends it to their smartphone or tablet.

[0578] Input: Tracking data, target participant identification information

[0579] Output: Speech prompt notification

[0580] Specific operation: The server uses a notification service (e.g., Firebase Cloud Messaging) to send a notification to the target participant's device.

[0581] Step 6:

[0582] Derailment detection and guide suggestions

[0583] The server uses natural language processing techniques (such as the Spacy library) to determine in real time whether the discussion is deviating from the planned topic, and if a digression is detected, it generates suggestions to bring the discussion back on topic and notifies the user.

[0584] Input: Transcription data, topic information

[0585] Output: Derailment notice, discussion guide suggestion

[0586] Specific operation: The server uses natural language processing technology to perform text analysis and analyze the direction of the discussion.

[0587] Step 7:

[0588] Visualizing key points

[0589] The device automatically extracts important points and decisions during the meeting and displays them on a whiteboard in real time, making it easier for participants to visually follow important matters during the meeting.

[0590] Input: Speech data, important points data

[0591] Output: Display (whiteboard)

[0592] Specific operation: The device uses the display output function to display important points on the whiteboard.

[0593] Step 8:

[0594] Data processing after the meeting

[0595] After the meeting ends, the server performs a final check and saves the minutes data, preparing it for distribution to participants.

[0596] Input: Final minutes data

[0597] Output: minutes for distribution

[0598] Specific operation: The server uses a database system to store the minutes data and export it in an appropriate format.

[0599] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0600] The present invention relates to a system and method that provides multiple functions for efficiently and productively conducting online meetings, in particular, a combination of an automated meeting record generator, a participant speech volume tracking unit, a speech promotion unit, a digression detection and discussion guide unit, a key point visualization unit, and an emotion engine that recognizes user emotions.

[0601] Initialization

[0602] The server creates an instance of the AI ​​facilitator at the start of a meeting. At this stage, it loads basic information about each user participating in the meeting (such as name, position, and past speech data) from a database. At the same time, the emotion engine is initialized, and emotional data is ready to be collected in real time through facial expression recognition and voice tone analysis.

[0603] Automatic meeting minutes creation

[0604] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, and the transcribed content and speaker identification information are sent to a server, which receives the data and automatically updates the minutes.

[0605] Encouraging participants to speak up

[0606] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them notifications encouraging them to speak. This promotes active participation by everyone in the meeting.

[0607] Derailment Detection and Discussion Guide

[0608] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user.

[0609] Visualization of key points

[0610] The device extracts important points and decisions from what is being said during the meeting and displays them on the whiteboard in real time. By visually displaying important points, users can easily follow the flow of the discussion, making meetings more efficient.

[0611] Introducing the Emotion Engine

[0612] The server uses an emotion engine to analyze the user's emotional state in real time. Emotions are identified from the user's facial expressions, tone of voice, and content of comments, and the emotional data is fed back to the system. For example, if a user is feeling stressed, this information is notified to other participants, encouraging them to take the discussion further and adjust its content accordingly.

[0613] Specific examples

[0614] Suppose a company is holding a project progress meeting. At the start of the meeting, the server initializes the AI ​​facilitator and emotion engine and loads the information of the participants (Tanaka, Yamada, Sato). As the meeting progresses, if Tanaka says, "This project is going well," this statement is transcribed in real time and sent to the server. The server records this statement in the minutes and simultaneously tracks the amount of speech Tanaka made.

[0615] When it becomes clear that Yamada has barely spoken, the server sends Yamada a notification saying, "You haven't spoken much. Do you have any input?" If the discussion veers away from project progress and starts talking about recent news, the server sends a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda."

[0616] The device displays the important point, "The project is progressing smoothly," on the whiteboard. On the other hand, if the emotion engine detects that Tanaka is feeling very stressed, the server sends a notification to other participants saying, "Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[0617] In this way, the present invention significantly improves the efficiency and productivity of online meetings. The introduction of an emotion engine allows smooth communication while taking into consideration the emotional state of users during a meeting.

[0618] The processing flow will be explained below.

[0619] Step 1: Initialize the system

[0620] The server instantiates the AI ​​facilitator and emotion engine at the start of a meeting. This initialization ensures that the entire system is ready to assist in running the meeting.

[0621] Step 2: Load participant information

[0622] The server loads basic information of participants (such as name, title, and past comment data) from the database. This information is used for subsequent comment tracking and sentiment analysis.

[0623] Step 3: Start speech recognition

[0624] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, instantly converting what is said during the meeting into text data.

[0625] Step 4: Sending speech data

[0626] The device sends the transcribed speech and the speaker's identity to the server, where it is ready to be recorded.

[0627] Step 5: Update the minutes

[0628] The server automatically updates the meeting minutes based on the received speech data, allowing the content of speech to be recorded in real time.

[0629] Step 6: Track your speech volume

[0630] The server tracks each participant's speech volume (number of times and duration) in real time, and the tracking data is used to encourage subsequent speech.

[0631] Step 7: Identify quiet participants

[0632] The server identifies participants who are quiet based on the tracking data. Participants who are below a certain threshold are identified.

[0633] Step 8: Send a prompt

[0634] The server generates a notification for participants who have not spoken much and sends it to the user to encourage them to speak up. Example of notification: "You haven't spoken much. Do you have any input?"

[0635] Step 9: Derailment detection with natural language processing

[0636] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the topic, and if there are signs of derailment, subsequent processing will be based on that information.

[0637] Step 10: Sending a discussion guide notification

[0638] If the server detects a digression, it generates a suggestion to bring the discussion back on topic and sends it to the user. Example notification: "The discussion is veering off-topic. Let's return to the main agenda."

[0639] Step 11: Extract and display key points

[0640] The device extracts important points and decisions from what is being said during the meeting and displays them on the whiteboard in real time in a format that is easy to follow visually.

[0641] Step 12: Collect and analyze emotion data

[0642] The server uses an emotion engine to collect emotional data from participants' facial expressions, tone of voice, and content of remarks, etc. Based on the analyzed emotional data, feedback is provided to the progress of the meeting in real time.

[0643] Step 13: Communicating Emotional State

[0644] When the server detects an important emotional state, such as a specific user being stressed, it notifies other participants of that information. Example of notification: "Mr. Tanaka seems to be feeling stressed. Please be considerate in the progress of the discussion."

[0645] Step 14: End the meeting and export data

[0646] At the end of the meeting, the server exports the accumulated minutes and emotion data in an editable format, which users can download and review as needed.

[0647] Through the above steps, this system incorporating emotion analysis not only enables online meetings to proceed efficiently and productively, but also ensures smooth communication that takes into consideration the emotions of participants.

[0648] Example 2

[0649] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0650] In online meetings, it is extremely difficult to automatically create minutes, track the amount of speech, prevent discussions from going off track, and visualize important points in real time. Another challenge is accurately understanding participants' emotional states and reflecting them in the progress of the discussion. By solving these challenges, it is necessary to realize efficient and productive online meetings and reduce user stress.

[0651] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0652] In this invention, the server includes an automated conference record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, and a sentiment analysis means, which make it possible to transcribe speeches made during a conference in real time, track the speech volume of each participant, prevent digressions and bring the discussion back to the main topic, visualize key points, and analyze sentiment data to reflect the progress of the conference.

[0653] "Automated meeting record generation means" is a technology that automatically records what is said in online meetings and generates minutes of the meeting.

[0654] The "participant speech volume tracking means" is a technology for measuring and recording the number of speeches and speech duration of each participant during a conference in real time.

[0655] The "speech promotion means" is a function for sending a notification to participants who do not speak much to encourage them to speak.

[0656] The "digression detection and discussion guidance means" is a technology for determining whether a discussion during a meeting has deviated from the planned topic and generating appropriate suggestions to bring the discussion back on topic.

[0657] "Key point visualization means" is a technology for extracting important points and decisions from what is said during a meeting and displaying them visually.

[0658] The "emotion analysis means" is a technology for analyzing the user's emotional state in real time using facial expression recognition technology and voice tone analysis technology.

[0659] "Speech recognition technology" is a technology for converting voice data into text data.

[0660] "Natural language processing technology" is a technology for analyzing text data, understanding its meaning, and processing it.

[0661] "Whiteboard software" is a digital tool for visually sharing information during meetings.

[0662] "Real time" is a time concept that refers to processing and display occurring almost simultaneously.

[0663] This invention relates to a system and method for efficiently and productively conducting online meetings, which combines an automated meeting record generation unit, a participant speech volume tracking unit, a speech promotion unit, a digression detection and discussion guide unit, a key point visualization unit, and a sentiment analysis unit. The operation of each unit will be described in detail below.

[0664] Initialization

[0665] The server instantiates an AI facilitator at the start of a meeting. This instantiation is done using a Python program that loads each participant's name, title, and past speech data from a database (e.g., MySQL). It also initializes an emotion engine using the Microsoft Azure Face API and Amazon Rekognition to recognize the user's facial expressions and analyze their speech tone.

[0666] Automatic meeting minutes creation

[0667] The device uses speech recognition technology to transcribe what is said during the meeting in real time using the Google Cloud Speech-to-Text API and IBM Watson Speech to Text. The transcribed content and speaker identification information are sent to a server, which then automatically updates the meeting minutes based on this data.

[0668] Encouraging participants to speak up

[0669] The server tracks each participant's speech volume (number of times and duration) in real time. This allows it to identify participants who are not speaking much and send them a notification to encourage them to speak up. For example, it uses the Slack API to send a message like, "You haven't spoken much. Do you have any input?"

[0670] Derailment Detection and Discussion Guide

[0671] The server uses natural language processing (NLP) technology to analyze the content of the meeting in real time. Specifically, it uses OpenAI GPT-3 and Google BERT to determine whether the discussion is deviating from the planned topic. If a deviation is detected, it generates a suggestion to the user, saying, "The discussion is veering off-topic. Let's return to the main agenda."

[0672] Visualization of key points

[0673] The device uses NLP technology to extract key points and decisions from what is being said, and the extracted content is displayed in real time on whiteboard software such as Miro or Microsoft Whiteboard.

[0674] Introducing sentiment analysis

[0675] The server uses emotion analysis technology to analyze the user's emotional state in real time. It uses Amazon Rekognition and Microsoft Azure Face API to analyze the user's facial expressions and voice tone, and identifies their emotional state based on that data. For example, if a user is feeling stressed, the server notifies other participants, saying, "Mr. Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[0676] Specific examples

[0677] Suppose a company is holding a project progress meeting. At the start of the meeting, the server initializes the AI ​​facilitator and sentiment analysis engine, and loads the information of the participants (Tanaka, Yamada, Sato) from the database. As the meeting progresses, if Tanaka says, "This project is going well," his remarks are transcribed in real time and sent to the server. The server updates the minutes based on this data, while simultaneously tracking the amount of speech Tanaka made.

[0678] If John has barely spoken, the server sends him a notification saying, "You haven't spoken much. Do you have any input?" If the discussion veers away from project progress and starts talking about recent news, the server sends him a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda."

[0679] The device displays the key point on the whiteboard: "The project is progressing smoothly." However, if the emotion analysis engine detects that Tanaka is feeling extremely stressed, it sends a notification to other participants saying, "Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[0680] Prompt Sentence Examples

[0681] "Describe a system that generates real-time online meeting minutes. Detail the specific hardware and software used and how the data is processed. Also, mention the role of sentiment analysis."

[0682] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0683] Step 1:

[0684] The server creates an instance of the AI ​​facilitator at the start of a meeting. As input, it retrieves the list of meeting participants and past speech data from a database (e.g., MySQL). The output is the initialization of the AI ​​facilitator and the initialization of the sentiment analysis engine. Specifically, it runs a Python program to load the names, titles, and past speech data of participants. It also initializes the sentiment analysis engine using the Microsoft Azure Face API and Amazon Rekognition.

[0685] Step 2:

[0686] The device uses speech recognition technology to transcribe what is said during a meeting in real time. The input is audio data collected during the meeting via a microphone. The output is transcribed text data, which is sent to the server via WebSocket or HTTP request. Specifically, it uses the Google Cloud Speech-to-Text API or IBM Watson Speech to Text to generate text data from the input audio data.

[0687] Step 3:

[0688] The server receives the submitted transcription data and automatically updates the minutes with speaker information. It receives the transcription data and speaker identification information as input. The output is the updated minutes data, which is shared with all participants in real time. Specifically, it runs a Python program that adds new utterances to the database and updates the current minutes.

[0689] Step 4:

[0690] The server tracks each participant's speech volume (number of times spoken and duration) in real time. It uses transcription data and time information as input. The output is speech volume data for each participant, which is displayed on a dashboard. Specifically, it runs a tracking algorithm, calculates the speech volume for each participant, and stores it in a database.

[0691] Step 5:

[0692] When a participant who has not spoken much is identified, the server sends a notification to that participant's device encouraging them to speak. Speech volume data is used as input. The output is a notification message encouraging them to speak. Specifically, the server uses the Slack API to send the message "You haven't spoken much. Do you have any input?"

[0693] Step 6:

[0694] The server uses natural language processing (NLP) technology to analyze the content of the meeting in real time. The input is the transcription data. The output is a judgment result on whether the discussion has gone off track. Specifically, it uses OpenAI GPT-3 and Google BERT to analyze the input data and determine whether there has been any deviation.

[0695] Step 7:

[0696] When it detects that the discussion is veering off-topic, the server generates a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda." and notifies the user. The deviation detection result is used as input. The output is a suggestion message to return the discussion. The specific operation is to generate a suggestion message and send it to the user via Slack or email.

[0697] Step 8:

[0698] The device extracts key points and decisions from what is being said. It uses transcription data as input. The output is key points that are displayed in the whiteboard software. Specifically, it uses NLP technology to extract key information and display it on Miro or Microsoft Whiteboard.

[0699] Step 9:

[0700] The server uses emotion analysis technology to analyze the user's emotional state in real time. Facial expression data and voice tone data are collected as input. The output is analyzed emotional data. Specifically, emotions are analyzed using Amazon Rekognition and Microsoft Azure Face API, and the results are reflected in the system.

[0701] Step 10:

[0702] Based on the analyzed emotional data, the server sends appropriate notifications to other participants according to the user's emotions. Emotional data is used as input. The output is a notification message based on the emotional state. A specific operation would be to generate a notification such as "Tanaka seems to be feeling stressed. Please be considerate in the progress of the discussion" and send it to other participants.

[0703] (Application example 2)

[0704] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0705] In modern factories, many meetings are held for production management and business improvement, but it is common for not all participants to speak up actively, and discussions to go off track. It is also difficult to grasp the emotional state of participants, making stress management and improving the quality of communication challenging. Therefore, a system that can ensure efficient and productive meeting progress while also taking into consideration the emotions of participants is needed.

[0706] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes an automated conference record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, and an emotion recognition engine. This allows for real-time transcription of comments made during a conference, promotes active participation by all participants, prevents the discussion from going off track, and visually displays key points, enabling the conference to proceed efficiently. Furthermore, by understanding participants' emotions in real time and taking into consideration any stress or dissatisfaction, it is possible to create a healthier communication environment.

[0707] An "automated meeting record generator" is a system that automatically records what is said in a meeting, transcribes it, and generates minutes in real time.

[0708] The "participant speech volume tracking means" is a system that records and manages the number of speeches and speech duration of each participant during a conference in real time.

[0709] The "speech promotion means" is a system that automatically sends a notification to participants who make few comments to encourage them to speak.

[0710] The "digression detection and discussion guidance means" is a system that detects when a meeting discussion deviates from the planned topic and provides instructions or suggestions to bring the discussion back to the original topic.

[0711] The "important point visualization means" is a system that extracts important statements and decisions made during a meeting and displays them visually.

[0712] The "emotion recognition engine" is a system that analyzes participants' emotions in real time based on their speech, facial expressions, tone of voice, etc., and provides feedback on the results.

[0713] The present invention provides a system for supporting efficient and productive meetings in factories, which is composed of a server, terminals, and users for data processing and interaction.

[0714] Initialization

[0715] The server creates an instance of the AI ​​facilitator when a meeting starts. Basic information about each user participating in the meeting (such as name, position, and past speech data) is loaded from the database. At the same time, the emotion recognition engine is initialized, and emotional data is collected in real time by recognizing the user's facial expressions and analyzing their voice tone.

[0716] Automatic meeting minutes creation

[0717] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, and the transcribed content and speaker identification information are sent to a server, which receives the data and automatically updates the minutes.

[0718] Encouraging participants to speak up

[0719] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them notifications encouraging them to speak. This promotes active participation by everyone in the meeting.

[0720] Derailment Detection and Discussion Guide

[0721] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user.

[0722] Visualization of key points

[0723] The device extracts important points and decisions from what is being said during the meeting and displays them in real time. By visually displaying important points, users can easily follow the flow of the discussion, making meetings more efficient.

[0724] Introduction of emotion recognition engineers

[0725] The server uses an emotion recognition engine to analyze the user's emotional state in real time. Emotions are identified from the user's facial expressions, tone of voice, and content of remarks, and the emotional data is fed back to the system. For example, if a user is feeling stressed, this information is notified to other participants, encouraging them to take the discussion further and adjust its content accordingly.

[0726] Specific examples

[0727] Suppose a project progress meeting is being held at a factory. At the start of the meeting, the server initializes the AI ​​facilitator and emotion recognition engine, and loads information about the participants (User A, User B, and User C). As the meeting progresses, if User A says, "This project is going well," the device transcribes this statement in real time and sends it to the server. The server records this statement in the minutes and simultaneously tracks the amount of speech that User A made.

[0728] If it becomes clear that User B has barely spoken, the server sends User B a notification asking, "Can I have your opinion?" If the discussion strays from the project's progress and begins talking about recent news, the server sends a suggestion saying, "The discussion is drifting. Let's get back to the original topic." The device then displays the key point: "Project progress is going well."

[0729] On the other hand, if the emotion recognition engine detects that User A is feeling extremely stressed, the server will send a notification to other participants saying, "User A is feeling stressed. Please be considerate in the discussion." This will allow the meeting to proceed efficiently and productively, and by taking into consideration the emotions of the participants, smooth communication will be achieved.

[0730] Prompt Sentence Examples

[0731] Please implement a function that allows the AI ​​facilitator to track the amount of speech participants make in online meetings and send notifications to encourage participants who are not speaking often. Also, please add a function that uses an emotion recognition engine to analyze the emotional state of participants and notify other participants if they are feeling stressed.

[0732] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0733] Step 1:

[0734] At the start of a meeting, the server instantiates an AI facilitator and loads basic participant information from a database. Input includes participant name, role, and past speech data. Based on this data, the server creates individual participant profiles as output and provides them to the AI ​​facilitator.

[0735] Step 2:

[0736] The server initializes the emotion recognition engine and prepares to collect the user's emotion data in real time. The inputs are facial expression recognition data, voice tone data, and speech content data, and the data is analyzed to output the user's emotional state.

[0737] Step 3:

[0738] The device uses speech recognition technology to transcribe speech in real time during a meeting. The input is the user's speech audio data, and the output is the transcribed text data. This allows the speech content to be recorded as text.

[0739] Step 4:

[0740] The device sends the transcribed speech and the speaker's identification information to the server. The input is the speech transcription data and the speaker ID, and the output is a packet containing these data. The server receives this data and automatically updates the minutes.

[0741] Step 5:

[0742] The server tracks the amount of speech of each participant in real time. The input is speech transcription data and speaker ID, and accumulates the number of speeches and speech duration in a database. The output is speech volume data for each participant.

[0743] Step 6:

[0744] The server sends a notification to participants who speak less to encourage them to speak. The input is the speech volume data of each participant, and the output is a notification message to participants who speak less.

[0745] Step 7:

[0746] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. The input is speech transcription data, and the output is a flag indicating whether the topic is deviating. If a derailment is detected, the server generates instructions to return to the original topic and notifies the user via their device.

[0747] Step 8:

[0748] The device extracts important points and decisions from what is said during the meeting and displays them on the screen. The input is the transcription data of the speech, and the output is a visual representation of the extracted important points. By visually displaying the important points, users can easily follow the flow of the discussion.

[0749] Step 9:

[0750] The server uses an emotion recognition engine to analyze the user's emotional state. The inputs are the user's facial expression data, voice tone data, and speech content data, and the output is emotional state data. For example, if the user is in a stressful state, that information is notified to other participants.

[0751] Step 10:

[0752] At the end of the meeting, the server saves all meeting minutes, emotion data, and speech volume data, and outputs them as a report if necessary. The input is all the data collected during the meeting, and the output is the final meeting report. This leaves a record to review later.

[0753] 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 a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0754] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0755] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0756] [Third embodiment]

[0757] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0758] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0759] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0760] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0761] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0762] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0763] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0764] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0765] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[0766] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0767] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0768] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. 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."

[0769] Initialization

[0770] When a meeting starts, the server creates an instance of the AI ​​facilitator. Next, it loads basic information about participants (such as their names, positions, and past speech data) from the database. It also initializes a minutes template and prepares the data structures necessary for the meeting to proceed.

[0771] Automatic meeting minutes creation

[0772] The device uses speech recognition technology to transcribe each statement in real time during the meeting. The transcribed content and speaker identification information are sent to the server as the meeting progresses. The server receives the data and automatically records it as meeting minutes.

[0773] Encouraging participants to speak up

[0774] The server tracks each participant's speech volume (number of times and duration) in real time. It identifies participants who speak less and sends them notifications encouraging them to speak. Notifications are automatically generated according to the progress of the conference and are delivered to users.

[0775] Derailment Detection and Discussion Guide

[0776] The server uses natural language processing technology to determine in real time whether a discussion has deviated from the planned topic. If a deviation from the topic is detected, the server generates suggestions to bring the discussion back on topic and notifies the user, thereby reducing the risk of the discussion dragging on unnecessarily or drifting off topic.

[0777] Visualization of key points

[0778] The device extracts important points and decisions from what is being said during the meeting and displays them on a whiteboard in real time. This whiteboard is constantly updated during the meeting, making it easier for users to visually follow the important points. This allows meetings to proceed efficiently without losing track of the flow of discussion or important points.

[0779] Specific examples

[0780] For example, let's say a company is holding a project progress meeting. When the meeting begins, the server launches the AI ​​facilitator and loads the data of the participants, "Tanaka," "Yamada," and "Sato." As the meeting progresses, if Tanaka says, "This project is going well," his remarks are transcribed in real time by the device. The transcription data and Tanaka's identification information are sent to the server and recorded in the minutes.

[0781] Next, tracking reveals that Yamada has barely spoken anything. The server sends Yamada a notification saying, "You haven't spoken much. Do you have any input?" This gives Yamada a chance to express his opinion.

[0782] If, at some point, Sato starts talking about recent news, the server will determine that the remark is unrelated to the project's progress and send the user a suggestion: "The discussion is veering off-topic. Let's return to the main agenda."

[0783] Furthermore, the device automatically displays important points in the meeting, such as "Project progress is going smoothly," on the whiteboard, allowing users to visually confirm important matters during the meeting in real time.

[0784] As described above, this system provides various functions to ensure that online meetings proceed efficiently and productively, contributing to improving the quality of meetings.

[0785] The processing flow will be explained below.

[0786] Step 1: Initialization

[0787] The server creates an instance of the AI ​​facilitator at the start of the meeting, which launches a system to assist in the progress of the online meeting.

[0788] Step 2: Load participant information

[0789] The server loads basic information of participants (such as name, position, and past comment data) from the database, which enables identification of each participant and support based on past data.

[0790] Step 3: Initialize the meeting minutes template

[0791] The server initializes a meeting template, which provides the data structures necessary for the minutes format and flow of the meeting.

[0792] Step 4: Start real-time transcription

[0793] As the meeting progresses, the device uses speech recognition technology to transcribe each utterance in real time, and the speech is automatically converted into text.

[0794] Step 5: Sending speech data

[0795] The device sends the transcribed speech and the speaker's identity to the server, where the speech is ready to be recorded.

[0796] Step 6: Update the minutes

[0797] The server automatically updates the meeting minutes based on the received speech data, ensuring that each speech is recorded appropriately.

[0798] Step 7: Track your speech volume

[0799] The server tracks each participant's speech volume (number of times and duration) in real time, allowing the activity level of participants to be monitored.

[0800] Step 8: Identify quiet participants

[0801] The server identifies participants who are less active: participants whose number of comments falls below a certain threshold.

[0802] Step 9: Send a prompt

[0803] The server generates a notification for quiet participants to speak up and sends it to the user, with a message such as "You haven't spoken much. Do you have any input?"

[0804] Step 10: Derailment detection

[0805] The server uses natural language processing technology to determine in real time whether the current comment is off-topic.

[0806] Step 11: Sending a discussion guide notification

[0807] If the server detects a digression, it generates a suggestion to bring the discussion back on topic and sends it to the user: "The discussion is veering off-topic. Let's return to the main agenda."

[0808] Step 12: Extracting key points

[0809] The device extracts important points and decisions from what is said during the meeting.

[0810] Step 13: Display on the Whiteboard

[0811] The device displays the extracted important points on a whiteboard in real time, allowing users to visually confirm the important points.

[0812] Step 14: End the meeting and export the minutes

[0813] At the end of the meeting, the server exports the accumulated minutes in an editable format, which users can review and modify.

[0814] Through these steps, this system enables online meetings to proceed efficiently and productively.

[0815] Example 1

[0816] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0817] Existing online meeting systems face the challenge of efficiently managing a wide range of tasks, including meeting progress, minutes creation, participant prompting, digression detection and discussion guidance, and visualization of key points. This can lead to poor meeting quality, redundant discussions, and participants being reluctant to speak up. Furthermore, creating meeting minutes manually is time-consuming and labor-intensive, and lacks accuracy.

[0818] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0819] In this invention, the server includes a meeting start detection means, an AI facilitator instance generation means, a participant information database loading means, a minutes template initialization means, a speech transcription means using speech recognition, a speech data transmission means, a speech data recording means, a participant speech volume tracking means, a speech promotion notification generation means, a speech promotion notification transmission means, a discussion digression detection means, a discussion guide notification generation means, a discussion guide notification transmission means, an important point extraction means, and a whiteboard display means. This makes it possible to manage the progress of the meeting, create minutes, encourage participants to speak, guide the discussion, and visualize important points in real time.

[0820] The "conference start detection means" is a means for recognizing the start of a conference and for the system to automatically start the necessary initial processing.

[0821] An "AI facilitator instance generation means" is a means for generating an instance of an AI model to smoothly proceed with a meeting.

[0822] The "participant information database loading means" is a means for loading basic information of conference participants from a database.

[0823] The "minutes template initialization means" is a means for initializing a template for creating minutes of a meeting.

[0824] A "means for transcribing speech using speech recognition" is a means for using speech recognition technology to transcribe speech during a meeting in real time.

[0825] The "utterance data transmission means" is a means for transmitting the transcribed utterance data to the server.

[0826] The "utterance data recording means" is a means for recording received utterance data in a database.

[0827] The "participant speech volume tracking means" is a means for tracking the speech volume of each participant during a conference in real time.

[0828] The "means for generating a speech prompting notification" is a means for generating a notification that prompts a participant who makes few comments to speak.

[0829] The "speech prompting notification sending means" is a means for sending the generated speech prompting notification to the relevant participant.

[0830] The "discussion deviation detection means" is a means for detecting whether the discussion in the meeting has deviated from the planned topic.

[0831] The "discussion guide notification generating means" is a means for generating a suggestion to bring the discussion back to the main topic when the discussion goes off track.

[0832] The "discussion guide notification sending means" is a means for sending the generated discussion guide notification to the participants.

[0833] The "means for extracting important points" is a means for extracting important points and decisions from the content of statements made during a meeting.

[0834] The "whiteboard display means" is a means for displaying the extracted important points on a whiteboard in real time.

[0835] The present invention is a system for conducting online meetings efficiently and productively. This system automates various processes from the start to the end of a meeting and combines multiple means to improve the quality of the meeting.

[0836] Initialization

[0837] When the server detects the start of a meeting, it creates an instance of an AI facilitator. This uses an AI framework such as TensorFlow or PyTorch. Next, it loads participant information from a database (e.g., MySQL or PostgreSQL). This includes participant names, positions, and past speech data. Next, the server initializes a minutes template. This template uses a predefined file in Markdown format or similar.

[0838] Automatic meeting minutes creation

[0839] During a meeting, the device uses speech recognition technology such as the Google Speech-to-Text API or IBM Watson Speech to Text to transcribe each utterance in real time. The transcribed utterances and the speaker's identity are then sent to the server via an HTTP POST request. The server receives this data and automatically records it in a database such as Elasticsearch.

[0840] Encouraging participants to speak up

[0841] The server uses MongoDB or similar to track the amount of speech each participant makes (number of times they make a comment and how long they speak). For participants who make little comments, the server uses the Slack API or Microsoft Teams API to generate and send notifications encouraging them to speak. These notifications are generated automatically and sent to users.

[0842] Derailment Detection and Discussion Guide

[0843] The server uses natural language processing techniques, such as Hugging Face's Transformers library, to determine in real time whether the discussion is deviating from the planned topic. If a deviation is detected, the server compares it with pre-defined agenda information, generates suggestions to bring the discussion back on topic, and communicates them to the user. This notification is sent automatically at the appropriate time.

[0844] Visualization of key points

[0845] The device uses natural language processing libraries such as SpaCy and NLTK to extract key points and decisions from what is being said during the meeting. The extracted key points are displayed in real time on a whiteboard service such as Miro Board or Microsoft Whiteboard. This display is constantly updated as the meeting progresses, making it easy for users to visually follow the key points.

[0846] Specific examples

[0847] For example, when a company holds a project status meeting, the server detects the start of the meeting through the Google Calendar API, creates an instance of an AI facilitator using TensorFlow, and then loads the data of the participants, "Tanaka," "Yamada," and "Sato," from a MySQL database.

[0848] As the meeting progresses, if Tanaka says, "This project is going well," the device will transcribe his speech in real time using the Google Speech-to-Text API. The transcription data and Tanaka's identification information will be sent to the server via an HTTP POST request and recorded in Elasticsearch.

[0849] Once MongoDB data tracks that Yamada has barely spoken, the server uses the Slack API to send a notification to Yamada saying, "You haven't spoken much. Do you have any input?"

[0850] If Sato starts talking about recent news midway through, the server will use Hugging Face's Transformers library to determine that his remarks are unrelated to the project's progress and send the user a suggestion: "The discussion is veering off-topic. Let's return to the main agenda."

[0851] The device automatically displays the key point of the meeting, such as "Project progress is going well," on the Miro board, allowing users to visually check important matters during the meeting in real time through the whiteboard.

[0852] Prompt Sentence Examples

[0853] Below is an example prompt for requesting minutes of a project status meeting:

[0854] A project progress meeting was held. The participants were Tanaka, Yamada, and Sato. Tanaka stated, "This project is progressing smoothly." Based on the minutes, please create and propose the following action items:

[0855] As described above, the present invention provides various functions for conducting online conferences efficiently and productively, thereby contributing to improving the quality of conferences.

[0856] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0857] Processing flow

[0858] Step 1:

[0859] The server detects when a meeting starts. To do this, it uses the Google Calendar API to monitor scheduled events. The input is the calendar information obtained from the API, and the output is a meeting start event.

[0860] Step 2:

[0861] The server creates an instance of the AI ​​facilitator. First, it loads the AI ​​model using the TensorFlow or PyTorch library and creates an instance. The input is the pre-trained AI model, and the output is the generated instance of the AI ​​facilitator.

[0862] Step 3:

[0863] The server loads participant information from a database. It issues SQL queries to a database such as MySQL or PostgreSQL to retrieve participant names, positions, and past comment data. The input is the database connection information and query, and the output is the retrieved participant information.

[0864] Step 4:

[0865] The server initializes the minutes template. It reads a predefined Markdown template file and prepares the initial state. The input is the path to the template file, and the output is the initialized minutes template.

[0866] Step 5:

[0867] The device transcribes speech during meetings in real time. It uses the Google Speech-to-Text API to convert audio data captured from the microphone into text data. The input is audio data, and the output is transcribed text data.

[0868] Step 6:

[0869] The device sends the transcribed speech data and the speaker's identification information to the server using an HTTP POST request. The input is the speech data and the identification information, and the output is the success status of the data transmission to the server.

[0870] Step 7:

[0871] The server records the received utterance data in Elasticsearch, analyzes the data, and adds it to the appropriate index. The input is the received utterance data, and the output is the index information stored in Elasticsearch.

[0872] Step 8:

[0873] The server tracks the amount of speech of each participant in real time. It records the speech data in MongoDB and executes aggregation queries to calculate the number of speeches and speech duration. The input is the received speech data, and the output is the tracked speech volume.

[0874] Step 9:

[0875] The server generates notifications to encourage participants who speak less to speak. The notification content is automatically generated using the Slack API or Microsoft Teams API. The input is the tracked speech volume data, and the output is the generated notification message.

[0876] Step 10:

[0877] The server sends the generated notification to the appropriate user using the Slack API or Microsoft Teams API. The input is the generated notification message, and the output is the success status of the notification sending.

[0878] Step 11:

[0879] The server detects deviations in the discussion. It uses Hugging Face's Transformers library to compare the content of comments with the agenda and determine whether or not a deviation has occurred. The input is the comment data and agenda information, and the output is the deviation detection result.

[0880] Step 12:

[0881] When a digression is detected, the server generates a suggestion to bring the discussion back to the main topic. The suggestion content is generated using a generative AI model. The input is the detected digression result, and the output is the generated suggestion message.

[0882] Step 13:

[0883] The server notifies the user of the proposal message. It sends the notification using the chat application's API. The input is the generated proposal message, and the output is the success status of sending the notification.

[0884] Step 14:

[0885] The device extracts important points from what is said during the meeting. It uses SpaCy and NLTK to analyze the speech and extract important points and decisions. The input is the speech data, and the output is the extracted important points.

[0886] Step 15:

[0887] The device displays the extracted key points on a whiteboard. It updates the display in real time using the APIs of Miro Board and Microsoft Whiteboard. The input is the extracted key points, and the output is the display result on the whiteboard.

[0888] Through the above processing steps, the system manages the progress of the meeting, creates minutes, encourages participants to speak up, directs the discussion, and visualizes important points.

[0889] (Application example 1)

[0890] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0891] Modern meeting systems lack features such as taking minutes, encouraging participants to speak up, and detecting derailed discussions. Some participants often don't speak much, or the discussion often goes off track, wasting time. Furthermore, participants can lose track of the flow of the discussion because they cannot visually track important points and decisions. These issues directly affect production efficiency, particularly in factory progress meetings, so there is a high need for a system that can quickly and efficiently conduct meetings.

[0892] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0893] In this invention, the server includes an automated meeting record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, a means for transcribing speeches made during a meeting in real time using speech recognition technology and recording the results, a means for sending notifications to participants who speak less to encourage them to speak, a means for determining in real time whether the discussion has deviated from the planned topic using natural language processing technology and generating suggestions, and a means for displaying key points and decisions on a whiteboard in real time, thereby enabling improvements in the efficiency and productivity of meetings.

[0894] The "automated meeting record generator" is a function that automatically recognizes statements made during a meeting and creates minutes in real time.

[0895] The "participant speech volume tracking means" is a function for measuring and recording the number of speeches and speech duration of each conference participant in real time.

[0896] The "speech promotion means" is a function for sending notifications to participants who do not speak much, encouraging them to speak.

[0897] "Derailment detection and discussion guidance means" is a function that uses natural language processing technology to determine whether a discussion has deviated from the planned topic and suggests an appropriate direction for the discussion.

[0898] The "important point visualization means" is a function that displays important points and decisions made during a meeting on a display device in real time, allowing participants to visually grasp important information.

[0899] "Speech recognition technology" is a technology for converting speech into text, and in this invention is a technology used to transcribe statements made during a meeting in real time.

[0900] "Real-time transcription means" is a function that uses voice recognition technology to instantly convert meeting speech into text and record it.

[0901] "Means for encouraging participants who speak infrequently to speak by sending notifications" is a function for tracking the volume of comments in real time and sending notifications to participants who speak infrequently to encourage them to speak.

[0902] "Natural language processing technology" is a technology for understanding and analyzing human language, and in this invention, it is a technology used to detect digressions in discussions and provide appropriate discussion guidance.

[0903] "A means of displaying important points and decisions on a whiteboard in real time" is a function that instantly reflects important statements and decisions made during a meeting on a display device such as a whiteboard, allowing all participants to visually confirm them.

[0904] The system embodying this invention is designed to streamline factory progress meetings and is composed of a server, terminals, and users with various functions. Specifically, it is composed of an automated meeting record generation means, a means for tracking the amount of speech by participants, a means for encouraging speech, a means for detecting deviations and guiding discussions, and a means for visualizing important points.

[0905] When a meeting starts, the server creates an instance of the AI ​​facilitator and loads basic information about participants (such as their names, positions, and past speech data) from a database. This database uses a system such as SQLite. It also initializes a minutes template and prepares the data structure necessary for the meeting to proceed.

[0906] The device uses speech recognition technology to transcribe speech in real time during a meeting. The transcribed speech and speaker identification information are sent to a server, which records the data as meeting minutes in real time. The speech recognition technology used here includes the speech_recognition library.

[0907] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them a notification to encourage them to speak. This notification is sent to an application on a smartphone or tablet.

[0908] The digression detection and discussion guidance mechanism uses natural language processing (NLP) technology to determine in real time whether a discussion has deviated from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user. The NLP technology used here includes the spacy library.

[0909] The important points visualization feature extracts important points and decisions from what is said during the meeting and displays them in real time on a whiteboard or display, allowing users to visually confirm important points during the meeting.

[0910] Examples:

[0911] Consider a situation where a meeting is being held in a factory to discuss the progress of a manufacturing process. When the meeting begins, the server launches the AI ​​facilitator and loads the data of the participants, "Participant A," "Participant B," and "Participant C." The device transcribes Participant A's statement, "Production line 1 is progressing as planned," in real time and sends it to the server. The server records this statement in the minutes in real time.

[0912] Next, if tracking reveals that participant B is not speaking much, the server sends a notification to participant B's device saying, "Please speak more." Furthermore, if participant C goes off on a tangent about a topic unrelated to the discussion, the server generates and notifies the participant B with a suggestion saying, "The conversation is going off topic. Let's return to the original topic."

[0913] Important comments are immediately displayed on displays in the factory, allowing all participants to share the latest information and conduct meetings efficiently.

[0914] Example prompts for generative AI models:

[0915] "Please propose an application that utilizes an AI facilitator with the functions of automatic minutes creation, prompting, digression detection, and visualization of key points in status meetings on a manufacturing line. Please also include a specific usage scenario and example program."

[0916] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0917] Step 1:

[0918] Execution of initialization process

[0919] The server creates an instance of the AI ​​facilitator when the meeting starts. It also loads basic information about participants (such as name, position, and past speech data) from a database. This information is stored using SQLite or similar. It also initializes a minutes template and prepares the data structures necessary for the meeting to proceed.

[0920] Input: Conference start signal, database (basic information of participants)

[0921] Output: An instance of the AI ​​facilitator, an initialized data structure, and a meeting minutes template.

[0922] Specific operation: The server uses a program such as Java or Python to access the conference management server and load participant data.

[0923] Step 2:

[0924] Real-time audio transcription

[0925] During the meeting, the device uses speech recognition technology (e.g., the speech_recognition library) to transcribe each utterance in real time, and then sends the transcribed utterances to the server.

[0926] Input: Audio data (speech during a meeting)

[0927] Output: Transcription data (text format), speaker identification information

[0928] Specific operation: The device receives voice input from the microphone, converts it into text data, and sends it to the server.

[0929] Step 3:

[0930] Real-time meeting minutes updates

[0931] The server updates the minutes in real time based on the received transcription data, thereby recording every word spoken during the meeting.

[0932] Input: Transcription data, speaker identification information

[0933] Output: Updated minutes

[0934] Specific operation: The server updates and stores the minutes data using a list or tree structure.

[0935] Step 4:

[0936] Participant speech volume tracking

[0937] The server tracks each participant's speech volume (number of times and duration) in real time and identifies participants who are not speaking much, allowing the server to provide appropriate feedback.

[0938] Input: Speech data, speaker identification information

[0939] Output: Tracking data (speech volume record)

[0940] Specific operation: The server measures the volume of messages using timestamp analysis and statistical functions.

[0941] Step 5:

[0942] Sending a prompt to speak

[0943] For participants who do not speak much, the server generates a notification encouraging them to speak and sends it to their smartphone or tablet.

[0944] Input: Tracking data, target participant identification information

[0945] Output: Speech prompt notification

[0946] Specific operation: The server uses a notification service (e.g., Firebase Cloud Messaging) to send a notification to the target participant's device.

[0947] Step 6:

[0948] Derailment detection and guide suggestions

[0949] The server uses natural language processing techniques (such as the Spacy library) to determine in real time whether the discussion is deviating from the planned topic, and if a digression is detected, it generates suggestions to bring the discussion back on topic and notifies the user.

[0950] Input: Transcription data, topic information

[0951] Output: Derailment notice, discussion guide suggestion

[0952] Specific operation: The server uses natural language processing technology to perform text analysis and analyze the direction of the discussion.

[0953] Step 7:

[0954] Visualizing key points

[0955] The device automatically extracts important points and decisions during the meeting and displays them on a whiteboard in real time, making it easier for participants to visually follow important matters during the meeting.

[0956] Input: Speech data, important points data

[0957] Output: Display (whiteboard)

[0958] Specific operation: The device uses the display output function to display important points on the whiteboard.

[0959] Step 8:

[0960] Data processing after the meeting

[0961] After the meeting ends, the server performs a final check and saves the minutes data, preparing it for distribution to participants.

[0962] Input: Final minutes data

[0963] Output: minutes for distribution

[0964] Specific operation: The server uses a database system to store the minutes data and export it in an appropriate format.

[0965] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0966] The present invention relates to a system and method that provides multiple functions for efficiently and productively conducting online meetings, in particular, a combination of an automated meeting record generator, a participant speech volume tracking unit, a speech promotion unit, a digression detection and discussion guide unit, a key point visualization unit, and an emotion engine that recognizes user emotions.

[0967] Initialization

[0968] The server creates an instance of the AI ​​facilitator at the start of a meeting. At this stage, it loads basic information about each user participating in the meeting (such as name, position, and past speech data) from a database. At the same time, the emotion engine is initialized, and emotional data is ready to be collected in real time through facial expression recognition and voice tone analysis.

[0969] Automatic meeting minutes creation

[0970] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, and the transcribed content and speaker identification information are sent to a server, which receives the data and automatically updates the minutes.

[0971] Encouraging participants to speak up

[0972] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them notifications encouraging them to speak. This promotes active participation by everyone in the meeting.

[0973] Derailment Detection and Discussion Guide

[0974] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user.

[0975] Visualization of key points

[0976] The device extracts important points and decisions from what is being said during the meeting and displays them on the whiteboard in real time. By visually displaying important points, users can easily follow the flow of the discussion, making meetings more efficient.

[0977] Introducing the Emotion Engine

[0978] The server uses an emotion engine to analyze the user's emotional state in real time. Emotions are identified from the user's facial expressions, tone of voice, and content of comments, and the emotional data is fed back to the system. For example, if a user is feeling stressed, this information is notified to other participants, encouraging them to take the discussion further and adjust its content accordingly.

[0979] Specific examples

[0980] Suppose a company is holding a project progress meeting. At the start of the meeting, the server initializes the AI ​​facilitator and emotion engine and loads the information of the participants (Tanaka, Yamada, Sato). As the meeting progresses, if Tanaka says, "This project is going well," this statement is transcribed in real time and sent to the server. The server records this statement in the minutes and simultaneously tracks the amount of speech Tanaka made.

[0981] When it becomes clear that Yamada has barely spoken, the server sends Yamada a notification saying, "You haven't spoken much. Do you have any input?" If the discussion veers away from project progress and starts talking about recent news, the server sends a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda."

[0982] The device displays the important point, "The project is progressing smoothly," on the whiteboard. On the other hand, if the emotion engine detects that Tanaka is feeling very stressed, the server sends a notification to other participants saying, "Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[0983] In this way, the present invention significantly improves the efficiency and productivity of online meetings. The introduction of an emotion engine allows smooth communication while taking into consideration the emotional state of users during a meeting.

[0984] The processing flow will be explained below.

[0985] Step 1: Initialize the system

[0986] The server instantiates the AI ​​facilitator and emotion engine at the start of a meeting. This initialization ensures that the entire system is ready to assist in running the meeting.

[0987] Step 2: Load participant information

[0988] The server loads basic information of participants (such as name, title, and past comment data) from the database. This information is used for subsequent comment tracking and sentiment analysis.

[0989] Step 3: Start speech recognition

[0990] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, instantly converting what is said during the meeting into text data.

[0991] Step 4: Sending speech data

[0992] The device sends the transcribed speech and the speaker's identity to the server, where it is ready to be recorded.

[0993] Step 5: Update the minutes

[0994] The server automatically updates the meeting minutes based on the received speech data, allowing the content of speech to be recorded in real time.

[0995] Step 6: Track your speech volume

[0996] The server tracks each participant's speech volume (number of times and duration) in real time, and the tracking data is used to encourage subsequent speech.

[0997] Step 7: Identify quiet participants

[0998] The server identifies participants who are quiet based on the tracking data. Participants who are below a certain threshold are identified.

[0999] Step 8: Send a prompt

[1000] The server generates a notification for participants who have not spoken much and sends it to the user to encourage them to speak up. Example of notification: "You haven't spoken much. Do you have any input?"

[1001] Step 9: Derailment detection with natural language processing

[1002] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the topic, and if there are signs of derailment, subsequent processing will be based on that information.

[1003] Step 10: Sending a discussion guide notification

[1004] If the server detects a digression, it generates a suggestion to bring the discussion back on topic and sends it to the user. Example notification: "The discussion is veering off-topic. Let's return to the main agenda."

[1005] Step 11: Extract and display key points

[1006] The device extracts important points and decisions from what is being said during the meeting and displays them on the whiteboard in real time in a format that is easy to follow visually.

[1007] Step 12: Collect and analyze emotion data

[1008] The server uses an emotion engine to collect emotional data from participants' facial expressions, tone of voice, and content of remarks, etc. Based on the analyzed emotional data, feedback is provided to the progress of the meeting in real time.

[1009] Step 13: Communicating Emotional State

[1010] When the server detects an important emotional state, such as a specific user being stressed, it notifies other participants of that information. Example of notification: "Mr. Tanaka seems to be feeling stressed. Please be considerate in the progress of the discussion."

[1011] Step 14: End the meeting and export data

[1012] At the end of the meeting, the server exports the accumulated minutes and emotion data in an editable format, which users can download and review as needed.

[1013] Through the above steps, this system incorporating emotion analysis not only enables online meetings to proceed efficiently and productively, but also ensures smooth communication that takes into consideration the emotions of participants.

[1014] Example 2

[1015] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1016] In online meetings, it is extremely difficult to automatically create minutes, track the amount of speech, prevent discussions from going off track, and visualize important points in real time. Another challenge is accurately understanding participants' emotional states and reflecting them in the progress of the discussion. By solving these challenges, it is necessary to realize efficient and productive online meetings and reduce user stress.

[1017] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1018] In this invention, the server includes an automated conference record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, and a sentiment analysis means, which make it possible to transcribe speeches made during a conference in real time, track the speech volume of each participant, prevent digressions and bring the discussion back to the main topic, visualize key points, and analyze sentiment data to reflect the progress of the conference.

[1019] "Automated meeting record generation means" is a technology that automatically records what is said in online meetings and generates minutes of the meeting.

[1020] The "participant speech volume tracking means" is a technology for measuring and recording the number of speeches and speech duration of each participant during a conference in real time.

[1021] The "speech promotion means" is a function for sending a notification to participants who do not speak much to encourage them to speak.

[1022] The "digression detection and discussion guidance means" is a technology for determining whether a discussion during a meeting has deviated from the planned topic and generating appropriate suggestions to bring the discussion back on topic.

[1023] "Key point visualization means" is a technology for extracting important points and decisions from what is said during a meeting and displaying them visually.

[1024] The "emotion analysis means" is a technology for analyzing the user's emotional state in real time using facial expression recognition technology and voice tone analysis technology.

[1025] "Speech recognition technology" is a technology for converting voice data into text data.

[1026] "Natural language processing technology" is a technology for analyzing text data, understanding its meaning, and processing it.

[1027] "Whiteboard software" is a digital tool for visually sharing information during meetings.

[1028] "Real time" is a time concept that refers to processing and display occurring almost simultaneously.

[1029] This invention relates to a system and method for efficiently and productively conducting online meetings, which combines an automated meeting record generation unit, a participant speech volume tracking unit, a speech promotion unit, a digression detection and discussion guide unit, a key point visualization unit, and a sentiment analysis unit. The operation of each unit will be described in detail below.

[1030] Initialization

[1031] The server instantiates an AI facilitator at the start of a meeting. This instantiation is done using a Python program that loads each participant's name, title, and past speech data from a database (e.g., MySQL). It also initializes an emotion engine using the Microsoft Azure Face API and Amazon Rekognition to recognize the user's facial expressions and analyze their speech tone.

[1032] Automatic meeting minutes creation

[1033] The device uses speech recognition technology to transcribe what is said during the meeting in real time using the Google Cloud Speech-to-Text API and IBM Watson Speech to Text. The transcribed content and speaker identification information are sent to a server, which then automatically updates the meeting minutes based on this data.

[1034] Encouraging participants to speak up

[1035] The server tracks each participant's speech volume (number of times and duration) in real time. This allows it to identify participants who are not speaking much and send them a notification to encourage them to speak up. For example, it uses the Slack API to send a message like, "You haven't spoken much. Do you have any input?"

[1036] Derailment Detection and Discussion Guide

[1037] The server uses natural language processing (NLP) technology to analyze the content of the meeting in real time. Specifically, it uses OpenAI GPT-3 and Google BERT to determine whether the discussion is deviating from the planned topic. If a deviation is detected, it generates a suggestion to the user, saying, "The discussion is veering off-topic. Let's return to the main agenda."

[1038] Visualization of key points

[1039] The device uses NLP technology to extract key points and decisions from what is being said, and the extracted content is displayed in real time on whiteboard software such as Miro or Microsoft Whiteboard.

[1040] Introducing sentiment analysis

[1041] The server uses emotion analysis technology to analyze the user's emotional state in real time. It uses Amazon Rekognition and Microsoft Azure Face API to analyze the user's facial expressions and voice tone, and identifies their emotional state based on that data. For example, if a user is feeling stressed, the server notifies other participants, saying, "Mr. Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[1042] Specific examples

[1043] Suppose a company is holding a project progress meeting. At the start of the meeting, the server initializes the AI ​​facilitator and sentiment analysis engine, and loads the information of the participants (Tanaka, Yamada, Sato) from the database. As the meeting progresses, if Tanaka says, "This project is going well," his remarks are transcribed in real time and sent to the server. The server updates the minutes based on this data, while simultaneously tracking the amount of speech Tanaka made.

[1044] If John has barely spoken, the server sends him a notification saying, "You haven't spoken much. Do you have any input?" If the discussion veers away from project progress and starts talking about recent news, the server sends him a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda."

[1045] The device displays the key point on the whiteboard: "The project is progressing smoothly." However, if the emotion analysis engine detects that Tanaka is feeling extremely stressed, it sends a notification to other participants saying, "Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[1046] Prompt Sentence Examples

[1047] "Describe a system that generates real-time online meeting minutes. Detail the specific hardware and software used and how the data is processed. Also, mention the role of sentiment analysis."

[1048] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1049] Step 1:

[1050] The server creates an instance of the AI ​​facilitator at the start of a meeting. As input, it retrieves the list of meeting participants and past speech data from a database (e.g., MySQL). The output is the initialization of the AI ​​facilitator and the initialization of the sentiment analysis engine. Specifically, it runs a Python program to load the names, titles, and past speech data of participants. It also initializes the sentiment analysis engine using the Microsoft Azure Face API and Amazon Rekognition.

[1051] Step 2:

[1052] The device uses speech recognition technology to transcribe what is said during a meeting in real time. The input is audio data collected during the meeting via a microphone. The output is transcribed text data, which is sent to the server via WebSocket or HTTP request. Specifically, it uses the Google Cloud Speech-to-Text API or IBM Watson Speech to Text to generate text data from the input audio data.

[1053] Step 3:

[1054] The server receives the submitted transcription data and automatically updates the minutes with speaker information. It receives the transcription data and speaker identification information as input. The output is the updated minutes data, which is shared with all participants in real time. Specifically, it runs a Python program that adds new utterances to the database and updates the current minutes.

[1055] Step 4:

[1056] The server tracks each participant's speech volume (number of times spoken and duration) in real time. It uses transcription data and time information as input. The output is speech volume data for each participant, which is displayed on a dashboard. Specifically, it runs a tracking algorithm, calculates the speech volume for each participant, and stores it in a database.

[1057] Step 5:

[1058] When a participant who has not spoken much is identified, the server sends a notification to that participant's device encouraging them to speak. Speech volume data is used as input. The output is a notification message encouraging them to speak. Specifically, the server uses the Slack API to send the message "You haven't spoken much. Do you have any input?"

[1059] Step 6:

[1060] The server uses natural language processing (NLP) technology to analyze the content of the meeting in real time. The input is the transcription data. The output is a judgment result on whether the discussion has gone off track. Specifically, it uses OpenAI GPT-3 and Google BERT to analyze the input data and determine whether there has been any deviation.

[1061] Step 7:

[1062] When it detects that the discussion is veering off-topic, the server generates a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda." and notifies the user. The deviation detection result is used as input. The output is a suggestion message to return the discussion. The specific operation is to generate a suggestion message and send it to the user via Slack or email.

[1063] Step 8:

[1064] The device extracts key points and decisions from what is being said. It uses transcription data as input. The output is key points that are displayed in the whiteboard software. Specifically, it uses NLP technology to extract key information and display it on Miro or Microsoft Whiteboard.

[1065] Step 9:

[1066] The server uses emotion analysis technology to analyze the user's emotional state in real time. Facial expression data and voice tone data are collected as input. The output is analyzed emotional data. Specifically, emotions are analyzed using Amazon Rekognition and Microsoft Azure Face API, and the results are reflected in the system.

[1067] Step 10:

[1068] Based on the analyzed emotional data, the server sends appropriate notifications to other participants according to the user's emotions. Emotional data is used as input. The output is a notification message based on the emotional state. A specific operation would be to generate a notification such as "Tanaka seems to be feeling stressed. Please be considerate in the progress of the discussion" and send it to other participants.

[1069] (Application example 2)

[1070] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1071] In modern factories, many meetings are held for production management and business improvement, but it is common for not all participants to speak up actively, and discussions to go off track. It is also difficult to grasp the emotional state of participants, making stress management and improving the quality of communication challenging. Therefore, a system that can ensure efficient and productive meeting progress while also taking into consideration the emotions of participants is needed.

[1072] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes an automated conference record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, and an emotion recognition engine. This allows for real-time transcription of comments made during a conference, promotes active participation by all participants, prevents the discussion from going off track, and visually displays key points, enabling the conference to proceed efficiently. Furthermore, by understanding participants' emotions in real time and taking into consideration any stress or dissatisfaction, it is possible to create a healthier communication environment.

[1073] An "automated meeting record generator" is a system that automatically records what is said in a meeting, transcribes it, and generates minutes in real time.

[1074] The "participant speech volume tracking means" is a system that records and manages the number of speeches and speech duration of each participant during a conference in real time.

[1075] The "speech promotion means" is a system that automatically sends a notification to participants who make few comments to encourage them to speak.

[1076] The "digression detection and discussion guidance means" is a system that detects when a meeting discussion deviates from the planned topic and provides instructions or suggestions to bring the discussion back to the original topic.

[1077] The "important point visualization means" is a system that extracts important statements and decisions made during a meeting and displays them visually.

[1078] The "emotion recognition engine" is a system that analyzes participants' emotions in real time based on their speech, facial expressions, tone of voice, etc., and provides feedback on the results.

[1079] The present invention provides a system for supporting efficient and productive meetings in factories, which is composed of a server, terminals, and users for data processing and interaction.

[1080] Initialization

[1081] The server creates an instance of the AI ​​facilitator when a meeting starts. Basic information about each user participating in the meeting (such as name, position, and past speech data) is loaded from the database. At the same time, the emotion recognition engine is initialized, and emotional data is collected in real time by recognizing the user's facial expressions and analyzing their voice tone.

[1082] Automatic meeting minutes creation

[1083] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, and the transcribed content and speaker identification information are sent to a server, which receives the data and automatically updates the minutes.

[1084] Encouraging participants to speak up

[1085] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them notifications encouraging them to speak. This promotes active participation by everyone in the meeting.

[1086] Derailment Detection and Discussion Guide

[1087] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user.

[1088] Visualization of key points

[1089] The device extracts important points and decisions from what is being said during the meeting and displays them in real time. By visually displaying important points, users can easily follow the flow of the discussion, making meetings more efficient.

[1090] Introduction of emotion recognition engineers

[1091] The server uses an emotion recognition engine to analyze the user's emotional state in real time. Emotions are identified from the user's facial expressions, tone of voice, and content of remarks, and the emotional data is fed back to the system. For example, if a user is feeling stressed, this information is notified to other participants, encouraging them to take the discussion further and adjust its content accordingly.

[1092] Specific examples

[1093] Suppose a project progress meeting is being held at a factory. At the start of the meeting, the server initializes the AI ​​facilitator and emotion recognition engine, and loads information about the participants (User A, User B, and User C). As the meeting progresses, if User A says, "This project is going well," the device transcribes this statement in real time and sends it to the server. The server records this statement in the minutes and simultaneously tracks the amount of speech that User A made.

[1094] If it becomes clear that User B has barely spoken, the server sends User B a notification asking, "Can I have your opinion?" If the discussion strays from the project's progress and begins talking about recent news, the server sends a suggestion saying, "The discussion is drifting. Let's get back to the original topic." The device then displays the key point: "Project progress is going well."

[1095] On the other hand, if the emotion recognition engine detects that User A is feeling extremely stressed, the server will send a notification to other participants saying, "User A is feeling stressed. Please be considerate in the discussion." This will allow the meeting to proceed efficiently and productively, and by taking into consideration the emotions of the participants, smooth communication will be achieved.

[1096] Prompt Sentence Examples

[1097] Please implement a function that allows the AI ​​facilitator to track the amount of speech participants make in online meetings and send notifications to encourage participants who are not speaking often. Also, please add a function that uses an emotion recognition engine to analyze the emotional state of participants and notify other participants if they are feeling stressed.

[1098] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1099] Step 1:

[1100] At the start of a meeting, the server instantiates an AI facilitator and loads basic participant information from a database. Input includes participant name, role, and past speech data. Based on this data, the server creates individual participant profiles as output and provides them to the AI ​​facilitator.

[1101] Step 2:

[1102] The server initializes the emotion recognition engine and prepares to collect the user's emotion data in real time. The inputs are facial expression recognition data, voice tone data, and speech content data, and the data is analyzed to output the user's emotional state.

[1103] Step 3:

[1104] The device uses speech recognition technology to transcribe speech in real time during a meeting. The input is the user's speech audio data, and the output is the transcribed text data. This allows the speech content to be recorded as text.

[1105] Step 4:

[1106] The device sends the transcribed speech and the speaker's identification information to the server. The input is the speech transcription data and the speaker ID, and the output is a packet containing these data. The server receives this data and automatically updates the minutes.

[1107] Step 5:

[1108] The server tracks the amount of speech of each participant in real time. The input is speech transcription data and speaker ID, and accumulates the number of speeches and speech duration in a database. The output is speech volume data for each participant.

[1109] Step 6:

[1110] The server sends a notification to participants who speak less to encourage them to speak. The input is the speech volume data of each participant, and the output is a notification message to participants who speak less.

[1111] Step 7:

[1112] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. The input is speech transcription data, and the output is a flag indicating whether the topic is deviating. If a derailment is detected, the server generates instructions to return to the original topic and notifies the user via their device.

[1113] Step 8:

[1114] The device extracts important points and decisions from what is said during the meeting and displays them on the screen. The input is the transcription data of the speech, and the output is a visual representation of the extracted important points. By visually displaying the important points, users can easily follow the flow of the discussion.

[1115] Step 9:

[1116] The server uses an emotion recognition engine to analyze the user's emotional state. The inputs are the user's facial expression data, voice tone data, and speech content data, and the output is emotional state data. For example, if the user is in a stressful state, that information is notified to other participants.

[1117] Step 10:

[1118] At the end of the meeting, the server saves all meeting minutes, emotion data, and speech volume data, and outputs them as a report if necessary. The input is all the data collected during the meeting, and the output is the final meeting report. This leaves a record to review later.

[1119] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1120] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1121] 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 the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1122] [Fourth embodiment]

[1123] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1124] 7, a 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.

[1125] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1126] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1127] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1128] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1129] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1130] The control object 443 includes a display device, LEDs in the eyes, and motors for driving 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 emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1131] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1132] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[1133] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1134] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1135] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1136] Initialization

[1137] When a meeting starts, the server creates an instance of the AI ​​facilitator. Next, it loads basic information about participants (such as their names, positions, and past speech data) from the database. It also initializes a minutes template and prepares the data structures necessary for the meeting to proceed.

[1138] Automatic meeting minutes creation

[1139] The device uses speech recognition technology to transcribe each statement in real time during the meeting. The transcribed content and speaker identification information are sent to the server as the meeting progresses. The server receives the data and automatically records it as meeting minutes.

[1140] Encouraging participants to speak up

[1141] The server tracks each participant's speech volume (number of times and duration) in real time. It identifies participants who speak less and sends them notifications encouraging them to speak. Notifications are automatically generated according to the progress of the conference and are delivered to users.

[1142] Derailment Detection and Discussion Guide

[1143] The server uses natural language processing technology to determine in real time whether a discussion has deviated from the planned topic. If a deviation from the topic is detected, the server generates suggestions to bring the discussion back on topic and notifies the user, thereby reducing the risk of the discussion dragging on unnecessarily or drifting off topic.

[1144] Visualization of key points

[1145] The device extracts important points and decisions from what is being said during the meeting and displays them on a whiteboard in real time. This whiteboard is constantly updated during the meeting, making it easier for users to visually follow the important points. This allows meetings to proceed efficiently without losing track of the flow of discussion or important points.

[1146] Specific examples

[1147] For example, let's say a company is holding a project progress meeting. When the meeting begins, the server launches the AI ​​facilitator and loads the data of the participants, "Tanaka," "Yamada," and "Sato." As the meeting progresses, if Tanaka says, "This project is going well," his remarks are transcribed in real time by the device. The transcription data and Tanaka's identification information are sent to the server and recorded in the minutes.

[1148] Next, tracking reveals that Yamada has barely spoken anything. The server sends Yamada a notification saying, "You haven't spoken much. Do you have any input?" This gives Yamada a chance to express his opinion.

[1149] If, at some point, Sato starts talking about recent news, the server will determine that the remark is unrelated to the project's progress and send the user a suggestion: "The discussion is veering off-topic. Let's return to the main agenda."

[1150] Furthermore, the device automatically displays important points in the meeting, such as "Project progress is going smoothly," on the whiteboard, allowing users to visually confirm important matters during the meeting in real time.

[1151] As described above, this system provides various functions to ensure that online meetings proceed efficiently and productively, contributing to improving the quality of meetings.

[1152] The processing flow will be explained below.

[1153] Step 1: Initialization

[1154] The server creates an instance of the AI ​​facilitator at the start of the meeting, which launches a system to assist in the progress of the online meeting.

[1155] Step 2: Load participant information

[1156] The server loads basic information of participants (such as name, position, and past comment data) from the database, which enables identification of each participant and support based on past data.

[1157] Step 3: Initialize the meeting minutes template

[1158] The server initializes a meeting template, which provides the data structures necessary for the minutes format and flow of the meeting.

[1159] Step 4: Start real-time transcription

[1160] As the meeting progresses, the device uses speech recognition technology to transcribe each utterance in real time, and the speech is automatically converted into text.

[1161] Step 5: Sending speech data

[1162] The device sends the transcribed speech and the speaker's identity to the server, where the speech is ready to be recorded.

[1163] Step 6: Update the minutes

[1164] The server automatically updates the meeting minutes based on the received speech data, ensuring that each speech is recorded appropriately.

[1165] Step 7: Track your speech volume

[1166] The server tracks each participant's speech volume (number of times and duration) in real time, allowing the activity level of participants to be monitored.

[1167] Step 8: Identify quiet participants

[1168] The server identifies participants who are less active: participants whose number of comments falls below a certain threshold.

[1169] Step 9: Send a prompt

[1170] The server generates a notification for quiet participants to speak up and sends it to the user, with a message such as "You haven't spoken much. Do you have any input?"

[1171] Step 10: Derailment detection

[1172] The server uses natural language processing technology to determine in real time whether the current comment is off-topic.

[1173] Step 11: Sending a discussion guide notification

[1174] If the server detects a digression, it generates a suggestion to bring the discussion back on topic and sends it to the user: "The discussion is veering off-topic. Let's return to the main agenda."

[1175] Step 12: Extracting key points

[1176] The device extracts important points and decisions from what is said during the meeting.

[1177] Step 13: Display on the Whiteboard

[1178] The device displays the extracted important points on a whiteboard in real time, allowing users to visually confirm the important points.

[1179] Step 14: End the meeting and export the minutes

[1180] At the end of the meeting, the server exports the accumulated minutes in an editable format, which users can review and modify.

[1181] Through these steps, this system enables online meetings to proceed efficiently and productively.

[1182] Example 1

[1183] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1184] Existing online meeting systems face the challenge of efficiently managing a wide range of tasks, including meeting progress, minutes creation, participant prompting, digression detection and discussion guidance, and visualization of key points. This can lead to poor meeting quality, redundant discussions, and participants being reluctant to speak up. Furthermore, creating meeting minutes manually is time-consuming and labor-intensive, and lacks accuracy.

[1185] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1186] In this invention, the server includes a meeting start detection means, an AI facilitator instance generation means, a participant information database loading means, a minutes template initialization means, a speech transcription means using speech recognition, a speech data transmission means, a speech data recording means, a participant speech volume tracking means, a speech promotion notification generation means, a speech promotion notification transmission means, a discussion digression detection means, a discussion guide notification generation means, a discussion guide notification transmission means, an important point extraction means, and a whiteboard display means. This makes it possible to manage the progress of the meeting, create minutes, encourage participants to speak, guide the discussion, and visualize important points in real time.

[1187] The "conference start detection means" is a means for recognizing the start of a conference and for the system to automatically start the necessary initial processing.

[1188] An "AI facilitator instance generation means" is a means for generating an instance of an AI model to smoothly proceed with a meeting.

[1189] The "participant information database loading means" is a means for loading basic information of conference participants from a database.

[1190] The "minutes template initialization means" is a means for initializing a template for creating minutes of a meeting.

[1191] A "means for transcribing speech using speech recognition" is a means for using speech recognition technology to transcribe speech during a meeting in real time.

[1192] The "utterance data transmission means" is a means for transmitting the transcribed utterance data to the server.

[1193] The "utterance data recording means" is a means for recording received utterance data in a database.

[1194] The "participant speech volume tracking means" is a means for tracking the speech volume of each participant during a conference in real time.

[1195] The "means for generating a speech prompting notification" is a means for generating a notification that prompts a participant who makes few comments to speak.

[1196] The "speech prompting notification sending means" is a means for sending the generated speech prompting notification to the relevant participant.

[1197] The "discussion deviation detection means" is a means for detecting whether the discussion in the meeting has deviated from the planned topic.

[1198] The "discussion guide notification generating means" is a means for generating a suggestion to bring the discussion back to the main topic when the discussion goes off track.

[1199] The "discussion guide notification sending means" is a means for sending the generated discussion guide notification to the participants.

[1200] The "means for extracting important points" is a means for extracting important points and decisions from the content of statements made during a meeting.

[1201] The "whiteboard display means" is a means for displaying the extracted important points on a whiteboard in real time.

[1202] The present invention is a system for conducting online meetings efficiently and productively. This system automates various processes from the start to the end of a meeting and combines multiple means to improve the quality of the meeting.

[1203] Initialization

[1204] When the server detects the start of a meeting, it creates an instance of an AI facilitator. This uses an AI framework such as TensorFlow or PyTorch. Next, it loads participant information from a database (e.g., MySQL or PostgreSQL). This includes participant names, positions, and past speech data. Next, the server initializes a minutes template. This template uses a predefined file in Markdown format or similar.

[1205] Automatic meeting minutes creation

[1206] During a meeting, the device uses speech recognition technology such as the Google Speech-to-Text API or IBM Watson Speech to Text to transcribe each utterance in real time. The transcribed utterances and the speaker's identity are then sent to the server via an HTTP POST request. The server receives this data and automatically records it in a database such as Elasticsearch.

[1207] Encouraging participants to speak up

[1208] The server uses MongoDB or similar to track the amount of speech each participant makes (number of times they make a comment and how long they speak). For participants who make little comments, the server uses the Slack API or Microsoft Teams API to generate and send notifications encouraging them to speak. These notifications are generated automatically and sent to users.

[1209] Derailment Detection and Discussion Guide

[1210] The server uses natural language processing techniques, such as Hugging Face's Transformers library, to determine in real time whether the discussion is deviating from the planned topic. If a deviation is detected, the server compares it with pre-defined agenda information, generates suggestions to bring the discussion back on topic, and communicates them to the user. This notification is sent automatically at the appropriate time.

[1211] Visualization of key points

[1212] The device uses natural language processing libraries such as SpaCy and NLTK to extract key points and decisions from what is being said during the meeting. The extracted key points are displayed in real time on a whiteboard service such as Miro Board or Microsoft Whiteboard. This display is constantly updated as the meeting progresses, making it easy for users to visually follow the key points.

[1213] Specific examples

[1214] For example, when a company holds a project status meeting, the server detects the start of the meeting through the Google Calendar API, creates an instance of an AI facilitator using TensorFlow, and then loads the data of the participants, "Tanaka," "Yamada," and "Sato," from a MySQL database.

[1215] As the meeting progresses, if Tanaka says, "This project is going well," the device will transcribe his speech in real time using the Google Speech-to-Text API. The transcription data and Tanaka's identification information will be sent to the server via an HTTP POST request and recorded in Elasticsearch.

[1216] Once MongoDB data tracks that Yamada has barely spoken, the server uses the Slack API to send a notification to Yamada saying, "You haven't spoken much. Do you have any input?"

[1217] If Sato starts talking about recent news midway through, the server will use Hugging Face's Transformers library to determine that his remarks are unrelated to the project's progress and send the user a suggestion: "The discussion is veering off-topic. Let's return to the main agenda."

[1218] The device automatically displays the key point of the meeting, such as "Project progress is going well," on the Miro board, allowing users to visually check important matters during the meeting in real time through the whiteboard.

[1219] Prompt Sentence Examples

[1220] Below is an example prompt for requesting minutes of a project status meeting:

[1221] A project progress meeting was held. The participants were Tanaka, Yamada, and Sato. Tanaka stated, "This project is progressing smoothly." Based on the minutes, please create and propose the following action items:

[1222] As described above, the present invention provides various functions for conducting online conferences efficiently and productively, thereby contributing to improving the quality of conferences.

[1223] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1224] Processing flow

[1225] Step 1:

[1226] The server detects when a meeting starts. To do this, it uses the Google Calendar API to monitor scheduled events. The input is the calendar information obtained from the API, and the output is a meeting start event.

[1227] Step 2:

[1228] The server creates an instance of the AI ​​facilitator. First, it loads the AI ​​model using the TensorFlow or PyTorch library and creates an instance. The input is the pre-trained AI model, and the output is the generated instance of the AI ​​facilitator.

[1229] Step 3:

[1230] The server loads participant information from a database. It issues SQL queries to a database such as MySQL or PostgreSQL to retrieve participant names, positions, and past comment data. The input is the database connection information and query, and the output is the retrieved participant information.

[1231] Step 4:

[1232] The server initializes the minutes template. It reads a predefined Markdown template file and prepares the initial state. The input is the path to the template file, and the output is the initialized minutes template.

[1233] Step 5:

[1234] The device transcribes speech during meetings in real time. It uses the Google Speech-to-Text API to convert audio data captured from the microphone into text data. The input is audio data, and the output is transcribed text data.

[1235] Step 6:

[1236] The device sends the transcribed speech data and the speaker's identification information to the server using an HTTP POST request. The input is the speech data and the identification information, and the output is the success status of the data transmission to the server.

[1237] Step 7:

[1238] The server records the received utterance data in Elasticsearch, analyzes the data, and adds it to the appropriate index. The input is the received utterance data, and the output is the index information stored in Elasticsearch.

[1239] Step 8:

[1240] The server tracks the amount of speech of each participant in real time. It records the speech data in MongoDB and executes aggregation queries to calculate the number of speeches and speech duration. The input is the received speech data, and the output is the tracked speech volume.

[1241] Step 9:

[1242] The server generates notifications to encourage participants who speak less to speak. The notification content is automatically generated using the Slack API or Microsoft Teams API. The input is the tracked speech volume data, and the output is the generated notification message.

[1243] Step 10:

[1244] The server sends the generated notification to the appropriate user using the Slack API or Microsoft Teams API. The input is the generated notification message, and the output is the success status of the notification sending.

[1245] Step 11:

[1246] The server detects deviations in the discussion. It uses Hugging Face's Transformers library to compare the content of comments with the agenda and determine whether or not a deviation has occurred. The input is the comment data and agenda information, and the output is the deviation detection result.

[1247] Step 12:

[1248] When a digression is detected, the server generates a suggestion to bring the discussion back to the main topic. The suggestion content is generated using a generative AI model. The input is the detected digression result, and the output is the generated suggestion message.

[1249] Step 13:

[1250] The server notifies the user of the proposal message. It sends the notification using the chat application's API. The input is the generated proposal message, and the output is the success status of sending the notification.

[1251] Step 14:

[1252] The device extracts important points from what is said during the meeting. It uses SpaCy and NLTK to analyze the speech and extract important points and decisions. The input is the speech data, and the output is the extracted important points.

[1253] Step 15:

[1254] The device displays the extracted key points on a whiteboard. It updates the display in real time using the APIs of Miro Board and Microsoft Whiteboard. The input is the extracted key points, and the output is the display result on the whiteboard.

[1255] Through the above processing steps, the system manages the progress of the meeting, creates minutes, encourages participants to speak up, directs the discussion, and visualizes important points.

[1256] (Application example 1)

[1257] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1258] Modern meeting systems lack features such as taking minutes, encouraging participants to speak up, and detecting derailed discussions. Some participants often don't speak much, or the discussion often goes off track, wasting time. Furthermore, participants can lose track of the flow of the discussion because they cannot visually track important points and decisions. These issues directly affect production efficiency, particularly in factory progress meetings, so there is a high need for a system that can quickly and efficiently conduct meetings.

[1259] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1260] In this invention, the server includes an automated meeting record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, a means for transcribing speeches made during a meeting in real time using speech recognition technology and recording the results, a means for sending notifications to participants who speak less to encourage them to speak, a means for determining in real time whether the discussion has deviated from the planned topic using natural language processing technology and generating suggestions, and a means for displaying key points and decisions on a whiteboard in real time, thereby enabling improvements in the efficiency and productivity of meetings.

[1261] The "automated meeting record generator" is a function that automatically recognizes statements made during a meeting and creates minutes in real time.

[1262] The "participant speech volume tracking means" is a function for measuring and recording the number of speeches and speech duration of each conference participant in real time.

[1263] The "speech promotion means" is a function for sending notifications to participants who do not speak much, encouraging them to speak.

[1264] "Derailment detection and discussion guidance means" is a function that uses natural language processing technology to determine whether a discussion has deviated from the planned topic and suggests an appropriate direction for the discussion.

[1265] The "important point visualization means" is a function that displays important points and decisions made during a meeting on a display device in real time, allowing participants to visually grasp important information.

[1266] "Speech recognition technology" is a technology for converting speech into text, and in this invention is a technology used to transcribe statements made during a meeting in real time.

[1267] "Real-time transcription means" is a function that uses voice recognition technology to instantly convert meeting speech into text and record it.

[1268] "Means for encouraging participants who speak infrequently to speak by sending notifications" is a function for tracking the volume of comments in real time and sending notifications to participants who speak infrequently to encourage them to speak.

[1269] "Natural language processing technology" is a technology for understanding and analyzing human language, and in this invention, it is a technology used to detect digressions in discussions and provide appropriate discussion guidance.

[1270] "A means of displaying important points and decisions on a whiteboard in real time" is a function that instantly reflects important statements and decisions made during a meeting on a display device such as a whiteboard, allowing all participants to visually confirm them.

[1271] The system embodying this invention is designed to streamline factory progress meetings and is composed of a server, terminals, and users with various functions. Specifically, it is composed of an automated meeting record generation means, a means for tracking the amount of speech by participants, a means for encouraging speech, a means for detecting deviations and guiding discussions, and a means for visualizing important points.

[1272] When a meeting starts, the server creates an instance of the AI ​​facilitator and loads basic information about participants (such as their names, positions, and past speech data) from a database. This database uses a system such as SQLite. It also initializes a minutes template and prepares the data structure necessary for the meeting to proceed.

[1273] The device uses speech recognition technology to transcribe speech in real time during a meeting. The transcribed speech and speaker identification information are sent to a server, which records the data as meeting minutes in real time. The speech recognition technology used here includes the speech_recognition library.

[1274] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them a notification to encourage them to speak. This notification is sent to an application on a smartphone or tablet.

[1275] The digression detection and discussion guidance mechanism uses natural language processing (NLP) technology to determine in real time whether a discussion has deviated from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user. The NLP technology used here includes the spacy library.

[1276] The important points visualization feature extracts important points and decisions from what is said during the meeting and displays them in real time on a whiteboard or display, allowing users to visually confirm important points during the meeting.

[1277] Examples:

[1278] Consider a situation where a meeting is being held in a factory to discuss the progress of a manufacturing process. When the meeting begins, the server launches the AI ​​facilitator and loads the data of the participants, "Participant A," "Participant B," and "Participant C." The device transcribes Participant A's statement, "Production line 1 is progressing as planned," in real time and sends it to the server. The server records this statement in the minutes in real time.

[1279] Next, if tracking reveals that participant B is not speaking much, the server sends a notification to participant B's device saying, "Please speak more." Furthermore, if participant C goes off on a tangent about a topic unrelated to the discussion, the server generates and notifies the participant B with a suggestion saying, "The conversation is going off topic. Let's return to the original topic."

[1280] Important comments are immediately displayed on displays in the factory, allowing all participants to share the latest information and conduct meetings efficiently.

[1281] Example prompts for generative AI models:

[1282] "Please propose an application that utilizes an AI facilitator with the functions of automatic minutes creation, prompting, digression detection, and visualization of key points in status meetings on a manufacturing line. Please also include a specific usage scenario and example program."

[1283] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1284] Step 1:

[1285] Execution of initialization process

[1286] The server creates an instance of the AI ​​facilitator when the meeting starts. It also loads basic information about participants (such as name, position, and past speech data) from a database. This information is stored using SQLite or similar. It also initializes a minutes template and prepares the data structures necessary for the meeting to proceed.

[1287] Input: Conference start signal, database (basic information of participants)

[1288] Output: An instance of the AI ​​facilitator, an initialized data structure, and a meeting minutes template.

[1289] Specific operation: The server uses a program such as Java or Python to access the conference management server and load participant data.

[1290] Step 2:

[1291] Real-time audio transcription

[1292] During the meeting, the device uses speech recognition technology (e.g., the speech_recognition library) to transcribe each utterance in real time, and then sends the transcribed utterances to the server.

[1293] Input: Audio data (speech during a meeting)

[1294] Output: Transcription data (text format), speaker identification information

[1295] Specific operation: The device receives voice input from the microphone, converts it into text data, and sends it to the server.

[1296] Step 3:

[1297] Real-time meeting minutes updates

[1298] The server updates the minutes in real time based on the received transcription data, thereby recording every word spoken during the meeting.

[1299] Input: Transcription data, speaker identification information

[1300] Output: Updated minutes

[1301] Specific operation: The server updates and stores the minutes data using a list or tree structure.

[1302] Step 4:

[1303] Participant speech volume tracking

[1304] The server tracks each participant's speech volume (number of times and duration) in real time and identifies participants who are not speaking much, allowing the server to provide appropriate feedback.

[1305] Input: Speech data, speaker identification information

[1306] Output: Tracking data (speech volume record)

[1307] Specific operation: The server measures the volume of messages using timestamp analysis and statistical functions.

[1308] Step 5:

[1309] Sending a prompt to speak

[1310] For participants who do not speak much, the server generates a notification encouraging them to speak and sends it to their smartphone or tablet.

[1311] Input: Tracking data, target participant identification information

[1312] Output: Speech prompt notification

[1313] Specific operation: The server uses a notification service (e.g., Firebase Cloud Messaging) to send a notification to the target participant's device.

[1314] Step 6:

[1315] Derailment detection and guide suggestions

[1316] The server uses natural language processing techniques (such as the Spacy library) to determine in real time whether the discussion is deviating from the planned topic, and if a digression is detected, it generates suggestions to bring the discussion back on topic and notifies the user.

[1317] Input: Transcription data, topic information

[1318] Output: Derailment notice, discussion guide suggestion

[1319] Specific operation: The server uses natural language processing technology to perform text analysis and analyze the direction of the discussion.

[1320] Step 7:

[1321] Visualizing key points

[1322] The device automatically extracts important points and decisions during the meeting and displays them on a whiteboard in real time, making it easier for participants to visually follow important matters during the meeting.

[1323] Input: Speech data, important points data

[1324] Output: Display (whiteboard)

[1325] Specific operation: The device uses the display output function to display important points on the whiteboard.

[1326] Step 8:

[1327] Data processing after the meeting

[1328] After the meeting ends, the server performs a final check and saves the minutes data, preparing it for distribution to participants.

[1329] Input: Final minutes data

[1330] Output: minutes for distribution

[1331] Specific operation: The server uses a database system to store the minutes data and export it in an appropriate format.

[1332] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1333] The present invention relates to a system and method that provides multiple functions for efficiently and productively conducting online meetings, in particular, a combination of an automated meeting record generator, a participant speech volume tracking unit, a speech promotion unit, a digression detection and discussion guide unit, a key point visualization unit, and an emotion engine that recognizes user emotions.

[1334] Initialization

[1335] The server creates an instance of the AI ​​facilitator at the start of a meeting. At this stage, it loads basic information about each user participating in the meeting (such as name, position, and past speech data) from a database. At the same time, the emotion engine is initialized, and emotional data is ready to be collected in real time through facial expression recognition and voice tone analysis.

[1336] Automatic meeting minutes creation

[1337] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, and the transcribed content and speaker identification information are sent to a server, which receives the data and automatically updates the minutes.

[1338] Encouraging participants to speak up

[1339] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them notifications encouraging them to speak. This promotes active participation by everyone in the meeting.

[1340] Derailment Detection and Discussion Guide

[1341] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user.

[1342] Visualization of key points

[1343] The device extracts important points and decisions from what is being said during the meeting and displays them on the whiteboard in real time. By visually displaying important points, users can easily follow the flow of the discussion, making meetings more efficient.

[1344] Introducing the Emotion Engine

[1345] The server uses an emotion engine to analyze the user's emotional state in real time. Emotions are identified from the user's facial expressions, tone of voice, and content of comments, and the emotional data is fed back to the system. For example, if a user is feeling stressed, this information is notified to other participants, encouraging them to take the discussion further and adjust its content accordingly.

[1346] Specific examples

[1347] Suppose a company is holding a project progress meeting. At the start of the meeting, the server initializes the AI ​​facilitator and emotion engine and loads the information of the participants (Tanaka, Yamada, Sato). As the meeting progresses, if Tanaka says, "This project is going well," this statement is transcribed in real time and sent to the server. The server records this statement in the minutes and simultaneously tracks the amount of speech Tanaka made.

[1348] When it becomes clear that Yamada has barely spoken, the server sends Yamada a notification saying, "You haven't spoken much. Do you have any input?" If the discussion veers away from project progress and starts talking about recent news, the server sends a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda."

[1349] The device displays the important point, "The project is progressing smoothly," on the whiteboard. On the other hand, if the emotion engine detects that Tanaka is feeling very stressed, the server sends a notification to other participants saying, "Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[1350] In this way, the present invention significantly improves the efficiency and productivity of online meetings. The introduction of an emotion engine allows smooth communication while taking into consideration the emotional state of users during a meeting.

[1351] The processing flow will be explained below.

[1352] Step 1: Initialize the system

[1353] The server instantiates the AI ​​facilitator and emotion engine at the start of a meeting. This initialization ensures that the entire system is ready to assist in running the meeting.

[1354] Step 2: Load participant information

[1355] The server loads basic information of participants (such as name, title, and past comment data) from the database. This information is used for subsequent comment tracking and sentiment analysis.

[1356] Step 3: Start speech recognition

[1357] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, instantly converting what is said during the meeting into text data.

[1358] Step 4: Sending speech data

[1359] The device sends the transcribed speech and the speaker's identity to the server, where it is ready to be recorded.

[1360] Step 5: Update the minutes

[1361] The server automatically updates the meeting minutes based on the received speech data, allowing the content of speech to be recorded in real time.

[1362] Step 6: Track your speech volume

[1363] The server tracks each participant's speech volume (number of times and duration) in real time, and the tracking data is used to encourage subsequent speech.

[1364] Step 7: Identify quiet participants

[1365] The server identifies participants who are quiet based on the tracking data. Participants who are below a certain threshold are identified.

[1366] Step 8: Send a prompt

[1367] The server generates a notification for participants who have not spoken much and sends it to the user to encourage them to speak up. Example of notification: "You haven't spoken much. Do you have any input?"

[1368] Step 9: Derailment detection with natural language processing

[1369] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the topic, and if there are signs of derailment, subsequent processing will be based on that information.

[1370] Step 10: Sending a discussion guide notification

[1371] If the server detects a digression, it generates a suggestion to bring the discussion back on topic and sends it to the user. Example notification: "The discussion is veering off-topic. Let's return to the main agenda."

[1372] Step 11: Extract and display key points

[1373] The device extracts important points and decisions from what is being said during the meeting and displays them on the whiteboard in real time in a format that is easy to follow visually.

[1374] Step 12: Collect and analyze emotion data

[1375] The server uses an emotion engine to collect emotional data from participants' facial expressions, tone of voice, and content of remarks, etc. Based on the analyzed emotional data, feedback is provided to the progress of the meeting in real time.

[1376] Step 13: Communicating Emotional State

[1377] When the server detects an important emotional state, such as a specific user being stressed, it notifies other participants of that information. Example of notification: "Mr. Tanaka seems to be feeling stressed. Please be considerate in the progress of the discussion."

[1378] Step 14: End the meeting and export data

[1379] At the end of the meeting, the server exports the accumulated minutes and emotion data in an editable format, which users can download and review as needed.

[1380] Through the above steps, this system incorporating emotion analysis not only enables online meetings to proceed efficiently and productively, but also ensures smooth communication that takes into consideration the emotions of participants.

[1381] Example 2

[1382] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1383] In online meetings, it is extremely difficult to automatically create minutes, track the amount of speech, prevent discussions from going off track, and visualize important points in real time. Another challenge is accurately understanding participants' emotional states and reflecting them in the progress of the discussion. By solving these challenges, it is necessary to realize efficient and productive online meetings and reduce user stress.

[1384] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1385] In this invention, the server includes an automated conference record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, and a sentiment analysis means, which make it possible to transcribe speeches made during a conference in real time, track the speech volume of each participant, prevent digressions and bring the discussion back to the main topic, visualize key points, and analyze sentiment data to reflect the progress of the conference.

[1386] "Automated meeting record generation means" is a technology that automatically records what is said in online meetings and generates minutes of the meeting.

[1387] The "participant speech volume tracking means" is a technology for measuring and recording the number of speeches and speech duration of each participant during a conference in real time.

[1388] The "speech promotion means" is a function for sending a notification to participants who do not speak much to encourage them to speak.

[1389] The "digression detection and discussion guidance means" is a technology for determining whether a discussion during a meeting has deviated from the planned topic and generating appropriate suggestions to bring the discussion back on topic.

[1390] "Key point visualization means" is a technology for extracting important points and decisions from what is said during a meeting and displaying them visually.

[1391] The "emotion analysis means" is a technology for analyzing the user's emotional state in real time using facial expression recognition technology and voice tone analysis technology.

[1392] "Speech recognition technology" is a technology for converting voice data into text data.

[1393] "Natural language processing technology" is a technology for analyzing text data, understanding its meaning, and processing it.

[1394] "Whiteboard software" is a digital tool for visually sharing information during meetings.

[1395] "Real time" is a time concept that refers to processing and display occurring almost simultaneously.

[1396] This invention relates to a system and method for efficiently and productively conducting online meetings, which combines an automated meeting record generation unit, a participant speech volume tracking unit, a speech promotion unit, a digression detection and discussion guide unit, a key point visualization unit, and a sentiment analysis unit. The operation of each unit will be described in detail below.

[1397] Initialization

[1398] The server instantiates an AI facilitator at the start of a meeting. This instantiation is done using a Python program that loads each participant's name, title, and past speech data from a database (e.g., MySQL). It also initializes an emotion engine using the Microsoft Azure Face API and Amazon Rekognition to recognize the user's facial expressions and analyze their speech tone.

[1399] Automatic meeting minutes creation

[1400] The device uses speech recognition technology to transcribe what is said during the meeting in real time using the Google Cloud Speech-to-Text API and IBM Watson Speech to Text. The transcribed content and speaker identification information are sent to a server, which then automatically updates the meeting minutes based on this data.

[1401] Encouraging participants to speak up

[1402] The server tracks each participant's speech volume (number of times and duration) in real time. This allows it to identify participants who are not speaking much and send them a notification to encourage them to speak up. For example, it uses the Slack API to send a message like, "You haven't spoken much. Do you have any input?"

[1403] Derailment Detection and Discussion Guide

[1404] The server uses natural language processing (NLP) technology to analyze the content of the meeting in real time. Specifically, it uses OpenAI GPT-3 and Google BERT to determine whether the discussion is deviating from the planned topic. If a deviation is detected, it generates a suggestion to the user, saying, "The discussion is veering off-topic. Let's return to the main agenda."

[1405] Visualization of key points

[1406] The device uses NLP technology to extract key points and decisions from what is being said, and the extracted content is displayed in real time on whiteboard software such as Miro or Microsoft Whiteboard.

[1407] Introducing sentiment analysis

[1408] The server uses emotion analysis technology to analyze the user's emotional state in real time. It uses Amazon Rekognition and Microsoft Azure Face API to analyze the user's facial expressions and voice tone, and identifies their emotional state based on that data. For example, if a user is feeling stressed, the server notifies other participants, saying, "Mr. Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[1409] Specific examples

[1410] Suppose a company is holding a project progress meeting. At the start of the meeting, the server initializes the AI ​​facilitator and sentiment analysis engine, and loads the information of the participants (Tanaka, Yamada, Sato) from the database. As the meeting progresses, if Tanaka says, "This project is going well," his remarks are transcribed in real time and sent to the server. The server updates the minutes based on this data, while simultaneously tracking the amount of speech Tanaka made.

[1411] If John has barely spoken, the server sends him a notification saying, "You haven't spoken much. Do you have any input?" If the discussion veers away from project progress and starts talking about recent news, the server sends him a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda."

[1412] The device displays the key point on the whiteboard: "The project is progressing smoothly." However, if the emotion analysis engine detects that Tanaka is feeling extremely stressed, it sends a notification to other participants saying, "Tanaka seems to be feeling stressed. Please be considerate in the discussion."

[1413] Prompt Sentence Examples

[1414] "Describe a system that generates real-time online meeting minutes. Detail the specific hardware and software used and how the data is processed. Also, mention the role of sentiment analysis."

[1415] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1416] Step 1:

[1417] The server creates an instance of the AI ​​facilitator at the start of a meeting. As input, it retrieves the list of meeting participants and past speech data from a database (e.g., MySQL). The output is the initialization of the AI ​​facilitator and the initialization of the sentiment analysis engine. Specifically, it runs a Python program to load the names, titles, and past speech data of participants. It also initializes the sentiment analysis engine using the Microsoft Azure Face API and Amazon Rekognition.

[1418] Step 2:

[1419] The device uses speech recognition technology to transcribe what is said during a meeting in real time. The input is audio data collected during the meeting via a microphone. The output is transcribed text data, which is sent to the server via WebSocket or HTTP request. Specifically, it uses the Google Cloud Speech-to-Text API or IBM Watson Speech to Text to generate text data from the input audio data.

[1420] Step 3:

[1421] The server receives the submitted transcription data and automatically updates the minutes with speaker information. It receives the transcription data and speaker identification information as input. The output is the updated minutes data, which is shared with all participants in real time. Specifically, it runs a Python program that adds new utterances to the database and updates the current minutes.

[1422] Step 4:

[1423] The server tracks each participant's speech volume (number of times spoken and duration) in real time. It uses transcription data and time information as input. The output is speech volume data for each participant, which is displayed on a dashboard. Specifically, it runs a tracking algorithm, calculates the speech volume for each participant, and stores it in a database.

[1424] Step 5:

[1425] When a participant who has not spoken much is identified, the server sends a notification to that participant's device encouraging them to speak. Speech volume data is used as input. The output is a notification message encouraging them to speak. Specifically, the server uses the Slack API to send the message "You haven't spoken much. Do you have any input?"

[1426] Step 6:

[1427] The server uses natural language processing (NLP) technology to analyze the content of the meeting in real time. The input is the transcription data. The output is a judgment result on whether the discussion has gone off track. Specifically, it uses OpenAI GPT-3 and Google BERT to analyze the input data and determine whether there has been any deviation.

[1428] Step 7:

[1429] When it detects that the discussion is veering off-topic, the server generates a suggestion saying, "The discussion is veering off-topic. Let's return to the main agenda." and notifies the user. The deviation detection result is used as input. The output is a suggestion message to return the discussion. The specific operation is to generate a suggestion message and send it to the user via Slack or email.

[1430] Step 8:

[1431] The device extracts key points and decisions from what is being said. It uses transcription data as input. The output is key points that are displayed in the whiteboard software. Specifically, it uses NLP technology to extract key information and display it on Miro or Microsoft Whiteboard.

[1432] Step 9:

[1433] The server uses emotion analysis technology to analyze the user's emotional state in real time. Facial expression data and voice tone data are collected as input. The output is analyzed emotional data. Specifically, emotions are analyzed using Amazon Rekognition and Microsoft Azure Face API, and the results are reflected in the system.

[1434] Step 10:

[1435] Based on the analyzed emotional data, the server sends appropriate notifications to other participants according to the user's emotions. Emotional data is used as input. The output is a notification message based on the emotional state. A specific operation would be to generate a notification such as "Tanaka seems to be feeling stressed. Please be considerate in the progress of the discussion" and send it to other participants.

[1436] (Application example 2)

[1437] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1438] In modern factories, many meetings are held for production management and business improvement, but it is common for not all participants to speak up actively, and discussions to go off track. It is also difficult to grasp the emotional state of participants, making stress management and improving the quality of communication challenging. Therefore, a system that can ensure efficient and productive meeting progress while also taking into consideration the emotions of participants is needed.

[1439] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes an automated conference record generation means, a participant speech volume tracking means, a speech promotion means, a digression detection and discussion guide means, a key point visualization means, and an emotion recognition engine. This allows for real-time transcription of comments made during a conference, promotes active participation by all participants, prevents the discussion from going off track, and visually displays key points, enabling the conference to proceed efficiently. Furthermore, by understanding participants' emotions in real time and taking into consideration any stress or dissatisfaction, it is possible to create a healthier communication environment.

[1440] An "automated meeting record generator" is a system that automatically records what is said in a meeting, transcribes it, and generates minutes in real time.

[1441] The "participant speech volume tracking means" is a system that records and manages the number of speeches and speech duration of each participant during a conference in real time.

[1442] The "speech promotion means" is a system that automatically sends a notification to participants who make few comments to encourage them to speak.

[1443] The "digression detection and discussion guidance means" is a system that detects when a meeting discussion deviates from the planned topic and provides instructions or suggestions to bring the discussion back to the original topic.

[1444] The "important point visualization means" is a system that extracts important statements and decisions made during a meeting and displays them visually.

[1445] The "emotion recognition engine" is a system that analyzes participants' emotions in real time based on their speech, facial expressions, tone of voice, etc., and provides feedback on the results.

[1446] The present invention provides a system for supporting efficient and productive meetings in factories, which is composed of a server, terminals, and users for data processing and interaction.

[1447] Initialization

[1448] The server creates an instance of the AI ​​facilitator when a meeting starts. Basic information about each user participating in the meeting (such as name, position, and past speech data) is loaded from the database. At the same time, the emotion recognition engine is initialized, and emotional data is collected in real time by recognizing the user's facial expressions and analyzing their voice tone.

[1449] Automatic meeting minutes creation

[1450] As the meeting progresses, the device uses voice recognition technology to transcribe each statement in real time, and the transcribed content and speaker identification information are sent to a server, which receives the data and automatically updates the minutes.

[1451] Encouraging participants to speak up

[1452] The server tracks each participant's speech volume (number of times and duration) in real time, identifies participants who are not speaking much, and sends them notifications encouraging them to speak. This promotes active participation by everyone in the meeting.

[1453] Derailment Detection and Discussion Guide

[1454] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. If a digression is detected, the server generates suggestions to bring the discussion back on topic and notifies the user.

[1455] Visualization of key points

[1456] The device extracts important points and decisions from what is being said during the meeting and displays them in real time. By visually displaying important points, users can easily follow the flow of the discussion, making meetings more efficient.

[1457] Introduction of emotion recognition engineers

[1458] The server uses an emotion recognition engine to analyze the user's emotional state in real time. Emotions are identified from the user's facial expressions, tone of voice, and content of remarks, and the emotional data is fed back to the system. For example, if a user is feeling stressed, this information is notified to other participants, encouraging them to take the discussion further and adjust its content accordingly.

[1459] Specific examples

[1460] Suppose a project progress meeting is being held at a factory. At the start of the meeting, the server initializes the AI ​​facilitator and emotion recognition engine, and loads information about the participants (User A, User B, and User C). As the meeting progresses, if User A says, "This project is going well," the device transcribes this statement in real time and sends it to the server. The server records this statement in the minutes and simultaneously tracks the amount of speech that User A made.

[1461] If it becomes clear that User B has barely spoken, the server sends User B a notification asking, "Can I have your opinion?" If the discussion strays from the project's progress and begins talking about recent news, the server sends a suggestion saying, "The discussion is drifting. Let's get back to the original topic." The device then displays the key point: "Project progress is going well."

[1462] On the other hand, if the emotion recognition engine detects that User A is feeling extremely stressed, the server will send a notification to other participants saying, "User A is feeling stressed. Please be considerate in the discussion." This will allow the meeting to proceed efficiently and productively, and by taking into consideration the emotions of the participants, smooth communication will be achieved.

[1463] Prompt Sentence Examples

[1464] Please implement a function that allows the AI ​​facilitator to track the amount of speech participants make in online meetings and send notifications to encourage participants who are not speaking often. Also, please add a function that uses an emotion recognition engine to analyze the emotional state of participants and notify other participants if they are feeling stressed.

[1465] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1466] Step 1:

[1467] At the start of a meeting, the server instantiates an AI facilitator and loads basic participant information from a database. Input includes participant name, role, and past speech data. Based on this data, the server creates individual participant profiles as output and provides them to the AI ​​facilitator.

[1468] Step 2:

[1469] The server initializes the emotion recognition engine and prepares to collect the user's emotion data in real time. The inputs are facial expression recognition data, voice tone data, and speech content data, and the data is analyzed to output the user's emotional state.

[1470] Step 3:

[1471] The device uses speech recognition technology to transcribe speech in real time during a meeting. The input is the user's speech audio data, and the output is the transcribed text data. This allows the speech content to be recorded as text.

[1472] Step 4:

[1473] The device sends the transcribed speech and the speaker's identification information to the server. The input is the speech transcription data and the speaker ID, and the output is a packet containing these data. The server receives this data and automatically updates the minutes.

[1474] Step 5:

[1475] The server tracks the amount of speech of each participant in real time. The input is speech transcription data and speaker ID, and accumulates the number of speeches and speech duration in a database. The output is speech volume data for each participant.

[1476] Step 6:

[1477] The server sends a notification to participants who speak less to encourage them to speak. The input is the speech volume data of each participant, and the output is a notification message to participants who speak less.

[1478] Step 7:

[1479] The server uses natural language processing technology to determine in real time whether the discussion is deviating from the planned topic. The input is speech transcription data, and the output is a flag indicating whether the topic is deviating. If a derailment is detected, the server generates instructions to return to the original topic and notifies the user via their device.

[1480] Step 8:

[1481] The device extracts important points and decisions from what is said during the meeting and displays them on the screen. The input is the transcription data of the speech, and the output is a visual representation of the extracted important points. By visually displaying the important points, users can easily follow the flow of the discussion.

[1482] Step 9:

[1483] The server uses an emotion recognition engine to analyze the user's emotional state. The inputs are the user's facial expression data, voice tone data, and speech content data, and the output is emotional state data. For example, if the user is in a stressful state, that information is notified to other participants.

[1484] Step 10:

[1485] At the end of the meeting, the server saves all meeting minutes, emotion data, and speech volume data, and outputs them as a report if necessary. The input is all the data collected during the meeting, and the output is the final meeting report. This leaves a record to review later.

[1486] 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 control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1487] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1488] 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 the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1489] The emotion identification model 59 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 an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1490] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1491] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1492] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1493] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1494] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1495] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1496] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1497] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1498] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1499] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1500] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1501] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1502] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1503] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1504] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1505] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1506] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1507] The following is further disclosed regarding the above embodiment.

[1508] (Claim 1)

[1509] an automated meeting record generation means;

[1510] A means for tracking participant speech volume;

[1511] means of encouraging speech;

[1512] digression detection and discussion guide means;

[1513] A means of visualizing key points;

[1514] A system including:

[1515] (Claim 2)

[1516] The system according to claim 1, characterized in that it uses speech recognition technology to transcribe statements made during a meeting in real time and records the results.

[1517] (Claim 3)

[1518] The system according to claim 1, characterized in that the number of times each participant speaks is tracked in real time, and a notification encouraging participants who speak less to speak is sent.

[1519] (Claim 4)

[1520] 10. The system of claim 1, further comprising the ability to use natural language processing to determine if a discussion topic has gone off track and to suggest returning to the main topic.

[1521] (Claim 5)

[1522] 2. The system according to claim 1, wherein the main points of the meeting are displayed on a whiteboard in real time, allowing participants to visually follow the flow of the discussion.

[1523] "Example 1"

[1524] (Claim 1)

[1525] A conference start detection means;

[1526] an AI facilitator instance generation means;

[1527] Participant information database loading means;

[1528] A minutes template initialization means;

[1529] A speech transcription means using speech recognition;

[1530] utterance data transmitting means;

[1531] utterance data recording means;

[1532] A means for tracking participant speech volume;

[1533] speech prompt notification generating means;

[1534] speech prompting notification sending means;

[1535] A discussion digression detection means;

[1536] a discussion guide notification generating means;

[1537] a means for sending a discussion guide notification;

[1538] An important point extraction means;

[1539] a whiteboard display means;

[1540] A system including:

[1541] (Claim 2)

[1542] The system according to claim 1, characterized in that it uses speech recognition technology to transcribe statements made during a meeting in real time and records the results.

[1543] (Claim 3)

[1544] The system according to claim 1, characterized in that the number of times each participant speaks and the duration of their speech are tracked in real time, and a notification is sent to participants who speak less to encourage them to speak.

[1545] "Application Example 1"

[1546] (Claim 1)

[1547] an automated meeting record generation means;

[1548] A means for tracking participant speech volume;

[1549] means of encouraging speech;

[1550] digression detection and discussion guide means;

[1551] A means of visualizing key points;

[1552] A means of transcribing statements made during meetings in real time using speech recognition technology and recording the results;

[1553] A means for sending notifications to participants who do not speak much to encourage them to speak;

[1554] a means for determining in real time whether a discussion has deviated from the planned topic using natural language processing technology and generating suggestions;

[1555] A way to display key points and decisions on a whiteboard in real time,

[1556] A system including:

[1557] (Claim 2)

[1558] The system according to claim 1, characterized in that the system is an application installed on a factory operation panel and has the function of loading basic information of participants and initializing a minutes template.

[1559] (Claim 3)

[1560] The system described in claim 1 is characterized in that, in manufacturing process progress meetings, the system transcribes each participant's speech data in real time, automatically creates minutes, sends notifications to participants who speak less to encourage them to speak up, and displays important points and decisions in real time on a factory display.

[1561] "Example 2: Combining Emotion Engines"

[1562] (Claim 1)

[1563] an automated meeting record generation means;

[1564] A means for tracking participant speech volume;

[1565] means of encouraging speech;

[1566] digression detection and discussion guide means;

[1567] A means of visualizing key points;

[1568] A sentiment analysis means;

[1569] A system including:

[1570] (Claim 2)

[1571] The system according to claim 1, characterized in that it uses speech recognition technology to transcribe statements made during a meeting in real time and records the results.

[1572] (Claim 3)

[1573] The system according to claim 1, characterized in that the number of times each participant speaks and the duration of their speech are tracked in real time, and a notification is sent to participants who speak less to encourage them to speak.

[1574] (Claim 4)

[1575] The system of claim 1, characterized in that it uses natural language processing technology to analyze participants' comments and generate suggestions to bring the discussion back on topic if the discussion d...

Claims

1. an automated meeting record generation means; A means for tracking participant speech volume; means of encouraging speech; digression detection and discussion guide means; A means of visualizing key points; A system including:

2. 2. The system according to claim 1, wherein speech recognition technology is used to transcribe statements made during a meeting in real time and the results are recorded.

3. 2. The system according to claim 1, wherein the number of times each participant speaks is tracked in real time, and a notification is sent to encourage a participant who has not spoken much to speak.

4. 10. The system of claim 1, further comprising the capability of using natural language processing to determine if a discussion topic has gone off track and suggest returning to the main topic.

5. 2. The system according to claim 1, wherein the main points of the meeting are displayed on a whiteboard in real time, allowing participants to visually follow the flow of the discussion.

Images