System

Abstract

A system is provided.SOLUTION: A system comprising: means for transcribing speech during a meeting in real time; means for storing past meeting minutes; means for comparing transcription data of current meeting speech with past minutes data; means for extracting variations and inconsistencies in speech content; means for visualizing extracted differences as graphs and charts; means for notifying users; means for collecting feedback from users; and means for storing collected feedback data.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 business, it's not uncommon for superiors and managers to make different statements from meeting to meeting, which can cause confusion in project progress and decision-making. In such situations, it can be difficult to verify the consistency of superiors' statements and hold them accountable without risking anger. Traditional methods require manually recording meeting minutes and comparing their contents, which is time-consuming, labor-intensive, and lacks accuracy. Therefore, there is a need for a simple and reliable way to monitor the consistency of statements. [Means for solving the problem]

[0005] The present invention provides a system that transcribes statements made during meetings in real time, saves past meeting minutes, compares the transcript data of current meeting statements with the past minutes data, and identifies variations or inconsistencies in the content of statements. This system includes a means for recording audio, a means for comparing the content of statements using a natural language processing algorithm, a means for visualizing the extracted differences as graphs or charts, a means for notifying users, a means for collecting feedback from users, and a means for saving the collected feedback data. This allows meeting participants to easily check the consistency of the statements made by their superiors or managers, and, if necessary, to question the reasons for variations in the statements and pursue responsibility.

[0006] A "meeting" is a place in a business or organization where multiple people gather to share information, discuss, and make decisions.

[0007] "Speech" refers to opinions or information expressed orally by participants in a meeting.

[0008] "Real-time" refers to processing and operations being carried out at the same time as real time, without delay.

[0009] "Transcription" is the process of converting audio data into text data.

[0010] "Minutes" are documents that record the contents of a meeting, including what was said and what decisions were made.

[0011] "Data" is a collection of information or signals, here in the form of text, audio, etc.

[0012] "Comparison" refers to contrasting two or more objects and finding differences and similarities.

[0013] A "natural language processing algorithm" is a technology or method that allows a computer to process human language.

[0014] "Notification" is the act of making specific information known to another person or system.

[0015] "User" refers to a person who uses the system or an end user.

[0016] "Feedback" refers to evaluations and comments on system usage and results.

[0017] A "server" is a device that provides functions such as data processing and storage to other computers over a network.

[0018] "Terminal" refers to an apparatus or device that is directly operated by a user, and in this case refers to equipment that records and operates during a conference.

[0019] "Preservation" is the act of storing information or data for later use.

[0020] "Visualization" is the process of presenting data and information in a visually easy-to-understand format, such as a graph or chart.

[0021] "Difference" refers to the difference that exists between two data sets. [Brief explanation of the drawings]

[0022] [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

[0023] 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.

[0024] First, the terms used in the following description will be explained.

[0025] 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).

[0026] 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.

[0027] 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.

[0028] 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.

[0029] 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."

[0030] [First embodiment]

[0031] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0032] 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.

[0033] 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).

[0034] 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.

[0035] 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.

[0036] 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.

[0037] 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.

[0038] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0039] 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.

[0040] 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.

[0041] 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.

[0042] 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."

[0043] This invention relates to a system that transcribes statements made during meetings in real time and compares them with past minutes. This system visualizes inconsistencies in the content of statements made during meetings and monitors the consistency of statements.

[0044] First, when a meeting begins, the user starts recording audio on their device. The device sends what is said in the meeting to a speech recognition engine in real time, which converts it into text. This text data is then sent directly to the server in real time and temporarily saved as provisional minutes.

[0045] The server then accesses a database of past meeting minutes and compares the current meeting's content with past meeting minutes using natural language processing algorithms that detect variations or inconsistencies in the content.

[0046] The server converts the detected differences into visual formats such as graphs and charts, allowing users to see the changes in the content at a glance. The server then sends this visualization data to users via email or a notification system.

[0047] Users can review the data they receive and, if necessary, ask their superiors or managers why the content of their comments was changed. The users enter their feedback into the system, which the server stores. This feedback data is used to analyze and improve future meetings.

[0048] Specific examples

[0049] For example, suppose that during Meeting A, a boss says, "We'll start a new project next week." Then during the next Meeting B, the same boss says, "We'll start a new project the week after next." This system works as follows:

[0050] 1. Recording and Transcription:

[0051] The device records what is said in Conference B and converts it into text in real time.

[0052] The generated text data is sent to the server.

[0053] 2. Comparing minutes:

[0054] The server retrieves the minutes data for Conference A from the database.

[0055] Using a natural language processing algorithm, the content of statements made in Meeting A and Meeting B is compared.

[0056] 3. Extract and visualize the differences:

[0057] The server detects the change in the content of the message and discovers the change from "next week" to "the week after next."

[0058] The differences are visualized as a graph and sent to the user.

[0059] 4. Getting feedback:

[0060] The user checks the notification and asks their boss why the comment has changed.

[0061] The obtained feedback is entered into the system and the server stores it in a database.

[0062] In this way, this system records what is being said in real time and compares it with past comments to clarify inconsistencies in what is being said, improving the transparency and efficiency of meetings.

[0063] The processing flow will be explained below.

[0064] Step 1:

[0065] Start audio recording

[0066] The terminal activates the audio recording function at the start of the conference.

[0067] The device records what is said in the meeting in real time via a microphone.

[0068] Step 2:

[0069] Real-time audio transcription

[0070] The device sends the recorded voice data to the voice recognition engine in real time.

[0071] The device receives the text data returned from the voice recognition engine and converts the spoken content into text.

[0072] Step 3:

[0073] Sending text data and generating provisional minutes

[0074] The device transmits the text of the speech in real time to the server.

[0075] The server temporarily stores the received text data as provisional minutes.

[0076] Step 4:

[0077] Obtaining past meeting minutes data

[0078] The server retrieves past meeting minutes data from the database.

[0079] The server prepares the acquired minutes data for comparison with today's minutes data.

[0080] Step 5:

[0081] Comparison of meeting minutes data

[0082] The server uses natural language processing (NLP) algorithms to compare today's minutes with past minutes.

[0083] The server detects changes in keywords and phrases in the content of comments.

[0084] Step 6:

[0085] Extracting Differences

[0086] The server identifies fluctuations and inconsistencies in what is being said.

[0087] The server builds a list of the identified differences and aggregates the data.

[0088] Step 7:

[0089] Visualizing differential data

[0090] The server generates visual representations such as graphs and charts based on the extracted differences.

[0091] The server formats the visualization data into a format that is easy for the user to understand.

[0092] Step 8:

[0093] User Notification

[0094] The server adds explanatory text to the visualized data and sends it to the specified user as an email or internal notification.

[0095] The user checks the received notification and understands changes in the content of the comments.

[0096] Step 9:

[0097] Get feedback

[0098] The user checks the notification and, if necessary, asks their superior or manager about the reason for the change in the comment content.

[0099] The user enters this feedback into the system.

[0100] Step 10:

[0101] Feedback Data Storage

[0102] The server receives the feedback data sent by the user and stores it in a database.

[0103] The server will prepare to use the data for future analysis and improvements.

[0104] These steps allow the system to efficiently record what is said during a meeting, compare it with past comments to identify inconsistencies and changes, and report them back to meeting participants, allowing users to understand the reasons for fluctuations and take action to improve the quality of the meeting.

[0105] Example 1

[0106] 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."

[0107] Previously, it was difficult to record statements made during meetings in real time and understand the consistency and changes in the content of those statements. As a result, important decision-making and information transmission in meetings became uncertain, making efficient business operations difficult. Furthermore, comparing past minutes with current statements was time-consuming and labor-intensive. This made ensuring the transparency and reliability of meeting content a challenge.

[0108] 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.

[0109] In this invention, the server includes: means for transcribing statements made during a meeting in real time; means for saving past meeting minutes; means for comparing the transcription data of current meeting statements with past meeting minutes data; means for extracting variations and inconsistencies in the content of statements; means for visualizing the extracted differences as graphs or charts; means for notifying users; means for collecting feedback from users; means for saving the collected feedback data; means for transmitting voice data to a voice recognition engine in real time and acquiring it as text data; and means for comparing multiple meeting statements and analyzing the consistency of the statement content with high accuracy using a natural language processing algorithm. This automates the real-time recording and comparison of meeting content and makes it possible to visually confirm the consistency and variations in the content of statements, enabling efficient and transparent meeting management.

[0110] "Means for transcribing statements made during a meeting in real time" refers to a function or device that instantly converts speech made during a meeting into text data using voice recognition technology.

[0111] "Means for storing minutes of past meetings" refers to a function or device that stores the contents of past meetings as text data in a database or the like.

[0112] "Means for comparing the current meeting transcript with past meeting minutes" refers to algorithms or software that compare the transcript generated from the most recent meeting with similarly stored meeting minutes from the past.

[0113] "Means for detecting variations or inconsistencies in statements" refers to algorithms or software that identify changes or inconsistencies in statements from the results of comparisons.

[0114] "Means for visualizing extracted differences as graphs or charts" refers to a function or device for displaying the difference information extracted through comparison in a visually easy-to-understand format.

[0115] "Means for notifying users" refers to a function or device for sending information to users via text message, email, or other means of communication.

[0116] "Means for collecting feedback from users" refers to a function or device that allows users to input comments and opinions into the system and collects that information.

[0117] "Means for storing collected feedback data" refers to a database or storage system for storing feedback information obtained from users.

[0118] "Means for sending voice data to a voice recognition engine in real time and obtaining it as text data" refers to a function or device that sends voice during a meeting to a voice recognition system (API, etc.) in real time and receives the resulting text data.

[0119] "Means of using natural language processing algorithms to compare the content of multiple meetings and analyze the consistency of the content with high accuracy" refers to algorithms and software that use natural language processing technology to perform detailed analysis of the content of statements in different meeting records and evaluate their consistency.

[0120] The present invention relates to a system for transcribing statements made during a meeting in real time and comparing the results with past minutes. Specific embodiments of the system are described below.

[0121] First, when a user starts a meeting, they begin recording audio on their device. The device can be a smartphone or laptop, and recording is done using the device's built-in microphone. The recorded audio data is sent directly to a speech recognition engine in real time. Specifically, Google Cloud Speech-to-Text API or IBM Watson Speech to Text can be used. The device divides the audio data into buffers of a certain size, makes an API call, and converts the audio into text data.

[0122] The generated text data is then sent from the device to the server. The server receives it and temporarily stores it in a database as provisional minutes. Databases such as MySQL or PostgreSQL are often used for this purpose. The server accesses the database of past meeting minutes and retrieves past minutes related to the current meeting. For example, it extracts the data by executing the SQL query "SELECT FROM meeting_minutes WHERE meeting_id = 'XX'".

[0123] The server then uses natural language processing algorithms to compare the current meeting's content with past meeting minutes. These algorithms use models such as BERT and GPT-3. These algorithms tokenize and preprocess the text data before feeding it into a comparison model. The algorithm generates a similarity score for each utterance and detects variations or inconsistencies in the content.

[0124] The server converts the detected differences into graphs and charts and displays them in a visually easy-to-understand format. Libraries such as Matplotlib and D3.js are used for this visualization. For example, it is possible to display the differences in the content of comments over time as a bar graph.

[0125] The visualized data is sent from the server to the user via email or a notification system. Emails are sent using the SMTP protocol, and notifications are sent using the Slack API, etc. Users can review the data and ask their superiors or managers why the comments were changed.

[0126] Finally, the user enters their feedback into a form on the system and submits it. The server receives this feedback and stores it in a database, for example using the SQL query "INSERT INTO feedback (meeting_id, feedback_text) VALUES (XX, 'YY')".

[0127] As a concrete example, let's consider a case where a boss says in Meeting A, "We'll start a new project next week," and then in the next Meeting B, "The new project will start the week after next." In this case, the following behavior will occur.

[0128] 1. Recording and Transcription:

[0129] The device records what is said in Meeting B and converts it into text in real time using Google Cloud Speech-to-Text.

[0130] This text data is sent to the server and temporarily stored.

[0131] 2. Comparing minutes:

[0132] The server retrieves the minutes data for Meeting A from the MySQL database.

[0133] Use the BERT model to compare what was said in Meeting A and Meeting B.

[0134] 3. Extract and visualize the differences:

[0135] The server detects the change in the content of the message from "next week" to "the week after next."

[0136] The differences are graphed using Matplotlib and notified to the user.

[0137] 4. Getting feedback:

[0138] The user checks the notification and asks their boss why the comment has changed.

[0139] The feedback obtained is entered into the system, and the server stores it in a MySQL database.

[0140] This system makes it easy to record meeting content in real time and compare it with past parliamentary sessions, and allows for visual confirmation of consistency and fluctuations in speech content, enabling efficient and transparent meeting management.

[0141] Example prompt sentence:

[0142] Example: "Write a Python script to create a system that transcribes what is said during meetings in real time and compares it with past transcripts to visualize fluctuations in what is said."

[0143] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0144] Step 1:

[0145] Start audio recording:

[0146] When the conference starts, the user starts recording audio on the terminal.

[0147] Specific behavior: The user launches a recording application on their laptop or smartphone and clicks the "Start Recording" button.

[0148] Input: User's voice.

[0149] Output: Recorded audio data.

[0150] Step 2:

[0151] Real-time transcription of audio data:

[0152] The device splits the audio data being recorded and sends it to the voice recognition engine.

[0153] Specific operation: The terminal program captures voice data in a certain buffer size, sends the data to, for example, the Google Cloud Speech-to-Text API, and receives the text data obtained from the API.

[0154] Input: Recorded audio data.

[0155] Output: Real-time converted text data.

[0156] Step 3:

[0157] Sending text data to the server and temporarily saving it:

[0158] The terminal sends the text data returned from the voice recognition engine to the server.

[0159] Specific operation: The terminal sends an HTTP POST request to the server, and the server temporarily stores the received data in a MySQL database.

[0160] Input: Real-time text data.

[0161] Output: Text data temporarily stored in the server database.

[0162] Step 4:

[0163] Obtaining past minutes data:

[0164] The server accesses a past minutes database to retrieve relevant past meeting minutes.

[0165] Specific operation: The server executes the query "SELECT FROM meeting_minutes WHERE meeting_id = 'XX'" to retrieve the data.

[0166] Input: Past meeting ID or date and time.

[0167] Output: The retrieved past minutes data.

[0168] Step 5:

[0169] Comparison of current conference speeches with past minutes:

[0170] The server uses a natural language processing algorithm to compare the acquired minutes data with the current meeting text data.

[0171] Specific operation: Using natural language processing models such as BERT or GPT-3, the text data is tokenized and a similarity score is calculated.

[0172] Input: Current meeting text data and past meeting minutes data.

[0173] Output: Comparison results showing variability and inconsistencies in what was said.

[0174] Step 6:

[0175] Visualizing the differences:

[0176] The server visualizes the differences in the detected statements in graphs and charts.

[0177] Specific operation: Using Matplotlib and D3.js, fluctuations are displayed using bar graphs and heat maps along the time axis.

[0178] Input: Difference data from comparison.

[0179] Output: Visually displayed difference graphs and charts.

[0180] Step 7:

[0181] Sending visualization data:

[0182] The server notifies the user of the visualized data.

[0183] Specific behavior: Sends an email using the SMTP protocol or a notification using the Slack API.

[0184] Input: Visualization data.

[0185] Output: Notification to the user (email or notification message).

[0186] Step 8:

[0187] Get and store feedback:

[0188] The user checks the notification and enters feedback into the system if necessary, and the server stores the feedback.

[0189] What happens: A user enters feedback into a form on the system and clicks the "Submit" button. The server receives the data and stores it in a MySQL database.

[0190] Input: User feedback.

[0191] Output: Feedback data stored in a database.

[0192] (Application example 1)

[0193] 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."

[0194] With conventional meeting management systems, it was difficult to record what was said during meetings in real time and compare it with past minutes to maintain consistency. Furthermore, in factories and other workplaces, it was difficult to grasp the consistency and fluctuations of work instructions, which led to a decline in work efficiency. To solve these problems, a real-time work instruction monitoring system was required.

[0195] 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.

[0196] In this invention, the server includes means for transcribing statements made during a meeting in real time, means for saving minutes of past meetings, means for comparing transcription data of current meeting statements with past minutes data, means for extracting variations and inconsistencies in the content of statements, means for visualizing the extracted differences as graphs or charts, means for notifying users, means for collecting feedback from users, means for saving the collected feedback data, means for acquiring work instructions through real-time speech recognition, means for comparing with a database of past work instructions, means for detecting and visualizing differences, means for displaying the detected differences on a display, and means for sending user feedback to a manager. This makes it possible to monitor the consistency of statements and work instructions in real time and improve transparency.

[0197] A "meeting" is a gathering of participants to discuss and make decisions about a specific topic.

[0198] "Transcription" is the process of converting audio into text, either in real time or at a later time.

[0199] "Minutes" are documents that record the contents of a meeting and the decisions made.

[0200] "Minutes data" refers to the contents of past minutes saved as digital data.

[0201] A "natural language processing algorithm" is a technology that allows computers to analyze and process human language.

[0202] "Graphs and charts" are diagrams and tables that visually represent data.

[0203] A "notification" is a message or alert that informs the user of information.

[0204] "Feedback" refers to opinions and evaluation information provided by users.

[0205] "Speech recognition" is a technology that converts speech into text.

[0206] "Work instructions" are information for instructing a specific work.

[0207] "Work Order Database" means a digital database for storing past and current work orders.

[0208] "Difference" refers to the difference between the objects being compared.

[0209] A "display" is a device for visually displaying information.

[0210] An "administrator" is a person in charge of running a system or project.

[0211] This invention is a system that transcribes statements made during meetings in real time and compares them with past minutes. Furthermore, this technology is applied to factory robots to provide a real-time work instruction monitoring system that monitors the consistency of work instructions.

[0212] 1. Hardware and software configuration:

[0213] Hardware: Microphones and displays built into smartphones, smart glasses, head-mounted displays, or robots.

[0214] Software: We use the "speech_recognition" library for speech recognition, the "spacy" library for natural language processing, the "matplotlib" library for data visualization, and the "json" library for data storage and processing.

[0215] 2. Data processing and calculation processing:

[0216] Speech recognition: The device uses a microphone to collect speech and instructions in real time during meetings or factory work, then converts the speech data into text using a speech recognition engine (for example, the "speech_recognition" library).

[0217] Data transmission and storage: The generated text data is transmitted to the server in real time and temporarily stored as provisional minutes or work instructions.

[0218] Comparison and analysis: The server accesses a database of past meeting minutes or work instructions and compares them with the current text data. This is done using natural language processing algorithms (e.g., the "spacy" library).

[0219] Difference detection and visualization: Detect changes in what is said or what is directed, and visualize the differences as graphs or charts (for example, graphs created with the “matplotlib” library).

[0220] Notification and feedback: The visualized data is displayed to the user on the screen and feedback is collected through a notification system if necessary. The feedback is sent back to the server and stored in a database.

[0221] 3. Example:

[0222] If a boss says in Meeting A, "We'll start a new project next week," and then in the next Meeting B, the boss changes his mind and says, "The new project will start the week after next," the system will detect the change in real time and notify the user.

[0223] On a factory floor, if a robot receives the work instruction "Please start adjusting the mechanical parts on production line 1," and then the instruction is changed to "Please start adjusting the mechanical parts on production line 3," the robot will similarly detect the change and display it on the screen.

[0224] Example prompt sentence:

[0225] Take current work orders and compare them with past work orders to detect changes.

[0226] This structure allows for greater transparency and consistency in meetings and factory operations.

[0227] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0228] Step 1:

[0229] The device uses a microphone to capture the voice of the person speaking or working on the device, and this voice data is collected in real time and used as input data for subsequent processing.

[0230] Step 2:

[0231] The device sends the acquired voice data to a voice recognition engine (for example, the "speech_recognition" library) and converts it into text data. This process outputs the voice data as transcribed text.

[0232] Step 3:

[0233] The device sends the generated text data to the server, which temporarily stores it and registers it as the current minutes or work instructions.

[0234] Step 4:

[0235] The server retrieves corresponding data from a database of past meeting minutes or work instructions. This past data is text data and is used for comparison with the current data.

[0236] Step 5:

[0237] The server uses natural language processing algorithms (e.g., the "spacy" library) to compare the current text data with past data, detecting variations or inconsistencies in the data and reporting the differences.

[0238] Step 6:

[0239] The server generates visualization data such as graphs and charts based on the detected differences. In this process, the data is processed to visually represent the degree and content of the differences, and the results are output as visualization data.

[0240] Step 7:

[0241] The server sends the visualization data to the terminal, which displays it on the display, allowing the user to check the differences.

[0242] Step 8:

[0243] The user can then input feedback about any discrepancies they find, which is then sent to the server and used for future analysis.

[0244] Step 9:

[0245] The server stores the collected feedback data in a database, which stores the feedback for future analysis and system improvement.

[0246] 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.

[0247] This invention combines an emotion engine with a system that transcribes statements made during meetings in real time and compares them with past minutes. This system not only visualizes inconsistencies in the content of statements made during meetings and monitors the consistency of statements, but also recognizes and analyzes the emotions of users and speakers.

[0248] First, when a meeting begins, the user starts recording audio on their device. The device sends what is said in the meeting to a speech recognition engine in real time, which converts it into text. This text data is then sent directly to the server in real time and temporarily saved as provisional minutes.

[0249] The server accesses a database of past meeting minutes and compares the current meeting's content with past meeting minutes using natural language processing algorithms that detect variations or inconsistencies in the content of the conversation.

[0250] Furthermore, the server uses an emotion engine to analyze the emotions of users and speakers in real time. The emotion engine recognizes emotions based on tone of voice, content of speech, and facial expression data (if a camera is used), and stores the results as analysis data.

[0251] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts, allowing users to see not only the changes in the content of comments but also the emotions expressed at the time of the comments at a glance. The server then sends this visualized data to users via email or a notification system.

[0252] Users can review the data they receive and, if necessary, ask their superiors or managers why the content of their comments was changed. The users enter their feedback into the system, which the server stores. This feedback data is used to analyze and improve future meetings.

[0253] Specific examples

[0254] For example, suppose that during Meeting A, a boss says, "We'll start a new project next week." Then during the next Meeting B, the same boss says, "We'll start a new project the week after next." This system works as follows:

[0255] 1. Recording and Transcription:

[0256] The device records what is said in Conference B and converts it into text in real time.

[0257] The generated text data is sent to the server.

[0258] 2. Comparing minutes:

[0259] The server retrieves the minutes data for Conference A from the database.

[0260] Using a natural language processing algorithm, the content of statements made in Meeting A and Meeting B is compared.

[0261] 3. Extract and visualize the differences:

[0262] The server detects the change in the content of the message and discovers the change from "next week" to "the week after next."

[0263] The differences are visualized as a graph and sent to the user.

[0264] 4. Emotion analysis:

[0265] The emotion engine analyzes the emotions expressed by the boss when he or she speaks and identifies emotions such as "anxiety."

[0266] Emotional data is also visualized as a graph and sent to the user.

[0267] 5. Getting feedback:

[0268] The user checks the notification and asks their boss why the comment has changed.

[0269] The obtained feedback is entered into the system and the server stores it in a database.

[0270] In this way, the system not only records what is being said in real time and compares it with past comments, but also analyzes emotions to provide deeper insights, allowing meeting participants to monitor both the consistency of what is being said and the emotional fluctuations, improving the transparency and efficiency of meetings.

[0271] The processing flow will be explained below.

[0272] Step 1:

[0273] Start audio recording

[0274] The terminal activates the audio recording function at the start of the conference.

[0275] The device records what is said in the meeting in real time via a microphone.

[0276] Step 2:

[0277] Real-time audio transcription

[0278] The device sends the recorded voice data to the voice recognition engine in real time.

[0279] The device receives the text data returned from the voice recognition engine and converts the spoken content into text.

[0280] Step 3:

[0281] Generate and save provisional minutes

[0282] The device transmits the text of the speech in real time to the server.

[0283] The server temporarily stores the received text data as provisional minutes.

[0284] Step 4:

[0285] Obtaining past meeting minutes data

[0286] The server retrieves past meeting minutes data from the database.

[0287] The server prepares the acquired minutes data for comparison with today's minutes data.

[0288] Step 5:

[0289] Comparison of meeting minutes data

[0290] The server uses natural language processing (NLP) algorithms to compare today's minutes with past minutes.

[0291] The server detects changes in keywords and phrases in the content of comments.

[0292] Step 6:

[0293] Emotion analysis

[0294] The server uses an emotion engine to analyze the emotions of speakers in real time during a conference.

[0295] The device sends the speaker's tone of voice and facial expression data (if necessary) to the emotion engine.

[0296] Step 7:

[0297] Storing Emotional Data

[0298] The server receives the analyzed emotion data and stores it together with the provisional minutes data.

[0299] Step 8:

[0300] Extracting Differences

[0301] The server identifies fluctuations and inconsistencies in what is being said.

[0302] The server builds a list of the identified differences and aggregates the data.

[0303] Step 9:

[0304] Visualization of differential data and emotion data

[0305] The server generates visual representations such as graphs and charts based on the extracted differences and emotion data.

[0306] The server formats the visualization data into a format that is easy for the user to understand.

[0307] Step 10:

[0308] User Notification

[0309] The server adds explanatory text to the visualized data and sends it to the specified user as an email or internal notification.

[0310] The user checks the received notification and understands changes in the content of the comments and the results of sentiment analysis.

[0311] Step 11:

[0312] Get feedback

[0313] The user checks the notification and, if necessary, asks their superior or manager about the reason for the change in the comment content.

[0314] The user enters this feedback into the system.

[0315] Step 12:

[0316] Feedback Data Storage

[0317] The server receives the feedback data sent by the user and stores it in a database.

[0318] The server will prepare to use the data for future analysis and improvements.

[0319] These steps allow the system to efficiently record what is said during a meeting, compare it with past comments, and analyze and visualize the emotions of the speakers. Based on this information, users can identify the reasons for fluctuations in comments and take action to improve the quality of the meeting.

[0320] Example 2

[0321] 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."

[0322] Conventional meeting minutes systems lack the functionality to transcribe speeches made during meetings in real time and compare them with past minutes. It is also difficult to visualize fluctuations or inconsistencies in the speech content of meeting participants. Furthermore, there is no way to analyze the emotions of speakers and gain a deeper understanding of the meeting's progress, so the transparency and efficiency of meetings are not sufficiently ensured. To solve these issues, a system is needed that not only transcribes speeches in real time, but also compares speech content and analyzes emotions, and visualizes the results for users.

[0323] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for transcribing statements made during a meeting in real time, means for saving past meeting minutes, means for comparing transcription data of current meeting statements with past meeting minutes data, means for analyzing the emotions of users and speakers, and means for visualizing the extracted differences and emotion analysis data as graphs or charts. This makes it possible to improve the transparency and efficiency of meetings by comparing the consistency of statements made during a meeting with the contents of past statements, analyzing the emotions of speakers, and visually providing this to the user.

[0324] "A means for transcribing statements made during a meeting in real time" is a system that records audio during a meeting in real time and instantly converts it into text using voice recognition technology.

[0325] The "means for storing minutes of past meetings" is a system that stores the contents of statements made in previous meetings as text data in a database or the like.

[0326] "Means for comparing transcription data of current meeting remarks with past minutes data" is a function that uses natural language processing technology to compare the content of remarks made in the most recent meeting with past meeting records, and detects fluctuations or lack of consistency in the content.

[0327] "Means to extract variations and lack of consistency in speech content" is a function that identifies how speech made during a meeting differs from past speech and whether consistency is maintained, and clearly shows the differences.

[0328] "Means for analyzing the emotions of users and speakers" refers to technology that analyzes and identifies the emotional state of speakers during a meeting based on their voice, text, and, if necessary, facial expression data.

[0329] "Means for visualizing extracted differences and sentiment analysis data as graphs or charts" refers to a system that converts fluctuations in speech content and sentiment analysis results into visual data, i.e., visual formats such as graphs and charts, and displays them.

[0330] "Means for notifying users" refers to a function for delivering analysis results and visualization data to users via email or a notification system.

[0331] "Means for collecting feedback from users" is a function that collects opinions and responses that users input to the system and saves them as data.

[0332] The "means for storing collected feedback data" is a function for storing the feedback information collected from users in a database or the like and managing it for use in later analysis and improvement.

[0333] This invention combines an emotion engine with a system that transcribes statements made during meetings in real time and compares them with past minutes. The system aims to improve the transparency and efficiency of meetings by visualizing fluctuations and inconsistencies in the content of statements and analyzing the emotions of speakers.

[0334] First, when a meeting begins, the user starts recording audio on their device. The device then uses a speech recognition engine such as the Google Cloud Speech-to-Text API to convert what is being said into text in real time. This text data is then sent to a server in real time and saved as temporary meeting minutes.

[0335] The server accesses a database of past meeting minutes and compares the current meeting's content with past minutes using natural language processing techniques such as BERT and GPT. This comparison detects fluctuations and inconsistencies in the content of the speech.

[0336] Furthermore, the server uses an emotion analysis engine such as IBM Watson Tone Analyzer to analyze the emotions of users and speakers in real time. This analysis involves recognizing emotions based on voice tone, speech content, and, in some cases, facial expression data captured by a camera, and saving the analysis data.

[0337] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts, allowing users to see not only the changes in the content of comments but also the emotions expressed at the time of the comments at a glance. The server then sends this visualized data to users via email or a notification system.

[0338] The user can review the data and, if necessary, ask their superiors or managers why the content was changed. The user then enters the feedback they received into the system, and the server stores this feedback data. This feedback data can be used to analyze and improve future meetings.

[0339] Specific examples

[0340] Examples of Meeting A and Meeting B

[0341] For example, if a boss says "We'll start a new project next week" during Meeting A, and then the same boss says "We'll start a new project the week after next" during a subsequent Meeting B, the system will operate as follows:

[0342] 1. Recording and Transcription:

[0343] The device records what is said in Conference B and converts it into text in real time. This text data is then sent to the server.

[0344] 2. Comparing minutes:

[0345] The server retrieves the minutes of Meeting A from the database and uses natural language processing technology to compare the content of statements made in Meeting A and Meeting B.

[0346] 3. Extract and visualize the differences:

[0347] The server detects changes in the content of comments and detects changes from "next week" to "the week after next." This is visualized as a graph and sent to the user.

[0348] 4. Emotion analysis:

[0349] The emotion engine analyzes the emotion expressed at the time of speech and identifies emotions such as "anxiety." This is also visualized as a graph and sent to the user.

[0350] 5. Getting feedback:

[0351] The user checks the notification and asks their superior why the comment has changed. The feedback is then entered into the system, and the server stores it in a database.

[0352] Prompt Sentence Examples

[0353] An example of a prompt to input to a generative AI model is as follows:

[0354] "Discuss a system that transcribes speech during a meeting in real time and compares it with past minutes. In this system, the user starts recording, and the data is converted to text through a speech recognition engine and sent to a server. The server compares the speech using a natural language processing algorithm and analyzes the speaker's sentiment using an emotion engine. Finally, the system visualizes the differences and the sentiment analysis results and notifies the user."

[0355] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0356] Step 1: Start recording and real-time transcription

[0357] When a meeting begins, users start recording audio on their device, which then uses a speech recognition engine (Google Cloud Speech-to-Text API) to convert what is being said into text in real time.

[0358] Input: Meeting audio

[0359] Data processing: The voice recognition engine converts the voice data into text data.

[0360] Output: Text data generated in real time

[0361] Specific operation: The user presses the "Start Recording" button in the application. The device collects audio data through the microphone and converts the speech into text using the Google Cloud Speech-to-Text API.

[0362] Step 2: Sending real-time data to the server

[0363] The device sends the generated text data in real time to the server, which temporarily stores the data as provisional minutes.

[0364] Input: Real-time generated text data

[0365] Data processing: Send text data to the server via the network

[0366] Output: Text data saved as provisional minutes

[0367] Specific operation: The terminal sends text data to the server at regular intervals, and the server temporarily stores the received data.

[0368] Step 3: Obtain and compare past minutes

[0369] The server accesses the database and compares the current meeting content with past meeting minutes using natural language processing techniques (such as BERT and GPT).

[0370] Input: Current meeting text data and past meeting minutes data

[0371] Data processing: Analyzing text data using natural language processing techniques to detect variations and inconsistencies in spoken content

[0372] Output: Data showing fluctuations and inconsistencies in speech

[0373] How it works: The server extracts keywords from the stored text data, queries the database to retrieve relevant past meeting transcripts, and then applies natural language processing techniques such as the BERT model to identify variations and inconsistencies in the speech.

[0374] Step 4: Perform sentiment analysis

[0375] The server uses an emotion engine (such as IBM Watson Tone Analyzer) to analyze the emotions of users and speakers in real time.

[0376] Input: Meeting speech text and voice data, and facial expression data if necessary

[0377] Data processing: Using an emotion engine to analyze the speaker's emotions from voice and text

[0378] Output: Parsed emotion data

[0379] Specific operation: The server inputs the acquired voice, text, and facial expression data into the emotion engine, and then quantifies and stores the analyzed emotional state.

[0380] Step 5: Visualize the difference and sentiment analysis data

[0381] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts.

[0382] Input: Data on fluctuations in speech content and sentiment analysis data

[0383] Data processing: Generate graphs and charts using visualization tools (e.g., D3.js)

[0384] Output: Visual data (graphs and charts)

[0385] Specific operation: The server uses visualization tools to generate graphs and charts that highlight fluctuations and emotional changes.

[0386] Step 6: Communicate data and get feedback

[0387] The server sends the generated visualization data to the user via email or a notification system. The user checks the notification and, if necessary, asks their superior or manager for the reason for the change in the comment content.

[0388] Input: Visual data

[0389] Data processing: Distributing data via email and notification systems

[0390] Output: Visual data provided to the user

[0391] Specific operation: The server generates the visual data in PDF or web link format and sends a notification to the user, who receives the notification and asks his / her boss for feedback.

[0392] Step 7: Save the feedback and use it in your next meeting

[0393] Users input feedback from their superiors and managers into the system, and the server stores this feedback data and uses it to analyze and improve the next meeting.

[0394] Input: User feedback information

[0395] Data processing: Save the feedback data to a database

[0396] Output: Saved feedback data

[0397] Specific operation: The user enters comments and reasons on the feedback input screen, and the server stores them in the database. The feedback data is then used as a reference for the next analysis.

[0398] (Application example 2)

[0399] 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."

[0400] Conventional autonomous vehicles lack the means to monitor driver and passenger conversations in real time and improve safety based on that information. They also lack the ability to detect risks such as stress and lack of attention while driving in advance. Furthermore, there is a lack of comparative analysis with past driving records, making it difficult to respond appropriately to consistency and fluctuations.

[0401] The identification process by the identification 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 means for transcribing conversation content in real time, means for saving past conversation records, means for comparing transcription data of the current conversation content with past conversation record data, means for extracting fluctuations and inconsistencies in the conversation content, means for visualizing the extracted differences and emotion analysis data as graphs or charts, means for notifying the user, means for collecting feedback from the user, and means for saving the collected feedback data. This makes it possible to monitor conversation content and emotional fluctuations while driving in real time, thereby improving safety.

[0402] A "means for transcribing conversation content in real time" is a means that has the function of converting conversation from voice to text in real time.

[0403] The "means for saving records of past conversations" refers to a means having a function for saving text data of conversations recorded in the past.

[0404] "Means for comparing transcription data of current conversation content with past conversation record data" refers to a means that has the function of comparing data currently transcribed in real time with previously saved conversation record data to verify changes and consistency in content.

[0405] The "means for extracting fluctuations and inconsistencies in conversation content" refers to a means having a function for detecting and extracting changes and inconsistencies in conversation content.

[0406] "Means for visualizing extracted differences and emotion analysis data as graphs or charts" refers to means that has the function of generating graphs or charts to visually represent detected differences and analyzed emotion data.

[0407] The "means for notifying the user" refers to a means having a function for notifying the user of detected information and analysis results.

[0408] The "means for collecting feedback from users" refers to means that have the function of collecting opinions and information provided by users.

[0409] The "means for storing collected feedback data" refers to a means having a function for storing collected user feedback data.

[0410] "Audio recording means" means a means capable of recording conversations within a vehicle and storing the recording for later use.

[0411] A "means for analyzing emotions" is a means that has the function of identifying and analyzing emotions from the content of a conversation and tone of voice.

[0412] "Means for comparing conversation content using natural language processing algorithms" means means that use natural language processing technology to compare current conversation content with past records and verify consistency and variation.

[0413] This invention combines a sentiment analysis engine with a system that transcribes conversation content in real time, compares it with past conversation records, and monitors conversation fluctuations and utterance consistency. To realize this system, the following means are required.

[0414] First, the device records conversations in the car in real time and transcribes them. The device captures audio using a built-in microphone and converts this audio into text data in real time. This process uses voice recognition technology using the speech_recognition library.

[0415] The resulting text data is immediately sent to a server and temporarily stored as a provisional conversation record. The server then accesses a database of past conversation records and compares the current conversation with the past records using a natural language processing algorithm (NLP). This comparison method uses Hugging Face's natural language processing model.

[0416] The server then uses an emotion analysis engine to analyze emotions from the content of the conversation and the tone of voice. Hugging Face's emotion analysis model identifies emotions such as "stress," "impatience," and "fatigue" based on the text data obtained. This allows the server to understand the emotional state of the driver and passengers.

[0417] Furthermore, the server visualizes the comparison results and sentiment analysis data by using the matplotlib library to generate graphs and charts of the detected differences and sentiment data, converting them into a visually easy-to-understand format.

[0418] This visualized data is notified to the user (driver or in-car supervisor) in real time. For example, if the driver repeatedly says "I'm tired," the system analyzes in real time and issues a warning when it detects the emotion of "fatigue." The user can check this notification and take appropriate action if necessary.

[0419] The system also includes a means for users to provide feedback. Users receive notifications from the system and can enter their opinions or additional information into the system. The collected feedback data is stored in a database by the server and used for future conversation analysis.

[0420] As a specific example, "If a driver on their way to the office repeatedly says 'I'm tired,' the system will use its sentiment analysis engine to determine that the driver is 'fatigued' and issue a real-time warning." This prompt would be written as follows:

[0421] "If a driver on the way to the office repeatedly says 'I'm tired,' the system should use its sentiment analysis engine to determine that the driver is 'fatigued,' and devise a program to issue a real-time warning."

[0422] As described above, this invention not only records conversations in real time and compares them with past records, but also analyzes emotions, thereby improving safety and efficiency while driving.

[0423] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0424] Step 1:

[0425] The device uses the in-car microphone to record audio in real time. The input is in-car audio data, and the output is a recorded audio file. Specifically, the device's microphone picks up the conversations between the driver and passengers and records them as audio data.

[0426] Step 2:

[0427] The device transcribes recorded audio in real time. The input is an audio file, and the output is transcribed text data. The device uses the speech_recognition library to analyze the audio and convert it to text.

[0428] Step 3:

[0429] The transcribed text data is sent to a server and temporarily saved as a provisional conversation record. The input is the transcribed data, and the output is the provisional text record data on the server. The device sends the data to the server via network communication.

[0430] Step 4:

[0431] The server accesses a database of past conversation records and compares the current conversation content with past records using a natural language processing algorithm. The input is the current conversation text data and the past conversation record data, and the output is the comparison result data. Specifically, the server uses Hugging Face's NLP model to detect differences and inconsistencies in the text.

[0432] Step 5:

[0433] The server uses an emotion analysis engine to analyze emotions from the content of the conversation and tone of voice. The input is transcribed text data, and the output is analyzed emotional data. The server uses Hugging Face's emotion analysis model to identify emotions such as "stress," "anxiety," and "fatigue."

[0434] Step 6:

[0435] The server visualizes the detected differences and sentiment analysis data as graphs or charts. The input is the comparison results and sentiment data, and the output is a visual graph or chart. The server uses the matplotlib library to visually represent the detected variations and sentiment.

[0436] Step 7:

[0437] The server generates visualization data and notifies the user in real time. The input is the visualized data, and the output is a notification message to the user. For example, if the driver repeatedly says "I'm tired," the server generates a warning message and sends it to the user.

[0438] Step 8:

[0439] The user checks the notification from the system and provides feedback. The input is the user's feedback, and the output is the feedback data. The user checks the notification content and enters their opinions and information corresponding to that content into the system.

[0440] Step 9:

[0441] The server stores the collected feedback data and uses it for future conversation analysis. The input is the feedback data and the output is the stored feedback database. The server stores these data in a database and makes them available for subsequent dialogue analysis.

[0442] The above are the processing steps for realizing the invention.

[0443] 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.

[0444] 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.

[0445] 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.

[0446] [Second embodiment]

[0447] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0448] 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.

[0449] 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).

[0450] 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.

[0451] 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.

[0452] 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).

[0453] 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.

[0454] 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.

[0455] 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.

[0456] 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.

[0457] 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.

[0458] 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."

[0459] This invention relates to a system that transcribes statements made during meetings in real time and compares them with past minutes. This system visualizes inconsistencies in the content of statements made during meetings and monitors the consistency of statements.

[0460] First, when a meeting begins, the user starts recording audio on their device. The device sends what is said in the meeting to a speech recognition engine in real time, which converts it into text. This text data is then sent directly to the server in real time and temporarily saved as provisional minutes.

[0461] The server then accesses a database of past meeting minutes and compares the current meeting's content with past meeting minutes using natural language processing algorithms that detect variations or inconsistencies in the content.

[0462] The server converts the detected differences into visual formats such as graphs and charts, allowing users to see the changes in the content at a glance. The server then sends this visualization data to users via email or a notification system.

[0463] Users can review the data they receive and, if necessary, ask their superiors or managers why the content of their comments was changed. The users enter their feedback into the system, which the server stores. This feedback data is used to analyze and improve future meetings.

[0464] Specific examples

[0465] For example, suppose that during Meeting A, a boss says, "We'll start a new project next week." Then during the next Meeting B, the same boss says, "We'll start a new project the week after next." This system works as follows:

[0466] 1. Recording and Transcription:

[0467] The device records what is said in Conference B and converts it into text in real time.

[0468] The generated text data is sent to the server.

[0469] 2. Comparing minutes:

[0470] The server retrieves the minutes data for Conference A from the database.

[0471] Using a natural language processing algorithm, the content of statements made in Meeting A and Meeting B is compared.

[0472] 3. Extract and visualize the differences:

[0473] The server detects the change in the content of the message and discovers the change from "next week" to "the week after next."

[0474] The differences are visualized as a graph and sent to the user.

[0475] 4. Getting feedback:

[0476] The user checks the notification and asks their boss why the comment has changed.

[0477] The obtained feedback is entered into the system and the server stores it in a database.

[0478] In this way, this system records what is being said in real time and compares it with past comments to clarify inconsistencies in what is being said, improving the transparency and efficiency of meetings.

[0479] The processing flow will be explained below.

[0480] Step 1:

[0481] Start audio recording

[0482] The terminal activates the audio recording function at the start of the conference.

[0483] The device records what is said in the meeting in real time via a microphone.

[0484] Step 2:

[0485] Real-time audio transcription

[0486] The device sends the recorded voice data to the voice recognition engine in real time.

[0487] The device receives the text data returned from the voice recognition engine and converts the spoken content into text.

[0488] Step 3:

[0489] Sending text data and generating provisional minutes

[0490] The device transmits the text of the speech in real time to the server.

[0491] The server temporarily stores the received text data as provisional minutes.

[0492] Step 4:

[0493] Obtaining past meeting minutes data

[0494] The server retrieves past meeting minutes data from the database.

[0495] The server prepares the acquired minutes data for comparison with today's minutes data.

[0496] Step 5:

[0497] Comparison of meeting minutes data

[0498] The server uses natural language processing (NLP) algorithms to compare today's minutes with past minutes.

[0499] The server detects changes in keywords and phrases in the content of comments.

[0500] Step 6:

[0501] Extracting Differences

[0502] The server identifies fluctuations and inconsistencies in what is being said.

[0503] The server builds a list of the identified differences and aggregates the data.

[0504] Step 7:

[0505] Visualizing differential data

[0506] The server generates visual representations such as graphs and charts based on the extracted differences.

[0507] The server formats the visualization data into a format that is easy for the user to understand.

[0508] Step 8:

[0509] User Notification

[0510] The server adds explanatory text to the visualized data and sends it to the specified user as an email or internal notification.

[0511] The user checks the received notification and understands changes in the content of the comments.

[0512] Step 9:

[0513] Get feedback

[0514] The user checks the notification and, if necessary, asks their superior or manager about the reason for the change in the comment content.

[0515] The user enters this feedback into the system.

[0516] Step 10:

[0517] Feedback Data Storage

[0518] The server receives the feedback data sent by the user and stores it in a database.

[0519] The server will prepare to use the data for future analysis and improvements.

[0520] These steps allow the system to efficiently record what is said during a meeting, compare it with past comments to identify inconsistencies and changes, and report them back to meeting participants, allowing users to understand the reasons for fluctuations and take action to improve the quality of the meeting.

[0521] Example 1

[0522] 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."

[0523] Previously, it was difficult to record statements made during meetings in real time and understand the consistency and changes in the content of those statements. As a result, important decision-making and information transmission in meetings became uncertain, making efficient business operations difficult. Furthermore, comparing past minutes with current statements was time-consuming and labor-intensive. This made ensuring the transparency and reliability of meeting content a challenge.

[0524] 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.

[0525] In this invention, the server includes: means for transcribing statements made during a meeting in real time; means for saving past meeting minutes; means for comparing the transcription data of current meeting statements with past meeting minutes data; means for extracting variations and inconsistencies in the content of statements; means for visualizing the extracted differences as graphs or charts; means for notifying users; means for collecting feedback from users; means for saving the collected feedback data; means for transmitting voice data to a voice recognition engine in real time and acquiring it as text data; and means for comparing multiple meeting statements and analyzing the consistency of the statement content with high accuracy using a natural language processing algorithm. This automates the real-time recording and comparison of meeting content and makes it possible to visually confirm the consistency and variations in the content of statements, enabling efficient and transparent meeting management.

[0526] "Means for transcribing statements made during a meeting in real time" refers to a function or device that instantly converts speech made during a meeting into text data using voice recognition technology.

[0527] "Means for storing minutes of past meetings" refers to a function or device that stores the contents of past meetings as text data in a database or the like.

[0528] "Means for comparing the current meeting transcript with past meeting minutes" refers to algorithms or software that compare the transcript generated from the most recent meeting with similarly stored meeting minutes from the past.

[0529] "Means for detecting variations or inconsistencies in statements" refers to algorithms or software that identify changes or inconsistencies in statements from the results of comparisons.

[0530] "Means for visualizing extracted differences as graphs or charts" refers to a function or device for displaying the difference information extracted through comparison in a visually easy-to-understand format.

[0531] "Means for notifying users" refers to a function or device for sending information to users via text message, email, or other means of communication.

[0532] "Means for collecting feedback from users" refers to a function or device that allows users to input comments and opinions into the system and collects that information.

[0533] "Means for storing collected feedback data" refers to a database or storage system for storing feedback information obtained from users.

[0534] "Means for sending voice data to a voice recognition engine in real time and obtaining it as text data" refers to a function or device that sends voice during a meeting to a voice recognition system (API, etc.) in real time and receives the resulting text data.

[0535] "Means of using natural language processing algorithms to compare the content of multiple meetings and analyze the consistency of the content with high accuracy" refers to algorithms and software that use natural language processing technology to perform detailed analysis of the content of statements in different meeting records and evaluate their consistency.

[0536] The present invention relates to a system for transcribing statements made during a meeting in real time and comparing the results with past minutes. Specific embodiments of the system are described below.

[0537] First, when a user starts a meeting, they begin recording audio on their device. The device can be a smartphone or laptop, and recording is done using the device's built-in microphone. The recorded audio data is sent directly to a speech recognition engine in real time. Specifically, Google Cloud Speech-to-Text API or IBM Watson Speech to Text can be used. The device divides the audio data into buffers of a certain size, makes an API call, and converts the audio into text data.

[0538] The generated text data is then sent from the device to the server. The server receives it and temporarily stores it in a database as provisional minutes. Databases such as MySQL or PostgreSQL are often used for this purpose. The server accesses the database of past meeting minutes and retrieves past minutes related to the current meeting. For example, it extracts the data by executing the SQL query "SELECT FROM meeting_minutes WHERE meeting_id = 'XX'".

[0539] The server then uses natural language processing algorithms to compare the current meeting's content with past meeting minutes. These algorithms use models such as BERT and GPT-3. These algorithms tokenize and preprocess the text data before feeding it into a comparison model. The algorithm generates a similarity score for each utterance and detects variations or inconsistencies in the content.

[0540] The server converts the detected differences into graphs and charts and displays them in a visually easy-to-understand format. Libraries such as Matplotlib and D3.js are used for this visualization. For example, it is possible to display the differences in the content of comments over time as a bar graph.

[0541] The visualized data is sent from the server to the user via email or a notification system. Emails are sent using the SMTP protocol, and notifications are sent using the Slack API, etc. Users can review the data and ask their superiors or managers why the comments were changed.

[0542] Finally, the user enters their feedback into a form on the system and submits it. The server receives this feedback and stores it in a database, for example using the SQL query "INSERT INTO feedback (meeting_id, feedback_text) VALUES (XX, 'YY')".

[0543] As a concrete example, let's consider a case where a boss says in Meeting A, "We'll start a new project next week," and then in the next Meeting B, "The new project will start the week after next." In this case, the following behavior will occur.

[0544] 1. Recording and Transcription:

[0545] The device records what is said in Meeting B and converts it into text in real time using Google Cloud Speech-to-Text.

[0546] This text data is sent to the server and temporarily stored.

[0547] 2. Comparing minutes:

[0548] The server retrieves the minutes data for Meeting A from the MySQL database.

[0549] Use the BERT model to compare what was said in Meeting A and Meeting B.

[0550] 3. Extract and visualize the differences:

[0551] The server detects the change in the content of the message from "next week" to "the week after next."

[0552] The differences are graphed using Matplotlib and notified to the user.

[0553] 4. Getting feedback:

[0554] The user checks the notification and asks their boss why the comment has changed.

[0555] The feedback obtained is entered into the system, and the server stores it in a MySQL database.

[0556] This system makes it easy to record meeting content in real time and compare it with past parliamentary sessions, and allows for visual confirmation of consistency and fluctuations in speech content, enabling efficient and transparent meeting management.

[0557] Example prompt sentence:

[0558] Example: "Write a Python script to create a system that transcribes what is said during meetings in real time and compares it with past transcripts to visualize fluctuations in what is said."

[0559] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0560] Step 1:

[0561] Start audio recording:

[0562] When the conference starts, the user starts recording audio on the terminal.

[0563] Specific behavior: The user launches a recording application on their laptop or smartphone and clicks the "Start Recording" button.

[0564] Input: User's voice.

[0565] Output: Recorded audio data.

[0566] Step 2:

[0567] Real-time transcription of audio data:

[0568] The device splits the audio data being recorded and sends it to the voice recognition engine.

[0569] Specific operation: The terminal program captures voice data in a certain buffer size, sends the data to, for example, the Google Cloud Speech-to-Text API, and receives the text data obtained from the API.

[0570] Input: Recorded audio data.

[0571] Output: Real-time converted text data.

[0572] Step 3:

[0573] Sending text data to the server and temporarily saving it:

[0574] The terminal sends the text data returned from the voice recognition engine to the server.

[0575] Specific operation: The terminal sends an HTTP POST request to the server, and the server temporarily stores the received data in a MySQL database.

[0576] Input: Real-time text data.

[0577] Output: Text data temporarily stored in the server database.

[0578] Step 4:

[0579] Obtaining past minutes data:

[0580] The server accesses a past minutes database to retrieve relevant past meeting minutes.

[0581] Specific operation: The server executes the query "SELECT FROM meeting_minutes WHERE meeting_id = 'XX'" to retrieve the data.

[0582] Input: Past meeting ID or date and time.

[0583] Output: The retrieved past minutes data.

[0584] Step 5:

[0585] Comparison of current conference speeches with past minutes:

[0586] The server uses a natural language processing algorithm to compare the acquired minutes data with the current meeting text data.

[0587] Specific operation: Using natural language processing models such as BERT or GPT-3, the text data is tokenized and a similarity score is calculated.

[0588] Input: Current meeting text data and past meeting minutes data.

[0589] Output: Comparison results showing variability and inconsistencies in what was said.

[0590] Step 6:

[0591] Visualizing the differences:

[0592] The server visualizes the differences in the detected statements in graphs and charts.

[0593] Specific operation: Using Matplotlib and D3.js, fluctuations are displayed using bar graphs and heat maps along the time axis.

[0594] Input: Difference data from comparison.

[0595] Output: Visually displayed difference graphs and charts.

[0596] Step 7:

[0597] Sending visualization data:

[0598] The server notifies the user of the visualized data.

[0599] Specific behavior: Sends an email using the SMTP protocol or a notification using the Slack API.

[0600] Input: Visualization data.

[0601] Output: Notification to the user (email or notification message).

[0602] Step 8:

[0603] Get and store feedback:

[0604] The user checks the notification and enters feedback into the system if necessary, and the server stores the feedback.

[0605] What happens: A user enters feedback into a form on the system and clicks the "Submit" button. The server receives the data and stores it in a MySQL database.

[0606] Input: User feedback.

[0607] Output: Feedback data stored in a database.

[0608] (Application example 1)

[0609] 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."

[0610] With conventional meeting management systems, it was difficult to record what was said during meetings in real time and compare it with past minutes to maintain consistency. Furthermore, in factories and other workplaces, it was difficult to grasp the consistency and fluctuations of work instructions, which led to a decline in work efficiency. To solve these problems, a real-time work instruction monitoring system was required.

[0611] 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.

[0612] In this invention, the server includes means for transcribing statements made during a meeting in real time, means for saving minutes of past meetings, means for comparing transcription data of current meeting statements with past minutes data, means for extracting variations and inconsistencies in the content of statements, means for visualizing the extracted differences as graphs or charts, means for notifying users, means for collecting feedback from users, means for saving the collected feedback data, means for acquiring work instructions through real-time speech recognition, means for comparing with a database of past work instructions, means for detecting and visualizing differences, means for displaying the detected differences on a display, and means for sending user feedback to a manager. This makes it possible to monitor the consistency of statements and work instructions in real time and improve transparency.

[0613] A "meeting" is a gathering of participants to discuss and make decisions about a specific topic.

[0614] "Transcription" is the process of converting audio into text, either in real time or at a later time.

[0615] "Minutes" are documents that record the contents of a meeting and the decisions made.

[0616] "Minutes data" refers to the contents of past minutes saved as digital data.

[0617] A "natural language processing algorithm" is a technology that allows computers to analyze and process human language.

[0618] "Graphs and charts" are diagrams and tables that visually represent data.

[0619] A "notification" is a message or alert that informs the user of information.

[0620] "Feedback" refers to opinions and evaluation information provided by users.

[0621] "Speech recognition" is a technology that converts speech into text.

[0622] "Work instructions" are information for instructing a specific work.

[0623] "Work Order Database" means a digital database for storing past and current work orders.

[0624] "Difference" refers to the difference between the objects being compared.

[0625] A "display" is a device for visually displaying information.

[0626] An "administrator" is a person in charge of running a system or project.

[0627] This invention is a system that transcribes statements made during meetings in real time and compares them with past minutes. Furthermore, this technology is applied to factory robots to provide a real-time work instruction monitoring system that monitors the consistency of work instructions.

[0628] 1. Hardware and software configuration:

[0629] Hardware: Microphones and displays built into smartphones, smart glasses, head-mounted displays, or robots.

[0630] Software: We use the "speech_recognition" library for speech recognition, the "spacy" library for natural language processing, the "matplotlib" library for data visualization, and the "json" library for data storage and processing.

[0631] 2. Data processing and calculation processing:

[0632] Speech recognition: The device uses a microphone to collect speech and instructions in real time during meetings or factory work, then converts the speech data into text using a speech recognition engine (for example, the "speech_recognition" library).

[0633] Data transmission and storage: The generated text data is transmitted to the server in real time and temporarily stored as provisional minutes or work instructions.

[0634] Comparison and analysis: The server accesses a database of past meeting minutes or work instructions and compares them with the current text data. This is done using natural language processing algorithms (e.g., the "spacy" library).

[0635] Difference detection and visualization: Detect changes in what is said or what is directed, and visualize the differences as graphs or charts (for example, graphs created with the “matplotlib” library).

[0636] Notification and feedback: The visualized data is displayed to the user on the screen and feedback is collected through a notification system if necessary. The feedback is sent back to the server and stored in a database.

[0637] 3. Example:

[0638] If a boss says in Meeting A, "We'll start a new project next week," and then in the next Meeting B, the boss changes his mind and says, "The new project will start the week after next," the system will detect the change in real time and notify the user.

[0639] On a factory floor, if a robot receives the work instruction "Please start adjusting the mechanical parts on production line 1," and then the instruction is changed to "Please start adjusting the mechanical parts on production line 3," the robot will similarly detect the change and display it on the screen.

[0640] Example prompt sentence:

[0641] Take current work orders and compare them with past work orders to detect changes.

[0642] This structure allows for greater transparency and consistency in meetings and factory operations.

[0643] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0644] Step 1:

[0645] The device uses a microphone to capture the voice of the person speaking or working on the device, and this voice data is collected in real time and used as input data for subsequent processing.

[0646] Step 2:

[0647] The device sends the acquired voice data to a voice recognition engine (for example, the "speech_recognition" library) and converts it into text data. This process outputs the voice data as transcribed text.

[0648] Step 3:

[0649] The device sends the generated text data to the server, which temporarily stores it and registers it as the current minutes or work instructions.

[0650] Step 4:

[0651] The server retrieves corresponding data from a database of past meeting minutes or work instructions. This past data is text data and is used for comparison with the current data.

[0652] Step 5:

[0653] The server uses natural language processing algorithms (e.g., the "spacy" library) to compare the current text data with past data, detecting variations or inconsistencies in the data and reporting the differences.

[0654] Step 6:

[0655] The server generates visualization data such as graphs and charts based on the detected differences. In this process, the data is processed to visually represent the degree and content of the differences, and the results are output as visualization data.

[0656] Step 7:

[0657] The server sends the visualization data to the terminal, which displays it on the display, allowing the user to check the differences.

[0658] Step 8:

[0659] The user can then input feedback about any discrepancies they find, which is then sent to the server and used for future analysis.

[0660] Step 9:

[0661] The server stores the collected feedback data in a database, which stores the feedback for future analysis and system improvement.

[0662] 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.

[0663] This invention combines an emotion engine with a system that transcribes statements made during meetings in real time and compares them with past minutes. This system not only visualizes inconsistencies in the content of statements made during meetings and monitors the consistency of statements, but also recognizes and analyzes the emotions of users and speakers.

[0664] First, when a meeting begins, the user starts recording audio on their device. The device sends what is said in the meeting to a speech recognition engine in real time, which converts it into text. This text data is then sent directly to the server in real time and temporarily saved as provisional minutes.

[0665] The server accesses a database of past meeting minutes and compares the current meeting's content with past meeting minutes using natural language processing algorithms that detect variations or inconsistencies in the content of the conversation.

[0666] Furthermore, the server uses an emotion engine to analyze the emotions of users and speakers in real time. The emotion engine recognizes emotions based on tone of voice, content of speech, and facial expression data (if a camera is used), and stores the results as analysis data.

[0667] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts, allowing users to see not only the changes in the content of comments but also the emotions expressed at the time of the comments at a glance. The server then sends this visualized data to users via email or a notification system.

[0668] Users can review the data they receive and, if necessary, ask their superiors or managers why the content of their comments was changed. The users enter their feedback into the system, which the server stores. This feedback data is used to analyze and improve future meetings.

[0669] Specific examples

[0670] For example, suppose that during Meeting A, a boss says, "We'll start a new project next week." Then during the next Meeting B, the same boss says, "We'll start a new project the week after next." This system works as follows:

[0671] 1. Recording and Transcription:

[0672] The device records what is said in Conference B and converts it into text in real time.

[0673] The generated text data is sent to the server.

[0674] 2. Comparing minutes:

[0675] The server retrieves the minutes data for Conference A from the database.

[0676] Using a natural language processing algorithm, the content of statements made in Meeting A and Meeting B is compared.

[0677] 3. Extract and visualize the differences:

[0678] The server detects the change in the content of the message and discovers the change from "next week" to "the week after next."

[0679] The differences are visualized as a graph and sent to the user.

[0680] 4. Emotion analysis:

[0681] The emotion engine analyzes the emotions expressed by the boss when he or she speaks and identifies emotions such as "anxiety."

[0682] Emotional data is also visualized as a graph and sent to the user.

[0683] 5. Getting feedback:

[0684] The user checks the notification and asks their boss why the comment has changed.

[0685] The obtained feedback is entered into the system and the server stores it in a database.

[0686] In this way, the system not only records what is being said in real time and compares it with past comments, but also analyzes emotions to provide deeper insights, allowing meeting participants to monitor both the consistency of what is being said and the emotional fluctuations, improving the transparency and efficiency of meetings.

[0687] The processing flow will be explained below.

[0688] Step 1:

[0689] Start audio recording

[0690] The terminal activates the audio recording function at the start of the conference.

[0691] The device records what is said in the meeting in real time via a microphone.

[0692] Step 2:

[0693] Real-time audio transcription

[0694] The device sends the recorded voice data to the voice recognition engine in real time.

[0695] The device receives the text data returned from the voice recognition engine and converts the spoken content into text.

[0696] Step 3:

[0697] Generate and save provisional minutes

[0698] The device transmits the text of the speech in real time to the server.

[0699] The server temporarily stores the received text data as provisional minutes.

[0700] Step 4:

[0701] Obtaining past meeting minutes data

[0702] The server retrieves past meeting minutes data from the database.

[0703] The server prepares the acquired minutes data for comparison with today's minutes data.

[0704] Step 5:

[0705] Comparison of meeting minutes data

[0706] The server uses natural language processing (NLP) algorithms to compare today's minutes with past minutes.

[0707] The server detects changes in keywords and phrases in the content of comments.

[0708] Step 6:

[0709] Emotion analysis

[0710] The server uses an emotion engine to analyze the emotions of speakers in real time during a conference.

[0711] The device sends the speaker's tone of voice and facial expression data (if necessary) to the emotion engine.

[0712] Step 7:

[0713] Storing Emotional Data

[0714] The server receives the analyzed emotion data and stores it together with the provisional minutes data.

[0715] Step 8:

[0716] Extracting Differences

[0717] The server identifies fluctuations and inconsistencies in what is being said.

[0718] The server builds a list of the identified differences and aggregates the data.

[0719] Step 9:

[0720] Visualization of differential data and emotion data

[0721] The server generates visual representations such as graphs and charts based on the extracted differences and emotion data.

[0722] The server formats the visualization data into a format that is easy for the user to understand.

[0723] Step 10:

[0724] User Notification

[0725] The server adds explanatory text to the visualized data and sends it to the specified user as an email or internal notification.

[0726] The user checks the received notification and understands changes in the content of the comments and the results of sentiment analysis.

[0727] Step 11:

[0728] Get feedback

[0729] The user checks the notification and, if necessary, asks their superior or manager about the reason for the change in the comment content.

[0730] The user enters this feedback into the system.

[0731] Step 12:

[0732] Feedback Data Storage

[0733] The server receives the feedback data sent by the user and stores it in a database.

[0734] The server will prepare to use the data for future analysis and improvements.

[0735] These steps allow the system to efficiently record what is said during a meeting, compare it with past comments, and analyze and visualize the emotions of the speakers. Based on this information, users can identify the reasons for fluctuations in comments and take action to improve the quality of the meeting.

[0736] Example 2

[0737] 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."

[0738] Conventional meeting minutes systems lack the functionality to transcribe speeches made during meetings in real time and compare them with past minutes. It is also difficult to visualize fluctuations or inconsistencies in the speech content of meeting participants. Furthermore, there is no way to analyze the emotions of speakers and gain a deeper understanding of the meeting's progress, so the transparency and efficiency of meetings are not sufficiently ensured. To solve these issues, a system is needed that not only transcribes speeches in real time, but also compares speech content and analyzes emotions, and visualizes the results for users.

[0739] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for transcribing statements made during a meeting in real time, means for saving past meeting minutes, means for comparing transcription data of current meeting statements with past meeting minutes data, means for analyzing the emotions of users and speakers, and means for visualizing the extracted differences and emotion analysis data as graphs or charts. This makes it possible to improve the transparency and efficiency of meetings by comparing the consistency of statements made during a meeting with the contents of past statements, analyzing the emotions of speakers, and visually providing this to the user.

[0740] "A means for transcribing statements made during a meeting in real time" is a system that records audio during a meeting in real time and instantly converts it into text using voice recognition technology.

[0741] The "means for storing minutes of past meetings" is a system that stores the contents of statements made in previous meetings as text data in a database or the like.

[0742] "Means for comparing transcription data of current meeting remarks with past minutes data" is a function that uses natural language processing technology to compare the content of remarks made in the most recent meeting with past meeting records, and detects fluctuations or lack of consistency in the content.

[0743] "Means to extract variations and lack of consistency in speech content" is a function that identifies how speech made during a meeting differs from past speech and whether consistency is maintained, and clearly shows the differences.

[0744] "Means for analyzing the emotions of users and speakers" refers to technology that analyzes and identifies the emotional state of speakers during a meeting based on their voice, text, and, if necessary, facial expression data.

[0745] "Means for visualizing extracted differences and sentiment analysis data as graphs or charts" refers to a system that converts fluctuations in speech content and sentiment analysis results into visual data, i.e., visual formats such as graphs and charts, and displays them.

[0746] "Means for notifying users" refers to a function for delivering analysis results and visualization data to users via email or a notification system.

[0747] "Means for collecting feedback from users" is a function that collects opinions and responses that users input to the system and saves them as data.

[0748] The "means for storing collected feedback data" is a function for storing the feedback information collected from users in a database or the like and managing it for use in later analysis and improvement.

[0749] This invention combines an emotion engine with a system that transcribes statements made during meetings in real time and compares them with past minutes. The system aims to improve the transparency and efficiency of meetings by visualizing fluctuations and inconsistencies in the content of statements and analyzing the emotions of speakers.

[0750] First, when a meeting begins, the user starts recording audio on their device. The device then uses a speech recognition engine such as the Google Cloud Speech-to-Text API to convert what is being said into text in real time. This text data is then sent to a server in real time and saved as temporary meeting minutes.

[0751] The server accesses a database of past meeting minutes and compares the current meeting's content with past minutes using natural language processing techniques such as BERT and GPT. This comparison detects fluctuations and inconsistencies in the content of the speech.

[0752] Furthermore, the server uses an emotion analysis engine such as IBM Watson Tone Analyzer to analyze the emotions of users and speakers in real time. This analysis involves recognizing emotions based on voice tone, speech content, and, in some cases, facial expression data captured by a camera, and saving the analysis data.

[0753] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts, allowing users to see not only the changes in the content of comments but also the emotions expressed at the time of the comments at a glance. The server then sends this visualized data to users via email or a notification system.

[0754] The user can review the data and, if necessary, ask their superiors or managers why the content was changed. The user then enters the feedback they received into the system, and the server stores this feedback data. This feedback data can be used to analyze and improve future meetings.

[0755] Specific examples

[0756] Examples of Meeting A and Meeting B

[0757] For example, if a boss says "We'll start a new project next week" during Meeting A, and then the same boss says "We'll start a new project the week after next" during a subsequent Meeting B, the system will operate as follows:

[0758] 1. Recording and Transcription:

[0759] The device records what is said in Conference B and converts it into text in real time. This text data is then sent to the server.

[0760] 2. Comparing minutes:

[0761] The server retrieves the minutes of Meeting A from the database and uses natural language processing technology to compare the content of statements made in Meeting A and Meeting B.

[0762] 3. Extract and visualize the differences:

[0763] The server detects changes in the content of comments and detects changes from "next week" to "the week after next." This is visualized as a graph and sent to the user.

[0764] 4. Emotion analysis:

[0765] The emotion engine analyzes the emotion expressed at the time of speech and identifies emotions such as "anxiety." This is also visualized as a graph and sent to the user.

[0766] 5. Getting feedback:

[0767] The user checks the notification and asks their superior why the comment has changed. The feedback is then entered into the system, and the server stores it in a database.

[0768] Prompt Sentence Examples

[0769] An example of a prompt to input to a generative AI model is as follows:

[0770] "Discuss a system that transcribes speech during a meeting in real time and compares it with past minutes. In this system, the user starts recording, and the data is converted to text through a speech recognition engine and sent to a server. The server compares the speech using a natural language processing algorithm and analyzes the speaker's sentiment using an emotion engine. Finally, the system visualizes the differences and the sentiment analysis results and notifies the user."

[0771] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0772] Step 1: Start recording and real-time transcription

[0773] When a meeting begins, users start recording audio on their device, which then uses a speech recognition engine (Google Cloud Speech-to-Text API) to convert what is being said into text in real time.

[0774] Input: Meeting audio

[0775] Data processing: The voice recognition engine converts the voice data into text data.

[0776] Output: Text data generated in real time

[0777] Specific operation: The user presses the "Start Recording" button in the application. The device collects audio data through the microphone and converts the speech into text using the Google Cloud Speech-to-Text API.

[0778] Step 2: Sending real-time data to the server

[0779] The device sends the generated text data in real time to the server, which temporarily stores the data as provisional minutes.

[0780] Input: Real-time generated text data

[0781] Data processing: Send text data to the server via the network

[0782] Output: Text data saved as provisional minutes

[0783] Specific operation: The terminal sends text data to the server at regular intervals, and the server temporarily stores the received data.

[0784] Step 3: Obtain and compare past minutes

[0785] The server accesses the database and compares the current meeting content with past meeting minutes using natural language processing techniques (such as BERT and GPT).

[0786] Input: Current meeting text data and past meeting minutes data

[0787] Data processing: Analyzing text data using natural language processing techniques to detect variations and inconsistencies in spoken content

[0788] Output: Data showing fluctuations and inconsistencies in speech

[0789] How it works: The server extracts keywords from the stored text data, queries the database to retrieve relevant past meeting transcripts, and then applies natural language processing techniques such as the BERT model to identify variations and inconsistencies in the speech.

[0790] Step 4: Perform sentiment analysis

[0791] The server uses an emotion engine (such as IBM Watson Tone Analyzer) to analyze the emotions of users and speakers in real time.

[0792] Input: Meeting speech text and voice data, and facial expression data if necessary

[0793] Data processing: Using an emotion engine to analyze the speaker's emotions from voice and text

[0794] Output: Parsed emotion data

[0795] Specific operation: The server inputs the acquired voice, text, and facial expression data into the emotion engine, and then quantifies and stores the analyzed emotional state.

[0796] Step 5: Visualize the difference and sentiment analysis data

[0797] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts.

[0798] Input: Data on fluctuations in speech content and sentiment analysis data

[0799] Data processing: Generate graphs and charts using visualization tools (e.g., D3.js)

[0800] Output: Visual data (graphs and charts)

[0801] Specific operation: The server uses visualization tools to generate graphs and charts that highlight fluctuations and emotional changes.

[0802] Step 6: Communicate data and get feedback

[0803] The server sends the generated visualization data to the user via email or a notification system. The user checks the notification and, if necessary, asks their superior or manager for the reason for the change in the comment content.

[0804] Input: Visual data

[0805] Data processing: Distributing data via email and notification systems

[0806] Output: Visual data provided to the user

[0807] Specific operation: The server generates the visual data in PDF or web link format and sends a notification to the user, who receives the notification and asks his / her boss for feedback.

[0808] Step 7: Save the feedback and use it in your next meeting

[0809] Users input feedback from their superiors and managers into the system, and the server stores this feedback data and uses it to analyze and improve the next meeting.

[0810] Input: User feedback information

[0811] Data processing: Save the feedback data to a database

[0812] Output: Saved feedback data

[0813] Specific operation: The user enters comments and reasons on the feedback input screen, and the server stores them in the database. The feedback data is then used as a reference for the next analysis.

[0814] (Application example 2)

[0815] 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."

[0816] Conventional autonomous vehicles lack the means to monitor driver and passenger conversations in real time and improve safety based on that information. They also lack the ability to detect risks such as stress and lack of attention while driving in advance. Furthermore, there is a lack of comparative analysis with past driving records, making it difficult to respond appropriately to consistency and fluctuations.

[0817] The identification process by the identification 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 means for transcribing conversation content in real time, means for saving past conversation records, means for comparing transcription data of the current conversation content with past conversation record data, means for extracting fluctuations and inconsistencies in the conversation content, means for visualizing the extracted differences and emotion analysis data as graphs or charts, means for notifying the user, means for collecting feedback from the user, and means for saving the collected feedback data. This makes it possible to monitor conversation content and emotional fluctuations while driving in real time, thereby improving safety.

[0818] A "means for transcribing conversation content in real time" is a means that has the function of converting conversation from voice to text in real time.

[0819] The "means for saving records of past conversations" refers to a means having a function for saving text data of conversations recorded in the past.

[0820] "Means for comparing transcription data of current conversation content with past conversation record data" refers to a means that has the function of comparing data currently transcribed in real time with previously saved conversation record data to verify changes and consistency in content.

[0821] The "means for extracting fluctuations and inconsistencies in conversation content" refers to a means having a function for detecting and extracting changes and inconsistencies in conversation content.

[0822] "Means for visualizing extracted differences and emotion analysis data as graphs or charts" refers to means that has the function of generating graphs or charts to visually represent detected differences and analyzed emotion data.

[0823] The "means for notifying the user" refers to a means having a function for notifying the user of detected information and analysis results.

[0824] The "means for collecting feedback from users" refers to means that have the function of collecting opinions and information provided by users.

[0825] The "means for storing collected feedback data" refers to a means having a function for storing collected user feedback data.

[0826] "Audio recording means" means a means capable of recording conversations within a vehicle and storing the recording for later use.

[0827] A "means for analyzing emotions" is a means that has the function of identifying and analyzing emotions from the content of a conversation and tone of voice.

[0828] "Means for comparing conversation content using natural language processing algorithms" means means that use natural language processing technology to compare current conversation content with past records and verify consistency and variation.

[0829] This invention combines a sentiment analysis engine with a system that transcribes conversation content in real time, compares it with past conversation records, and monitors conversation fluctuations and utterance consistency. To realize this system, the following means are required.

[0830] First, the device records conversations in the car in real time and transcribes them. The device captures audio using a built-in microphone and converts this audio into text data in real time. This process uses voice recognition technology using the speech_recognition library.

[0831] The resulting text data is immediately sent to a server and temporarily stored as a provisional conversation record. The server then accesses a database of past conversation records and compares the current conversation with the past records using a natural language processing algorithm (NLP). This comparison method uses Hugging Face's natural language processing model.

[0832] The server then uses an emotion analysis engine to analyze emotions from the content of the conversation and the tone of voice. Hugging Face's emotion analysis model identifies emotions such as "stress," "impatience," and "fatigue" based on the text data obtained. This allows the server to understand the emotional state of the driver and passengers.

[0833] Furthermore, the server visualizes the comparison results and sentiment analysis data by using the matplotlib library to generate graphs and charts of the detected differences and sentiment data, converting them into a visually easy-to-understand format.

[0834] This visualized data is notified to the user (driver or in-car supervisor) in real time. For example, if the driver repeatedly says "I'm tired," the system analyzes in real time and issues a warning when it detects the emotion of "fatigue." The user can check this notification and take appropriate action if necessary.

[0835] The system also includes a means for users to provide feedback. Users receive notifications from the system and can enter their opinions or additional information into the system. The collected feedback data is stored in a database by the server and used for future conversation analysis.

[0836] As a specific example, "If a driver on their way to the office repeatedly says 'I'm tired,' the system will use its sentiment analysis engine to determine that the driver is 'fatigued' and issue a real-time warning." This prompt would be written as follows:

[0837] "If a driver on the way to the office repeatedly says 'I'm tired,' the system should use its sentiment analysis engine to determine that the driver is 'fatigued,' and devise a program to issue a real-time warning."

[0838] As described above, this invention not only records conversations in real time and compares them with past records, but also analyzes emotions, thereby improving safety and efficiency while driving.

[0839] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0840] Step 1:

[0841] The device uses the in-car microphone to record audio in real time. The input is in-car audio data, and the output is a recorded audio file. Specifically, the device's microphone picks up the conversations between the driver and passengers and records them as audio data.

[0842] Step 2:

[0843] The device transcribes recorded audio in real time. The input is an audio file, and the output is transcribed text data. The device uses the speech_recognition library to analyze the audio and convert it to text.

[0844] Step 3:

[0845] The transcribed text data is sent to a server and temporarily saved as a provisional conversation record. The input is the transcribed data, and the output is the provisional text record data on the server. The device sends the data to the server via network communication.

[0846] Step 4:

[0847] The server accesses a database of past conversation records and compares the current conversation content with past records using a natural language processing algorithm. The input is the current conversation text data and the past conversation record data, and the output is the comparison result data. Specifically, the server uses Hugging Face's NLP model to detect differences and inconsistencies in the text.

[0848] Step 5:

[0849] The server uses an emotion analysis engine to analyze emotions from the content of the conversation and tone of voice. The input is transcribed text data, and the output is analyzed emotional data. The server uses Hugging Face's emotion analysis model to identify emotions such as "stress," "anxiety," and "fatigue."

[0850] Step 6:

[0851] The server visualizes the detected differences and sentiment analysis data as graphs or charts. The input is the comparison results and sentiment data, and the output is a visual graph or chart. The server uses the matplotlib library to visually represent the detected variations and sentiment.

[0852] Step 7:

[0853] The server generates visualization data and notifies the user in real time. The input is the visualized data, and the output is a notification message to the user. For example, if the driver repeatedly says "I'm tired," the server generates a warning message and sends it to the user.

[0854] Step 8:

[0855] The user checks the notification from the system and provides feedback. The input is the user's feedback, and the output is the feedback data. The user checks the notification content and enters their opinions and information corresponding to that content into the system.

[0856] Step 9:

[0857] The server stores the collected feedback data and uses it for future conversation analysis. The input is the feedback data and the output is the stored feedback database. The server stores these data in a database and makes them available for subsequent dialogue analysis.

[0858] The above are the processing steps for realizing the invention.

[0859] 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.

[0860] 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.

[0861] 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.

[0862] [Third embodiment]

[0863] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0864] 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.

[0865] 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).

[0866] 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.

[0867] 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.

[0868] 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).

[0869] 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.

[0870] 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.

[0871] 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.

[0872] 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.

[0873] 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.

[0874] 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."

[0875] This invention relates to a system that transcribes statements made during meetings in real time and compares them with past minutes. This system visualizes inconsistencies in the content of statements made during meetings and monitors the consistency of statements.

[0876] First, when a meeting begins, the user starts recording audio on their device. The device sends what is said in the meeting to a speech recognition engine in real time, which converts it into text. This text data is then sent directly to the server in real time and temporarily saved as provisional minutes.

[0877] The server then accesses a database of past meeting minutes and compares the current meeting's content with past meeting minutes using natural language processing algorithms that detect variations or inconsistencies in the content.

[0878] The server converts the detected differences into visual formats such as graphs and charts, allowing users to see the changes in the content at a glance. The server then sends this visualization data to users via email or a notification system.

[0879] Users can review the data they receive and, if necessary, ask their superiors or managers why the content of their comments was changed. The users enter their feedback into the system, which the server stores. This feedback data is used to analyze and improve future meetings.

[0880] Specific examples

[0881] For example, suppose that during Meeting A, a boss says, "We'll start a new project next week." Then during the next Meeting B, the same boss says, "We'll start a new project the week after next." This system works as follows:

[0882] 1. Recording and Transcription:

[0883] The device records what is said in Conference B and converts it into text in real time.

[0884] The generated text data is sent to the server.

[0885] 2. Comparing minutes:

[0886] The server retrieves the minutes data for Conference A from the database.

[0887] Using a natural language processing algorithm, the content of statements made in Meeting A and Meeting B is compared.

[0888] 3. Extract and visualize the differences:

[0889] The server detects the change in the content of the message and discovers the change from "next week" to "the week after next."

[0890] The differences are visualized as a graph and sent to the user.

[0891] 4. Getting feedback:

[0892] The user checks the notification and asks their boss why the comment has changed.

[0893] The obtained feedback is entered into the system and the server stores it in a database.

[0894] In this way, this system records what is being said in real time and compares it with past comments to clarify inconsistencies in what is being said, improving the transparency and efficiency of meetings.

[0895] The processing flow will be explained below.

[0896] Step 1:

[0897] Start audio recording

[0898] The terminal activates the audio recording function at the start of the conference.

[0899] The device records what is said in the meeting in real time via a microphone.

[0900] Step 2:

[0901] Real-time audio transcription

[0902] The device sends the recorded voice data to the voice recognition engine in real time.

[0903] The device receives the text data returned from the voice recognition engine and converts the spoken content into text.

[0904] Step 3:

[0905] Sending text data and generating provisional minutes

[0906] The device transmits the text of the speech in real time to the server.

[0907] The server temporarily stores the received text data as provisional minutes.

[0908] Step 4:

[0909] Obtaining past meeting minutes data

[0910] The server retrieves past meeting minutes data from the database.

[0911] The server prepares the acquired minutes data for comparison with today's minutes data.

[0912] Step 5:

[0913] Comparison of meeting minutes data

[0914] The server uses natural language processing (NLP) algorithms to compare today's minutes with past minutes.

[0915] The server detects changes in keywords and phrases in the content of comments.

[0916] Step 6:

[0917] Extracting Differences

[0918] The server identifies fluctuations and inconsistencies in what is being said.

[0919] The server builds a list of the identified differences and aggregates the data.

[0920] Step 7:

[0921] Visualizing differential data

[0922] The server generates visual representations such as graphs and charts based on the extracted differences.

[0923] The server formats the visualization data into a format that is easy for the user to understand.

[0924] Step 8:

[0925] User Notification

[0926] The server adds explanatory text to the visualized data and sends it to the specified user as an email or internal notification.

[0927] The user checks the received notification and understands changes in the content of the comments.

[0928] Step 9:

[0929] Get feedback

[0930] The user checks the notification and, if necessary, asks their superior or manager about the reason for the change in the comment content.

[0931] The user enters this feedback into the system.

[0932] Step 10:

[0933] Feedback Data Storage

[0934] The server receives the feedback data sent by the user and stores it in a database.

[0935] The server will prepare to use the data for future analysis and improvements.

[0936] These steps allow the system to efficiently record what is said during a meeting, compare it with past comments to identify inconsistencies and changes, and report them back to meeting participants, allowing users to understand the reasons for fluctuations and take action to improve the quality of the meeting.

[0937] Example 1

[0938] 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."

[0939] Previously, it was difficult to record statements made during meetings in real time and understand the consistency and changes in the content of those statements. As a result, important decision-making and information transmission in meetings became uncertain, making efficient business operations difficult. Furthermore, comparing past minutes with current statements was time-consuming and labor-intensive. This made ensuring the transparency and reliability of meeting content a challenge.

[0940] 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.

[0941] In this invention, the server includes: means for transcribing statements made during a meeting in real time; means for saving past meeting minutes; means for comparing the transcription data of current meeting statements with past meeting minutes data; means for extracting variations and inconsistencies in the content of statements; means for visualizing the extracted differences as graphs or charts; means for notifying users; means for collecting feedback from users; means for saving the collected feedback data; means for transmitting voice data to a voice recognition engine in real time and acquiring it as text data; and means for comparing multiple meeting statements and analyzing the consistency of the statement content with high accuracy using a natural language processing algorithm. This automates the real-time recording and comparison of meeting content and makes it possible to visually confirm the consistency and variations in the content of statements, enabling efficient and transparent meeting management.

[0942] "Means for transcribing statements made during a meeting in real time" refers to a function or device that instantly converts speech made during a meeting into text data using voice recognition technology.

[0943] "Means for storing minutes of past meetings" refers to a function or device that stores the contents of past meetings as text data in a database or the like.

[0944] "Means for comparing the current meeting transcript with past meeting minutes" refers to algorithms or software that compare the transcript generated from the most recent meeting with similarly stored meeting minutes from the past.

[0945] "Means for detecting variations or inconsistencies in statements" refers to algorithms or software that identify changes or inconsistencies in statements from the results of comparisons.

[0946] "Means for visualizing extracted differences as graphs or charts" refers to a function or device for displaying the difference information extracted through comparison in a visually easy-to-understand format.

[0947] "Means for notifying users" refers to a function or device for sending information to users via text message, email, or other means of communication.

[0948] "Means for collecting feedback from users" refers to a function or device that allows users to input comments and opinions into the system and collects that information.

[0949] "Means for storing collected feedback data" refers to a database or storage system for storing feedback information obtained from users.

[0950] "Means for sending voice data to a voice recognition engine in real time and obtaining it as text data" refers to a function or device that sends voice during a meeting to a voice recognition system (API, etc.) in real time and receives the resulting text data.

[0951] "Means of using natural language processing algorithms to compare the content of multiple meetings and analyze the consistency of the content with high accuracy" refers to algorithms and software that use natural language processing technology to perform detailed analysis of the content of statements in different meeting records and evaluate their consistency.

[0952] The present invention relates to a system for transcribing statements made during a meeting in real time and comparing the results with past minutes. Specific embodiments of the system are described below.

[0953] First, when a user starts a meeting, they begin recording audio on their device. The device can be a smartphone or laptop, and recording is done using the device's built-in microphone. The recorded audio data is sent directly to a speech recognition engine in real time. Specifically, Google Cloud Speech-to-Text API or IBM Watson Speech to Text can be used. The device divides the audio data into buffers of a certain size, makes an API call, and converts the audio into text data.

[0954] The generated text data is then sent from the device to the server. The server receives it and temporarily stores it in a database as provisional minutes. Databases such as MySQL or PostgreSQL are often used for this purpose. The server accesses the database of past meeting minutes and retrieves past minutes related to the current meeting. For example, it extracts the data by executing the SQL query "SELECT FROM meeting_minutes WHERE meeting_id = 'XX'".

[0955] The server then uses natural language processing algorithms to compare the current meeting's content with past meeting minutes. These algorithms use models such as BERT and GPT-3. These algorithms tokenize and preprocess the text data before feeding it into a comparison model. The algorithm generates a similarity score for each utterance and detects variations or inconsistencies in the content.

[0956] The server converts the detected differences into graphs and charts and displays them in a visually easy-to-understand format. Libraries such as Matplotlib and D3.js are used for this visualization. For example, it is possible to display the differences in the content of comments over time as a bar graph.

[0957] The visualized data is sent from the server to the user via email or a notification system. Emails are sent using the SMTP protocol, and notifications are sent using the Slack API, etc. Users can review the data and ask their superiors or managers why the comments were changed.

[0958] Finally, the user enters their feedback into a form on the system and submits it. The server receives this feedback and stores it in a database, for example using the SQL query "INSERT INTO feedback (meeting_id, feedback_text) VALUES (XX, 'YY')".

[0959] As a concrete example, let's consider a case where a boss says in Meeting A, "We'll start a new project next week," and then in the next Meeting B, "The new project will start the week after next." In this case, the following behavior will occur.

[0960] 1. Recording and Transcription:

[0961] The device records what is said in Meeting B and converts it into text in real time using Google Cloud Speech-to-Text.

[0962] This text data is sent to the server and temporarily stored.

[0963] 2. Comparing minutes:

[0964] The server retrieves the minutes data for Meeting A from the MySQL database.

[0965] Use the BERT model to compare what was said in Meeting A and Meeting B.

[0966] 3. Extract and visualize the differences:

[0967] The server detects the change in the content of the message from "next week" to "the week after next."

[0968] The differences are graphed using Matplotlib and notified to the user.

[0969] 4. Getting feedback:

[0970] The user checks the notification and asks their boss why the comment has changed.

[0971] The feedback obtained is entered into the system, and the server stores it in a MySQL database.

[0972] This system makes it easy to record meeting content in real time and compare it with past parliamentary sessions, and allows for visual confirmation of consistency and fluctuations in speech content, enabling efficient and transparent meeting management.

[0973] Example prompt sentence:

[0974] Example: "Write a Python script to create a system that transcribes what is said during meetings in real time and compares it with past transcripts to visualize fluctuations in what is said."

[0975] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0976] Step 1:

[0977] Start audio recording:

[0978] When the conference starts, the user starts recording audio on the terminal.

[0979] Specific behavior: The user launches a recording application on their laptop or smartphone and clicks the "Start Recording" button.

[0980] Input: User's voice.

[0981] Output: Recorded audio data.

[0982] Step 2:

[0983] Real-time transcription of audio data:

[0984] The device splits the audio data being recorded and sends it to the voice recognition engine.

[0985] Specific operation: The terminal program captures voice data in a certain buffer size, sends the data to, for example, the Google Cloud Speech-to-Text API, and receives the text data obtained from the API.

[0986] Input: Recorded audio data.

[0987] Output: Real-time converted text data.

[0988] Step 3:

[0989] Sending text data to the server and temporarily saving it:

[0990] The terminal sends the text data returned from the voice recognition engine to the server.

[0991] Specific operation: The terminal sends an HTTP POST request to the server, and the server temporarily stores the received data in a MySQL database.

[0992] Input: Real-time text data.

[0993] Output: Text data temporarily stored in the server database.

[0994] Step 4:

[0995] Obtaining past minutes data:

[0996] The server accesses a past minutes database to retrieve relevant past meeting minutes.

[0997] Specific operation: The server executes the query "SELECT FROM meeting_minutes WHERE meeting_id = 'XX'" to retrieve the data.

[0998] Input: Past meeting ID or date and time.

[0999] Output: The retrieved past minutes data.

[1000] Step 5:

[1001] Comparison of current conference speeches with past minutes:

[1002] The server uses a natural language processing algorithm to compare the acquired minutes data with the current meeting text data.

[1003] Specific operation: Using natural language processing models such as BERT or GPT-3, the text data is tokenized and a similarity score is calculated.

[1004] Input: Current meeting text data and past meeting minutes data.

[1005] Output: Comparison results showing variability and inconsistencies in what was said.

[1006] Step 6:

[1007] Visualizing the differences:

[1008] The server visualizes the differences in the detected statements in graphs and charts.

[1009] Specific operation: Using Matplotlib and D3.js, fluctuations are displayed using bar graphs and heat maps along the time axis.

[1010] Input: Difference data from comparison.

[1011] Output: Visually displayed difference graphs and charts.

[1012] Step 7:

[1013] Sending visualization data:

[1014] The server notifies the user of the visualized data.

[1015] Specific behavior: Sends an email using the SMTP protocol or a notification using the Slack API.

[1016] Input: Visualization data.

[1017] Output: Notification to the user (email or notification message).

[1018] Step 8:

[1019] Get and store feedback:

[1020] The user checks the notification and enters feedback into the system if necessary, and the server stores the feedback.

[1021] What happens: A user enters feedback into a form on the system and clicks the "Submit" button. The server receives the data and stores it in a MySQL database.

[1022] Input: User feedback.

[1023] Output: Feedback data stored in a database.

[1024] (Application example 1)

[1025] 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."

[1026] With conventional meeting management systems, it was difficult to record what was said during meetings in real time and compare it with past minutes to maintain consistency. Furthermore, in factories and other workplaces, it was difficult to grasp the consistency and fluctuations of work instructions, which led to a decline in work efficiency. To solve these problems, a real-time work instruction monitoring system was required.

[1027] 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.

[1028] In this invention, the server includes means for transcribing statements made during a meeting in real time, means for saving minutes of past meetings, means for comparing transcription data of current meeting statements with past minutes data, means for extracting variations and inconsistencies in the content of statements, means for visualizing the extracted differences as graphs or charts, means for notifying users, means for collecting feedback from users, means for saving the collected feedback data, means for acquiring work instructions through real-time speech recognition, means for comparing with a database of past work instructions, means for detecting and visualizing differences, means for displaying the detected differences on a display, and means for sending user feedback to a manager. This makes it possible to monitor the consistency of statements and work instructions in real time and improve transparency.

[1029] A "meeting" is a gathering of participants to discuss and make decisions about a specific topic.

[1030] "Transcription" is the process of converting audio into text, either in real time or at a later time.

[1031] "Minutes" are documents that record the contents of a meeting and the decisions made.

[1032] "Minutes data" refers to the contents of past minutes saved as digital data.

[1033] A "natural language processing algorithm" is a technology that allows computers to analyze and process human language.

[1034] "Graphs and charts" are diagrams and tables that visually represent data.

[1035] A "notification" is a message or alert that informs the user of information.

[1036] "Feedback" refers to opinions and evaluation information provided by users.

[1037] "Speech recognition" is a technology that converts speech into text.

[1038] "Work instructions" are information for instructing a specific work.

[1039] "Work Order Database" means a digital database for storing past and current work orders.

[1040] "Difference" refers to the difference between the objects being compared.

[1041] A "display" is a device for visually displaying information.

[1042] An "administrator" is a person in charge of running a system or project.

[1043] This invention is a system that transcribes statements made during meetings in real time and compares them with past minutes. Furthermore, this technology is applied to factory robots to provide a real-time work instruction monitoring system that monitors the consistency of work instructions.

[1044] 1. Hardware and software configuration:

[1045] Hardware: Microphones and displays built into smartphones, smart glasses, head-mounted displays, or robots.

[1046] Software: We use the "speech_recognition" library for speech recognition, the "spacy" library for natural language processing, the "matplotlib" library for data visualization, and the "json" library for data storage and processing.

[1047] 2. Data processing and calculation processing:

[1048] Speech recognition: The device uses a microphone to collect speech and instructions in real time during meetings or factory work, then converts the speech data into text using a speech recognition engine (for example, the "speech_recognition" library).

[1049] Data transmission and storage: The generated text data is transmitted to the server in real time and temporarily stored as provisional minutes or work instructions.

[1050] Comparison and analysis: The server accesses a database of past meeting minutes or work instructions and compares them with the current text data. This is done using natural language processing algorithms (e.g., the "spacy" library).

[1051] Difference detection and visualization: Detect changes in what is said or what is directed, and visualize the differences as graphs or charts (for example, graphs created with the “matplotlib” library).

[1052] Notification and feedback: The visualized data is displayed to the user on the screen and feedback is collected through a notification system if necessary. The feedback is sent back to the server and stored in a database.

[1053] 3. Example:

[1054] If a boss says in Meeting A, "We'll start a new project next week," and then in the next Meeting B, the boss changes his mind and says, "The new project will start the week after next," the system will detect the change in real time and notify the user.

[1055] On a factory floor, if a robot receives the work instruction "Please start adjusting the mechanical parts on production line 1," and then the instruction is changed to "Please start adjusting the mechanical parts on production line 3," the robot will similarly detect the change and display it on the screen.

[1056] Example prompt sentence:

[1057] Take current work orders and compare them with past work orders to detect changes.

[1058] This structure allows for greater transparency and consistency in meetings and factory operations.

[1059] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1060] Step 1:

[1061] The device uses a microphone to capture the voice of the person speaking or working on the device, and this voice data is collected in real time and used as input data for subsequent processing.

[1062] Step 2:

[1063] The device sends the acquired voice data to a voice recognition engine (for example, the "speech_recognition" library) and converts it into text data. This process outputs the voice data as transcribed text.

[1064] Step 3:

[1065] The device sends the generated text data to the server, which temporarily stores it and registers it as the current minutes or work instructions.

[1066] Step 4:

[1067] The server retrieves corresponding data from a database of past meeting minutes or work instructions. This past data is text data and is used for comparison with the current data.

[1068] Step 5:

[1069] The server uses natural language processing algorithms (e.g., the "spacy" library) to compare the current text data with past data, detecting variations or inconsistencies in the data and reporting the differences.

[1070] Step 6:

[1071] The server generates visualization data such as graphs and charts based on the detected differences. In this process, the data is processed to visually represent the degree and content of the differences, and the results are output as visualization data.

[1072] Step 7:

[1073] The server sends the visualization data to the terminal, which displays it on the display, allowing the user to check the differences.

[1074] Step 8:

[1075] The user can then input feedback about any discrepancies they find, which is then sent to the server and used for future analysis.

[1076] Step 9:

[1077] The server stores the collected feedback data in a database, which stores the feedback for future analysis and system improvement.

[1078] 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.

[1079] This invention combines an emotion engine with a system that transcribes statements made during meetings in real time and compares them with past minutes. This system not only visualizes inconsistencies in the content of statements made during meetings and monitors the consistency of statements, but also recognizes and analyzes the emotions of users and speakers.

[1080] First, when a meeting begins, the user starts recording audio on their device. The device sends what is said in the meeting to a speech recognition engine in real time, which converts it into text. This text data is then sent directly to the server in real time and temporarily saved as provisional minutes.

[1081] The server accesses a database of past meeting minutes and compares the current meeting's content with past meeting minutes using natural language processing algorithms that detect variations or inconsistencies in the content of the conversation.

[1082] Furthermore, the server uses an emotion engine to analyze the emotions of users and speakers in real time. The emotion engine recognizes emotions based on tone of voice, content of speech, and facial expression data (if a camera is used), and stores the results as analysis data.

[1083] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts, allowing users to see not only the changes in the content of comments but also the emotions expressed at the time of the comments at a glance. The server then sends this visualized data to users via email or a notification system.

[1084] Users can review the data they receive and, if necessary, ask their superiors or managers why the content of their comments was changed. The users enter their feedback into the system, which the server stores. This feedback data is used to analyze and improve future meetings.

[1085] Specific examples

[1086] For example, suppose that during Meeting A, a boss says, "We'll start a new project next week." Then during the next Meeting B, the same boss says, "We'll start a new project the week after next." This system works as follows:

[1087] 1. Recording and Transcription:

[1088] The device records what is said in Conference B and converts it into text in real time.

[1089] The generated text data is sent to the server.

[1090] 2. Comparing minutes:

[1091] The server retrieves the minutes data for Conference A from the database.

[1092] Using a natural language processing algorithm, the content of statements made in Meeting A and Meeting B is compared.

[1093] 3. Extract and visualize the differences:

[1094] The server detects the change in the content of the message and discovers the change from "next week" to "the week after next."

[1095] The differences are visualized as a graph and sent to the user.

[1096] 4. Emotion analysis:

[1097] The emotion engine analyzes the emotions expressed by the boss when he or she speaks and identifies emotions such as "anxiety."

[1098] Emotional data is also visualized as a graph and sent to the user.

[1099] 5. Getting feedback:

[1100] The user checks the notification and asks their boss why the comment has changed.

[1101] The obtained feedback is entered into the system and the server stores it in a database.

[1102] In this way, the system not only records what is being said in real time and compares it with past comments, but also analyzes emotions to provide deeper insights, allowing meeting participants to monitor both the consistency of what is being said and the emotional fluctuations, improving the transparency and efficiency of meetings.

[1103] The processing flow will be explained below.

[1104] Step 1:

[1105] Start audio recording

[1106] The terminal activates the audio recording function at the start of the conference.

[1107] The device records what is said in the meeting in real time via a microphone.

[1108] Step 2:

[1109] Real-time audio transcription

[1110] The device sends the recorded voice data to the voice recognition engine in real time.

[1111] The device receives the text data returned from the voice recognition engine and converts the spoken content into text.

[1112] Step 3:

[1113] Generate and save provisional minutes

[1114] The device transmits the text of the speech in real time to the server.

[1115] The server temporarily stores the received text data as provisional minutes.

[1116] Step 4:

[1117] Obtaining past meeting minutes data

[1118] The server retrieves past meeting minutes data from the database.

[1119] The server prepares the acquired minutes data for comparison with today's minutes data.

[1120] Step 5:

[1121] Comparison of meeting minutes data

[1122] The server uses natural language processing (NLP) algorithms to compare today's minutes with past minutes.

[1123] The server detects changes in keywords and phrases in the content of comments.

[1124] Step 6:

[1125] Emotion analysis

[1126] The server uses an emotion engine to analyze the emotions of speakers in real time during a conference.

[1127] The device sends the speaker's tone of voice and facial expression data (if necessary) to the emotion engine.

[1128] Step 7:

[1129] Storing Emotional Data

[1130] The server receives the analyzed emotion data and stores it together with the provisional minutes data.

[1131] Step 8:

[1132] Extracting Differences

[1133] The server identifies fluctuations and inconsistencies in what is being said.

[1134] The server builds a list of the identified differences and aggregates the data.

[1135] Step 9:

[1136] Visualization of differential data and emotion data

[1137] The server generates visual representations such as graphs and charts based on the extracted differences and emotion data.

[1138] The server formats the visualization data into a format that is easy for the user to understand.

[1139] Step 10:

[1140] User Notification

[1141] The server adds explanatory text to the visualized data and sends it to the specified user as an email or internal notification.

[1142] The user checks the received notification and understands changes in the content of the comments and the results of sentiment analysis.

[1143] Step 11:

[1144] Get feedback

[1145] The user checks the notification and, if necessary, asks their superior or manager about the reason for the change in the comment content.

[1146] The user enters this feedback into the system.

[1147] Step 12:

[1148] Feedback Data Storage

[1149] The server receives the feedback data sent by the user and stores it in a database.

[1150] The server will prepare to use the data for future analysis and improvements.

[1151] These steps allow the system to efficiently record what is said during a meeting, compare it with past comments, and analyze and visualize the emotions of the speakers. Based on this information, users can identify the reasons for fluctuations in comments and take action to improve the quality of the meeting.

[1152] Example 2

[1153] 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."

[1154] Conventional meeting minutes systems lack the functionality to transcribe speeches made during meetings in real time and compare them with past minutes. It is also difficult to visualize fluctuations or inconsistencies in the speech content of meeting participants. Furthermore, there is no way to analyze the emotions of speakers and gain a deeper understanding of the meeting's progress, so the transparency and efficiency of meetings are not sufficiently ensured. To solve these issues, a system is needed that not only transcribes speeches in real time, but also compares speech content and analyzes emotions, and visualizes the results for users.

[1155] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for transcribing statements made during a meeting in real time, means for saving past meeting minutes, means for comparing transcription data of current meeting statements with past meeting minutes data, means for analyzing the emotions of users and speakers, and means for visualizing the extracted differences and emotion analysis data as graphs or charts. This makes it possible to improve the transparency and efficiency of meetings by comparing the consistency of statements made during a meeting with the contents of past statements, analyzing the emotions of speakers, and visually providing this to the user.

[1156] "A means for transcribing statements made during a meeting in real time" is a system that records audio during a meeting in real time and instantly converts it into text using voice recognition technology.

[1157] The "means for storing minutes of past meetings" is a system that stores the contents of statements made in previous meetings as text data in a database or the like.

[1158] "Means for comparing transcription data of current meeting remarks with past minutes data" is a function that uses natural language processing technology to compare the content of remarks made in the most recent meeting with past meeting records, and detects fluctuations or lack of consistency in the content.

[1159] "Means to extract variations and lack of consistency in speech content" is a function that identifies how speech made during a meeting differs from past speech and whether consistency is maintained, and clearly shows the differences.

[1160] "Means for analyzing the emotions of users and speakers" refers to technology that analyzes and identifies the emotional state of speakers during a meeting based on their voice, text, and, if necessary, facial expression data.

[1161] "Means for visualizing extracted differences and sentiment analysis data as graphs or charts" refers to a system that converts fluctuations in speech content and sentiment analysis results into visual data, i.e., visual formats such as graphs and charts, and displays them.

[1162] "Means for notifying users" refers to a function for delivering analysis results and visualization data to users via email or a notification system.

[1163] "Means for collecting feedback from users" is a function that collects opinions and responses that users input to the system and saves them as data.

[1164] The "means for storing collected feedback data" is a function for storing the feedback information collected from users in a database or the like and managing it for use in later analysis and improvement.

[1165] This invention combines an emotion engine with a system that transcribes statements made during meetings in real time and compares them with past minutes. The system aims to improve the transparency and efficiency of meetings by visualizing fluctuations and inconsistencies in the content of statements and analyzing the emotions of speakers.

[1166] First, when a meeting begins, the user starts recording audio on their device. The device then uses a speech recognition engine such as the Google Cloud Speech-to-Text API to convert what is being said into text in real time. This text data is then sent to a server in real time and saved as temporary meeting minutes.

[1167] The server accesses a database of past meeting minutes and compares the current meeting's content with past minutes using natural language processing techniques such as BERT and GPT. This comparison detects fluctuations and inconsistencies in the content of the speech.

[1168] Furthermore, the server uses an emotion analysis engine such as IBM Watson Tone Analyzer to analyze the emotions of users and speakers in real time. This analysis involves recognizing emotions based on voice tone, speech content, and, in some cases, facial expression data captured by a camera, and saving the analysis data.

[1169] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts, allowing users to see not only the changes in the content of comments but also the emotions expressed at the time of the comments at a glance. The server then sends this visualized data to users via email or a notification system.

[1170] The user can review the data and, if necessary, ask their superiors or managers why the content was changed. The user then enters the feedback they received into the system, and the server stores this feedback data. This feedback data can be used to analyze and improve future meetings.

[1171] Specific examples

[1172] Examples of Meeting A and Meeting B

[1173] For example, if a boss says "We'll start a new project next week" during Meeting A, and then the same boss says "We'll start a new project the week after next" during a subsequent Meeting B, the system will operate as follows:

[1174] 1. Recording and Transcription:

[1175] The device records what is said in Conference B and converts it into text in real time. This text data is then sent to the server.

[1176] 2. Comparing minutes:

[1177] The server retrieves the minutes of Meeting A from the database and uses natural language processing technology to compare the content of statements made in Meeting A and Meeting B.

[1178] 3. Extract and visualize the differences:

[1179] The server detects changes in the content of comments and detects changes from "next week" to "the week after next." This is visualized as a graph and sent to the user.

[1180] 4. Emotion analysis:

[1181] The emotion engine analyzes the emotion expressed at the time of speech and identifies emotions such as "anxiety." This is also visualized as a graph and sent to the user.

[1182] 5. Getting feedback:

[1183] The user checks the notification and asks their superior why the comment has changed. The feedback is then entered into the system, and the server stores it in a database.

[1184] Prompt Sentence Examples

[1185] An example of a prompt to input to a generative AI model is as follows:

[1186] "Discuss a system that transcribes speech during a meeting in real time and compares it with past minutes. In this system, the user starts recording, and the data is converted to text through a speech recognition engine and sent to a server. The server compares the speech using a natural language processing algorithm and analyzes the speaker's sentiment using an emotion engine. Finally, the system visualizes the differences and the sentiment analysis results and notifies the user."

[1187] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1188] Step 1: Start recording and real-time transcription

[1189] When a meeting begins, users start recording audio on their device, which then uses a speech recognition engine (Google Cloud Speech-to-Text API) to convert what is being said into text in real time.

[1190] Input: Meeting audio

[1191] Data processing: The voice recognition engine converts the voice data into text data.

[1192] Output: Text data generated in real time

[1193] Specific operation: The user presses the "Start Recording" button in the application. The device collects audio data through the microphone and converts the speech into text using the Google Cloud Speech-to-Text API.

[1194] Step 2: Sending real-time data to the server

[1195] The device sends the generated text data in real time to the server, which temporarily stores the data as provisional minutes.

[1196] Input: Real-time generated text data

[1197] Data processing: Send text data to the server via the network

[1198] Output: Text data saved as provisional minutes

[1199] Specific operation: The terminal sends text data to the server at regular intervals, and the server temporarily stores the received data.

[1200] Step 3: Obtain and compare past minutes

[1201] The server accesses the database and compares the current meeting content with past meeting minutes using natural language processing techniques (such as BERT and GPT).

[1202] Input: Current meeting text data and past meeting minutes data

[1203] Data processing: Analyzing text data using natural language processing techniques to detect variations and inconsistencies in spoken content

[1204] Output: Data showing fluctuations and inconsistencies in speech

[1205] How it works: The server extracts keywords from the stored text data, queries the database to retrieve relevant past meeting transcripts, and then applies natural language processing techniques such as the BERT model to identify variations and inconsistencies in the speech.

[1206] Step 4: Perform sentiment analysis

[1207] The server uses an emotion engine (such as IBM Watson Tone Analyzer) to analyze the emotions of users and speakers in real time.

[1208] Input: Meeting speech text and voice data, and facial expression data if necessary

[1209] Data processing: Using an emotion engine to analyze the speaker's emotions from voice and text

[1210] Output: Parsed emotion data

[1211] Specific operation: The server inputs the acquired voice, text, and facial expression data into the emotion engine, and then quantifies and stores the analyzed emotional state.

[1212] Step 5: Visualize the difference and sentiment analysis data

[1213] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts.

[1214] Input: Data on fluctuations in speech content and sentiment analysis data

[1215] Data processing: Generate graphs and charts using visualization tools (e.g., D3.js)

[1216] Output: Visual data (graphs and charts)

[1217] Specific operation: The server uses visualization tools to generate graphs and charts that highlight fluctuations and emotional changes.

[1218] Step 6: Communicate data and get feedback

[1219] The server sends the generated visualization data to the user via email or a notification system. The user checks the notification and, if necessary, asks their superior or manager for the reason for the change in the comment content.

[1220] Input: Visual data

[1221] Data processing: Distributing data via email and notification systems

[1222] Output: Visual data provided to the user

[1223] Specific operation: The server generates the visual data in PDF or web link format and sends a notification to the user, who receives the notification and asks his / her boss for feedback.

[1224] Step 7: Save the feedback and use it in your next meeting

[1225] Users input feedback from their superiors and managers into the system, and the server stores this feedback data and uses it to analyze and improve the next meeting.

[1226] Input: User feedback information

[1227] Data processing: Save the feedback data to a database

[1228] Output: Saved feedback data

[1229] Specific operation: The user enters comments and reasons on the feedback input screen, and the server stores them in the database. The feedback data is then used as a reference for the next analysis.

[1230] (Application example 2)

[1231] 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."

[1232] Conventional autonomous vehicles lack the means to monitor driver and passenger conversations in real time and improve safety based on that information. They also lack the ability to detect risks such as stress and lack of attention while driving in advance. Furthermore, there is a lack of comparative analysis with past driving records, making it difficult to respond appropriately to consistency and fluctuations.

[1233] The identification process by the identification 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 means for transcribing conversation content in real time, means for saving past conversation records, means for comparing transcription data of the current conversation content with past conversation record data, means for extracting fluctuations and inconsistencies in the conversation content, means for visualizing the extracted differences and emotion analysis data as graphs or charts, means for notifying the user, means for collecting feedback from the user, and means for saving the collected feedback data. This makes it possible to monitor conversation content and emotional fluctuations while driving in real time, thereby improving safety.

[1234] A "means for transcribing conversation content in real time" is a means that has the function of converting conversation from voice to text in real time.

[1235] The "means for saving records of past conversations" refers to a means having a function for saving text data of conversations recorded in the past.

[1236] "Means for comparing transcription data of current conversation content with past conversation record data" refers to a means that has the function of comparing data currently transcribed in real time with previously saved conversation record data to verify changes and consistency in content.

[1237] The "means for extracting fluctuations and inconsistencies in conversation content" refers to a means having a function for detecting and extracting changes and inconsistencies in conversation content.

[1238] "Means for visualizing extracted differences and emotion analysis data as graphs or charts" refers to means that has the function of generating graphs or charts to visually represent detected differences and analyzed emotion data.

[1239] The "means for notifying the user" refers to a means having a function for notifying the user of detected information and analysis results.

[1240] The "means for collecting feedback from users" refers to means that have the function of collecting opinions and information provided by users.

[1241] The "means for storing collected feedback data" refers to a means having a function for storing collected user feedback data.

[1242] "Audio recording means" means a means capable of recording conversations within a vehicle and storing the recording for later use.

[1243] A "means for analyzing emotions" is a means that has the function of identifying and analyzing emotions from the content of a conversation and tone of voice.

[1244] "Means for comparing conversation content using natural language processing algorithms" means means that use natural language processing technology to compare current conversation content with past records and verify consistency and variation.

[1245] This invention combines a sentiment analysis engine with a system that transcribes conversation content in real time, compares it with past conversation records, and monitors conversation fluctuations and utterance consistency. To realize this system, the following means are required.

[1246] First, the device records conversations in the car in real time and transcribes them. The device captures audio using a built-in microphone and converts this audio into text data in real time. This process uses voice recognition technology using the speech_recognition library.

[1247] The resulting text data is immediately sent to a server and temporarily stored as a provisional conversation record. The server then accesses a database of past conversation records and compares the current conversation with the past records using a natural language processing algorithm (NLP). This comparison method uses Hugging Face's natural language processing model.

[1248] The server then uses an emotion analysis engine to analyze emotions from the content of the conversation and the tone of voice. Hugging Face's emotion analysis model identifies emotions such as "stress," "impatience," and "fatigue" based on the text data obtained. This allows the server to understand the emotional state of the driver and passengers.

[1249] Furthermore, the server visualizes the comparison results and sentiment analysis data by using the matplotlib library to generate graphs and charts of the detected differences and sentiment data, converting them into a visually easy-to-understand format.

[1250] This visualized data is notified to the user (driver or in-car supervisor) in real time. For example, if the driver repeatedly says "I'm tired," the system analyzes in real time and issues a warning when it detects the emotion of "fatigue." The user can check this notification and take appropriate action if necessary.

[1251] The system also includes a means for users to provide feedback. Users receive notifications from the system and can enter their opinions or additional information into the system. The collected feedback data is stored in a database by the server and used for future conversation analysis.

[1252] As a specific example, "If a driver on their way to the office repeatedly says 'I'm tired,' the system will use its sentiment analysis engine to determine that the driver is 'fatigued' and issue a real-time warning." This prompt would be written as follows:

[1253] "If a driver on the way to the office repeatedly says 'I'm tired,' the system should use its sentiment analysis engine to determine that the driver is 'fatigued,' and devise a program to issue a real-time warning."

[1254] As described above, this invention not only records conversations in real time and compares them with past records, but also analyzes emotions, thereby improving safety and efficiency while driving.

[1255] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1256] Step 1:

[1257] The device uses the in-car microphone to record audio in real time. The input is in-car audio data, and the output is a recorded audio file. Specifically, the device's microphone picks up the conversations between the driver and passengers and records them as audio data.

[1258] Step 2:

[1259] The device transcribes recorded audio in real time. The input is an audio file, and the output is transcribed text data. The device uses the speech_recognition library to analyze the audio and convert it to text.

[1260] Step 3:

[1261] The transcribed text data is sent to a server and temporarily saved as a provisional conversation record. The input is the transcribed data, and the output is the provisional text record data on the server. The device sends the data to the server via network communication.

[1262] Step 4:

[1263] The server accesses a database of past conversation records and compares the current conversation content with past records using a natural language processing algorithm. The input is the current conversation text data and the past conversation record data, and the output is the comparison result data. Specifically, the server uses Hugging Face's NLP model to detect differences and inconsistencies in the text.

[1264] Step 5:

[1265] The server uses an emotion analysis engine to analyze emotions from the content of the conversation and tone of voice. The input is transcribed text data, and the output is analyzed emotional data. The server uses Hugging Face's emotion analysis model to identify emotions such as "stress," "anxiety," and "fatigue."

[1266] Step 6:

[1267] The server visualizes the detected differences and sentiment analysis data as graphs or charts. The input is the comparison results and sentiment data, and the output is a visual graph or chart. The server uses the matplotlib library to visually represent the detected variations and sentiment.

[1268] Step 7:

[1269] The server generates visualization data and notifies the user in real time. The input is the visualized data, and the output is a notification message to the user. For example, if the driver repeatedly says "I'm tired," the server generates a warning message and sends it to the user.

[1270] Step 8:

[1271] The user checks the notification from the system and provides feedback. The input is the user's feedback, and the output is the feedback data. The user checks the notification content and enters their opinions and information corresponding to that content into the system.

[1272] Step 9:

[1273] The server stores the collected feedback data and uses it for future conversation analysis. The input is the feedback data and the output is the stored feedback database. The server stores these data in a database and makes them available for subsequent dialogue analysis.

[1274] The above are the processing steps for realizing the invention.

[1275] 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.

[1276] 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.

[1277] 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.

[1278] [Fourth embodiment]

[1279] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1280] 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.

[1281] 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).

[1282] 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.

[1283] 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.

[1284] 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).

[1285] 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.

[1286] 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.

[1287] 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.

[1288] 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.

[1289] 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.

[1290] 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.

[1291] 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."

[1292] This invention relates to a system that transcribes statements made during meetings in real time and compares them with past minutes. This system visualizes inconsistencies in the content of statements made during meetings and monitors the consistency of statements.

[1293] First, when a meeting begins, the user starts recording audio on their device. The device sends what is said in the meeting to a speech recognition engine in real time, which converts it into text. This text data is then sent directly to the server in real time and temporarily saved as provisional minutes.

[1294] The server then accesses a database of past meeting minutes and compares the current meeting's content with past meeting minutes using natural language processing algorithms that detect variations or inconsistencies in the content.

[1295] The server converts the detected differences into visual formats such as graphs and charts, allowing users to see the changes in the content at a glance. The server then sends this visualization data to users via email or a notification system.

[1296] Users can review the data they receive and, if necessary, ask their superiors or managers why the content of their comments was changed. The users enter their feedback into the system, which the server stores. This feedback data is used to analyze and improve future meetings.

[1297] Specific examples

[1298] For example, suppose that during Meeting A, a boss says, "We'll start a new project next week." Then during the next Meeting B, the same boss says, "We'll start a new project the week after next." This system works as follows:

[1299] 1. Recording and Transcription:

[1300] The device records what is said in Conference B and converts it into text in real time.

[1301] The generated text data is sent to the server.

[1302] 2. Comparing minutes:

[1303] The server retrieves the minutes data for Conference A from the database.

[1304] Using a natural language processing algorithm, the content of statements made in Meeting A and Meeting B is compared.

[1305] 3. Extract and visualize the differences:

[1306] The server detects the change in the content of the message and discovers the change from "next week" to "the week after next."

[1307] The differences are visualized as a graph and sent to the user.

[1308] 4. Getting feedback:

[1309] The user checks the notification and asks their boss why the comment has changed.

[1310] The obtained feedback is entered into the system and the server stores it in a database.

[1311] In this way, this system records what is being said in real time and compares it with past comments to clarify inconsistencies in what is being said, improving the transparency and efficiency of meetings.

[1312] The processing flow will be explained below.

[1313] Step 1:

[1314] Start audio recording

[1315] The terminal activates the audio recording function at the start of the conference.

[1316] The device records what is said in the meeting in real time via a microphone.

[1317] Step 2:

[1318] Real-time audio transcription

[1319] The device sends the recorded voice data to the voice recognition engine in real time.

[1320] The device receives the text data returned from the voice recognition engine and converts the spoken content into text.

[1321] Step 3:

[1322] Sending text data and generating provisional minutes

[1323] The device transmits the text of the speech in real time to the server.

[1324] The server temporarily stores the received text data as provisional minutes.

[1325] Step 4:

[1326] Obtaining past meeting minutes data

[1327] The server retrieves past meeting minutes data from the database.

[1328] The server prepares the acquired minutes data for comparison with today's minutes data.

[1329] Step 5:

[1330] Comparison of meeting minutes data

[1331] The server uses natural language processing (NLP) algorithms to compare today's minutes with past minutes.

[1332] The server detects changes in keywords and phrases in the content of comments.

[1333] Step 6:

[1334] Extracting Differences

[1335] The server identifies fluctuations and inconsistencies in what is being said.

[1336] The server builds a list of the identified differences and aggregates the data.

[1337] Step 7:

[1338] Visualizing differential data

[1339] The server generates visual representations such as graphs and charts based on the extracted differences.

[1340] The server formats the visualization data into a format that is easy for the user to understand.

[1341] Step 8:

[1342] User Notification

[1343] The server adds explanatory text to the visualized data and sends it to the specified user as an email or internal notification.

[1344] The user checks the received notification and understands changes in the content of the comments.

[1345] Step 9:

[1346] Get feedback

[1347] The user checks the notification and, if necessary, asks their superior or manager about the reason for the change in the comment content.

[1348] The user enters this feedback into the system.

[1349] Step 10:

[1350] Feedback Data Storage

[1351] The server receives the feedback data sent by the user and stores it in a database.

[1352] The server will prepare to use the data for future analysis and improvements.

[1353] These steps allow the system to efficiently record what is said during a meeting, compare it with past comments to identify inconsistencies and changes, and report them back to meeting participants, allowing users to understand the reasons for fluctuations and take action to improve the quality of the meeting.

[1354] Example 1

[1355] 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."

[1356] Previously, it was difficult to record statements made during meetings in real time and understand the consistency and changes in the content of those statements. As a result, important decision-making and information transmission in meetings became uncertain, making efficient business operations difficult. Furthermore, comparing past minutes with current statements was time-consuming and labor-intensive. This made ensuring the transparency and reliability of meeting content a challenge.

[1357] 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.

[1358] In this invention, the server includes: means for transcribing statements made during a meeting in real time; means for saving past meeting minutes; means for comparing the transcription data of current meeting statements with past meeting minutes data; means for extracting variations and inconsistencies in the content of statements; means for visualizing the extracted differences as graphs or charts; means for notifying users; means for collecting feedback from users; means for saving the collected feedback data; means for transmitting voice data to a voice recognition engine in real time and acquiring it as text data; and means for comparing multiple meeting statements and analyzing the consistency of the statement content with high accuracy using a natural language processing algorithm. This automates the real-time recording and comparison of meeting content and makes it possible to visually confirm the consistency and variations in the content of statements, enabling efficient and transparent meeting management.

[1359] "Means for transcribing statements made during a meeting in real time" refers to a function or device that instantly converts speech made during a meeting into text data using voice recognition technology.

[1360] "Means for storing minutes of past meetings" refers to a function or device that stores the contents of past meetings as text data in a database or the like.

[1361] "Means for comparing the current meeting transcript with past meeting minutes" refers to algorithms or software that compare the transcript generated from the most recent meeting with similarly stored meeting minutes from the past.

[1362] "Means for detecting variations or inconsistencies in statements" refers to algorithms or software that identify changes or inconsistencies in statements from the results of comparisons.

[1363] "Means for visualizing extracted differences as graphs or charts" refers to a function or device for displaying the difference information extracted through comparison in a visually easy-to-understand format.

[1364] "Means for notifying users" refers to a function or device for sending information to users via text message, email, or other means of communication.

[1365] "Means for collecting feedback from users" refers to a function or device that allows users to input comments and opinions into the system and collects that information.

[1366] "Means for storing collected feedback data" refers to a database or storage system for storing feedback information obtained from users.

[1367] "Means for sending voice data to a voice recognition engine in real time and obtaining it as text data" refers to a function or device that sends voice during a meeting to a voice recognition system (API, etc.) in real time and receives the resulting text data.

[1368] "Means of using natural language processing algorithms to compare the content of multiple meetings and analyze the consistency of the content with high accuracy" refers to algorithms and software that use natural language processing technology to perform detailed analysis of the content of statements in different meeting records and evaluate their consistency.

[1369] The present invention relates to a system for transcribing statements made during a meeting in real time and comparing the results with past minutes. Specific embodiments of the system are described below.

[1370] First, when a user starts a meeting, they begin recording audio on their device. The device can be a smartphone or laptop, and recording is done using the device's built-in microphone. The recorded audio data is sent directly to a speech recognition engine in real time. Specifically, Google Cloud Speech-to-Text API or IBM Watson Speech to Text can be used. The device divides the audio data into buffers of a certain size, makes an API call, and converts the audio into text data.

[1371] The generated text data is then sent from the device to the server. The server receives it and temporarily stores it in a database as provisional minutes. Databases such as MySQL or PostgreSQL are often used for this purpose. The server accesses the database of past meeting minutes and retrieves past minutes related to the current meeting. For example, it extracts the data by executing the SQL query "SELECT FROM meeting_minutes WHERE meeting_id = 'XX'".

[1372] The server then uses natural language processing algorithms to compare the current meeting's content with past meeting minutes. These algorithms use models such as BERT and GPT-3. These algorithms tokenize and preprocess the text data before feeding it into a comparison model. The algorithm generates a similarity score for each utterance and detects variations or inconsistencies in the content.

[1373] The server converts the detected differences into graphs and charts and displays them in a visually easy-to-understand format. Libraries such as Matplotlib and D3.js are used for this visualization. For example, it is possible to display the differences in the content of comments over time as a bar graph.

[1374] The visualized data is sent from the server to the user via email or a notification system. Emails are sent using the SMTP protocol, and notifications are sent using the Slack API, etc. Users can review the data and ask their superiors or managers why the comments were changed.

[1375] Finally, the user enters their feedback into a form on the system and submits it. The server receives this feedback and stores it in a database, for example using the SQL query "INSERT INTO feedback (meeting_id, feedback_text) VALUES (XX, 'YY')".

[1376] As a concrete example, let's consider a case where a boss says in Meeting A, "We'll start a new project next week," and then in the next Meeting B, "The new project will start the week after next." In this case, the following behavior will occur.

[1377] 1. Recording and Transcription:

[1378] The device records what is said in Meeting B and converts it into text in real time using Google Cloud Speech-to-Text.

[1379] This text data is sent to the server and temporarily stored.

[1380] 2. Comparing minutes:

[1381] The server retrieves the minutes data for Meeting A from the MySQL database.

[1382] Use the BERT model to compare what was said in Meeting A and Meeting B.

[1383] 3. Extract and visualize the differences:

[1384] The server detects the change in the content of the message from "next week" to "the week after next."

[1385] The differences are graphed using Matplotlib and notified to the user.

[1386] 4. Getting feedback:

[1387] The user checks the notification and asks their boss why the comment has changed.

[1388] The feedback obtained is entered into the system, and the server stores it in a MySQL database.

[1389] This system makes it easy to record meeting content in real time and compare it with past parliamentary sessions, and allows for visual confirmation of consistency and fluctuations in speech content, enabling efficient and transparent meeting management.

[1390] Example prompt sentence:

[1391] Example: "Write a Python script to create a system that transcribes what is said during meetings in real time and compares it with past transcripts to visualize fluctuations in what is said."

[1392] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1393] Step 1:

[1394] Start audio recording:

[1395] When the conference starts, the user starts recording audio on the terminal.

[1396] Specific behavior: The user launches a recording application on their laptop or smartphone and clicks the "Start Recording" button.

[1397] Input: User's voice.

[1398] Output: Recorded audio data.

[1399] Step 2:

[1400] Real-time transcription of audio data:

[1401] The device splits the audio data being recorded and sends it to the voice recognition engine.

[1402] Specific operation: The terminal program captures voice data in a certain buffer size, sends the data to, for example, the Google Cloud Speech-to-Text API, and receives the text data obtained from the API.

[1403] Input: Recorded audio data.

[1404] Output: Real-time converted text data.

[1405] Step 3:

[1406] Sending text data to the server and temporarily saving it:

[1407] The terminal sends the text data returned from the voice recognition engine to the server.

[1408] Specific operation: The terminal sends an HTTP POST request to the server, and the server temporarily stores the received data in a MySQL database.

[1409] Input: Real-time text data.

[1410] Output: Text data temporarily stored in the server database.

[1411] Step 4:

[1412] Obtaining past minutes data:

[1413] The server accesses a past minutes database to retrieve relevant past meeting minutes.

[1414] Specific operation: The server executes the query "SELECT FROM meeting_minutes WHERE meeting_id = 'XX'" to retrieve the data.

[1415] Input: Past meeting ID or date and time.

[1416] Output: The retrieved past minutes data.

[1417] Step 5:

[1418] Comparison of current conference speeches with past minutes:

[1419] The server uses a natural language processing algorithm to compare the acquired minutes data with the current meeting text data.

[1420] Specific operation: Using natural language processing models such as BERT or GPT-3, the text data is tokenized and a similarity score is calculated.

[1421] Input: Current meeting text data and past meeting minutes data.

[1422] Output: Comparison results showing variability and inconsistencies in what was said.

[1423] Step 6:

[1424] Visualizing the differences:

[1425] The server visualizes the differences in the detected statements in graphs and charts.

[1426] Specific operation: Using Matplotlib and D3.js, fluctuations are displayed using bar graphs and heat maps along the time axis.

[1427] Input: Difference data from comparison.

[1428] Output: Visually displayed difference graphs and charts.

[1429] Step 7:

[1430] Sending visualization data:

[1431] The server notifies the user of the visualized data.

[1432] Specific behavior: Sends an email using the SMTP protocol or a notification using the Slack API.

[1433] Input: Visualization data.

[1434] Output: Notification to the user (email or notification message).

[1435] Step 8:

[1436] Get and store feedback:

[1437] The user checks the notification and enters feedback into the system if necessary, and the server stores the feedback.

[1438] What happens: A user enters feedback into a form on the system and clicks the "Submit" button. The server receives the data and stores it in a MySQL database.

[1439] Input: User feedback.

[1440] Output: Feedback data stored in a database.

[1441] (Application example 1)

[1442] 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."

[1443] With conventional meeting management systems, it was difficult to record what was said during meetings in real time and compare it with past minutes to maintain consistency. Furthermore, in factories and other workplaces, it was difficult to grasp the consistency and fluctuations of work instructions, which led to a decline in work efficiency. To solve these problems, a real-time work instruction monitoring system was required.

[1444] 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.

[1445] In this invention, the server includes means for transcribing statements made during a meeting in real time, means for saving minutes of past meetings, means for comparing transcription data of current meeting statements with past minutes data, means for extracting variations and inconsistencies in the content of statements, means for visualizing the extracted differences as graphs or charts, means for notifying users, means for collecting feedback from users, means for saving the collected feedback data, means for acquiring work instructions through real-time speech recognition, means for comparing with a database of past work instructions, means for detecting and visualizing differences, means for displaying the detected differences on a display, and means for sending user feedback to a manager. This makes it possible to monitor the consistency of statements and work instructions in real time and improve transparency.

[1446] A "meeting" is a gathering of participants to discuss and make decisions about a specific topic.

[1447] "Transcription" is the process of converting audio into text, either in real time or at a later time.

[1448] "Minutes" are documents that record the contents of a meeting and the decisions made.

[1449] "Minutes data" refers to the contents of past minutes saved as digital data.

[1450] A "natural language processing algorithm" is a technology that allows computers to analyze and process human language.

[1451] "Graphs and charts" are diagrams and tables that visually represent data.

[1452] A "notification" is a message or alert that informs the user of information.

[1453] "Feedback" refers to opinions and evaluation information provided by users.

[1454] "Speech recognition" is a technology that converts speech into text.

[1455] "Work instructions" are information for instructing a specific work.

[1456] "Work Order Database" means a digital database for storing past and current work orders.

[1457] "Difference" refers to the difference between the objects being compared.

[1458] A "display" is a device for visually displaying information.

[1459] An "administrator" is a person in charge of running a system or project.

[1460] This invention is a system that transcribes statements made during meetings in real time and compares them with past minutes. Furthermore, this technology is applied to factory robots to provide a real-time work instruction monitoring system that monitors the consistency of work instructions.

[1461] 1. Hardware and software configuration:

[1462] Hardware: Microphones and displays built into smartphones, smart glasses, head-mounted displays, or robots.

[1463] Software: We use the "speech_recognition" library for speech recognition, the "spacy" library for natural language processing, the "matplotlib" library for data visualization, and the "json" library for data storage and processing.

[1464] 2. Data processing and calculation processing:

[1465] Speech recognition: The device uses a microphone to collect speech and instructions in real time during meetings or factory work, then converts the speech data into text using a speech recognition engine (for example, the "speech_recognition" library).

[1466] Data transmission and storage: The generated text data is transmitted to the server in real time and temporarily stored as provisional minutes or work instructions.

[1467] Comparison and analysis: The server accesses a database of past meeting minutes or work instructions and compares them with the current text data. This is done using natural language processing algorithms (e.g., the "spacy" library).

[1468] Difference detection and visualization: Detect changes in what is said or what is directed, and visualize the differences as graphs or charts (for example, graphs created with the “matplotlib” library).

[1469] Notification and feedback: The visualized data is displayed to the user on the screen and feedback is collected through a notification system if necessary. The feedback is sent back to the server and stored in a database.

[1470] 3. Example:

[1471] If a boss says in Meeting A, "We'll start a new project next week," and then in the next Meeting B, the boss changes his mind and says, "The new project will start the week after next," the system will detect the change in real time and notify the user.

[1472] On a factory floor, if a robot receives the work instruction "Please start adjusting the mechanical parts on production line 1," and then the instruction is changed to "Please start adjusting the mechanical parts on production line 3," the robot will similarly detect the change and display it on the screen.

[1473] Example prompt sentence:

[1474] Take current work orders and compare them with past work orders to detect changes.

[1475] This structure allows for greater transparency and consistency in meetings and factory operations.

[1476] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1477] Step 1:

[1478] The device uses a microphone to capture the voice of the person speaking or working on the device, and this voice data is collected in real time and used as input data for subsequent processing.

[1479] Step 2:

[1480] The device sends the acquired voice data to a voice recognition engine (for example, the "speech_recognition" library) and converts it into text data. This process outputs the voice data as transcribed text.

[1481] Step 3:

[1482] The device sends the generated text data to the server, which temporarily stores it and registers it as the current minutes or work instructions.

[1483] Step 4:

[1484] The server retrieves corresponding data from a database of past meeting minutes or work instructions. This past data is text data and is used for comparison with the current data.

[1485] Step 5:

[1486] The server uses natural language processing algorithms (e.g., the "spacy" library) to compare the current text data with past data, detecting variations or inconsistencies in the data and reporting the differences.

[1487] Step 6:

[1488] The server generates visualization data such as graphs and charts based on the detected differences. In this process, the data is processed to visually represent the degree and content of the differences, and the results are output as visualization data.

[1489] Step 7:

[1490] The server sends the visualization data to the terminal, which displays it on the display, allowing the user to check the differences.

[1491] Step 8:

[1492] The user can then input feedback about any discrepancies they find, which is then sent to the server and used for future analysis.

[1493] Step 9:

[1494] The server stores the collected feedback data in a database, which stores the feedback for future analysis and system improvement.

[1495] 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.

[1496] This invention combines an emotion engine with a system that transcribes statements made during meetings in real time and compares them with past minutes. This system not only visualizes inconsistencies in the content of statements made during meetings and monitors the consistency of statements, but also recognizes and analyzes the emotions of users and speakers.

[1497] First, when a meeting begins, the user starts recording audio on their device. The device sends what is said in the meeting to a speech recognition engine in real time, which converts it into text. This text data is then sent directly to the server in real time and temporarily saved as provisional minutes.

[1498] The server accesses a database of past meeting minutes and compares the current meeting's content with past meeting minutes using natural language processing algorithms that detect variations or inconsistencies in the content of the conversation.

[1499] Furthermore, the server uses an emotion engine to analyze the emotions of users and speakers in real time. The emotion engine recognizes emotions based on tone of voice, content of speech, and facial expression data (if a camera is used), and stores the results as analysis data.

[1500] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts, allowing users to see not only the changes in the content of comments but also the emotions expressed at the time of the comments at a glance. The server then sends this visualized data to users via email or a notification system.

[1501] Users can review the data they receive and, if necessary, ask their superiors or managers why the content of their comments was changed. The users enter their feedback into the system, which the server stores. This feedback data is used to analyze and improve future meetings.

[1502] Specific examples

[1503] For example, suppose that during Meeting A, a boss says, "We'll start a new project next week." Then during the next Meeting B, the same boss says, "We'll start a new project the week after next." This system works as follows:

[1504] 1. Recording and Transcription:

[1505] The device records what is said in Conference B and converts it into text in real time.

[1506] The generated text data is sent to the server.

[1507] 2. Comparing minutes:

[1508] The server retrieves the minutes data for Conference A from the database.

[1509] Using a natural language processing algorithm, the content of statements made in Meeting A and Meeting B is compared.

[1510] 3. Extract and visualize the differences:

[1511] The server detects the change in the content of the message and discovers the change from "next week" to "the week after next."

[1512] The differences are visualized as a graph and sent to the user.

[1513] 4. Emotion analysis:

[1514] The emotion engine analyzes the emotions expressed by the boss when he or she speaks and identifies emotions such as "anxiety."

[1515] Emotional data is also visualized as a graph and sent to the user.

[1516] 5. Getting feedback:

[1517] The user checks the notification and asks their boss why the comment has changed.

[1518] The obtained feedback is entered into the system and the server stores it in a database.

[1519] In this way, the system not only records what is being said in real time and compares it with past comments, but also analyzes emotions to provide deeper insights, allowing meeting participants to monitor both the consistency of what is being said and the emotional fluctuations, improving the transparency and efficiency of meetings.

[1520] The processing flow will be explained below.

[1521] Step 1:

[1522] Start audio recording

[1523] The terminal activates the audio recording function at the start of the conference.

[1524] The device records what is said in the meeting in real time via a microphone.

[1525] Step 2:

[1526] Real-time audio transcription

[1527] The device sends the recorded voice data to the voice recognition engine in real time.

[1528] The device receives the text data returned from the voice recognition engine and converts the spoken content into text.

[1529] Step 3:

[1530] Generate and save provisional minutes

[1531] The device transmits the text of the speech in real time to the server.

[1532] The server temporarily stores the received text data as provisional minutes.

[1533] Step 4:

[1534] Obtaining past meeting minutes data

[1535] The server retrieves past meeting minutes data from the database.

[1536] The server prepares the acquired minutes data for comparison with today's minutes data.

[1537] Step 5:

[1538] Comparison of meeting minutes data

[1539] The server uses natural language processing (NLP) algorithms to compare today's minutes with past minutes.

[1540] The server detects changes in keywords and phrases in the content of comments.

[1541] Step 6:

[1542] Emotion analysis

[1543] The server uses an emotion engine to analyze the emotions of speakers in real time during a conference.

[1544] The device sends the speaker's tone of voice and facial expression data (if necessary) to the emotion engine.

[1545] Step 7:

[1546] Storing Emotional Data

[1547] The server receives the analyzed emotion data and stores it together with the provisional minutes data.

[1548] Step 8:

[1549] Extracting Differences

[1550] The server identifies fluctuations and inconsistencies in what is being said.

[1551] The server builds a list of the identified differences and aggregates the data.

[1552] Step 9:

[1553] Visualization of differential data and emotion data

[1554] The server generates visual representations such as graphs and charts based on the extracted differences and emotion data.

[1555] The server formats the visualization data into a format that is easy for the user to understand.

[1556] Step 10:

[1557] User Notification

[1558] The server adds explanatory text to the visualized data and sends it to the specified user as an email or internal notification.

[1559] The user checks the received notification and understands changes in the content of the comments and the results of sentiment analysis.

[1560] Step 11:

[1561] Get feedback

[1562] The user checks the notification and, if necessary, asks their superior or manager about the reason for the change in the comment content.

[1563] The user enters this feedback into the system.

[1564] Step 12:

[1565] Feedback Data Storage

[1566] The server receives the feedback data sent by the user and stores it in a database.

[1567] The server will prepare to use the data for future analysis and improvements.

[1568] These steps allow the system to efficiently record what is said during a meeting, compare it with past comments, and analyze and visualize the emotions of the speakers. Based on this information, users can identify the reasons for fluctuations in comments and take action to improve the quality of the meeting.

[1569] Example 2

[1570] 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."

[1571] Conventional meeting minutes systems lack the functionality to transcribe speeches made during meetings in real time and compare them with past minutes. It is also difficult to visualize fluctuations or inconsistencies in the speech content of meeting participants. Furthermore, there is no way to analyze the emotions of speakers and gain a deeper understanding of the meeting's progress, so the transparency and efficiency of meetings are not sufficiently ensured. To solve these issues, a system is needed that not only transcribes speeches in real time, but also compares speech content and analyzes emotions, and visualizes the results for users.

[1572] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for transcribing statements made during a meeting in real time, means for saving past meeting minutes, means for comparing transcription data of current meeting statements with past meeting minutes data, means for analyzing the emotions of users and speakers, and means for visualizing the extracted differences and emotion analysis data as graphs or charts. This makes it possible to improve the transparency and efficiency of meetings by comparing the consistency of statements made during a meeting with the contents of past statements, analyzing the emotions of speakers, and visually providing this to the user.

[1573] "A means for transcribing statements made during a meeting in real time" is a system that records audio during a meeting in real time and instantly converts it into text using voice recognition technology.

[1574] The "means for storing minutes of past meetings" is a system that stores the contents of statements made in previous meetings as text data in a database or the like.

[1575] "Means for comparing transcription data of current meeting remarks with past minutes data" is a function that uses natural language processing technology to compare the content of remarks made in the most recent meeting with past meeting records, and detects fluctuations or lack of consistency in the content.

[1576] "Means to extract variations and lack of consistency in speech content" is a function that identifies how speech made during a meeting differs from past speech and whether consistency is maintained, and clearly shows the differences.

[1577] "Means for analyzing the emotions of users and speakers" refers to technology that analyzes and identifies the emotional state of speakers during a meeting based on their voice, text, and, if necessary, facial expression data.

[1578] "Means for visualizing extracted differences and sentiment analysis data as graphs or charts" refers to a system that converts fluctuations in speech content and sentiment analysis results into visual data, i.e., visual formats such as graphs and charts, and displays them.

[1579] "Means for notifying users" refers to a function for delivering analysis results and visualization data to users via email or a notification system.

[1580] "Means for collecting feedback from users" is a function that collects opinions and responses that users input to the system and saves them as data.

[1581] The "means for storing collected feedback data" is a function for storing the feedback information collected from users in a database or the like and managing it for use in later analysis and improvement.

[1582] This invention combines an emotion engine with a system that transcribes statements made during meetings in real time and compares them with past minutes. The system aims to improve the transparency and efficiency of meetings by visualizing fluctuations and inconsistencies in the content of statements and analyzing the emotions of speakers.

[1583] First, when a meeting begins, the user starts recording audio on their device. The device then uses a speech recognition engine such as the Google Cloud Speech-to-Text API to convert what is being said into text in real time. This text data is then sent to a server in real time and saved as temporary meeting minutes.

[1584] The server accesses a database of past meeting minutes and compares the current meeting's content with past minutes using natural language processing techniques such as BERT and GPT. This comparison detects fluctuations and inconsistencies in the content of the speech.

[1585] Furthermore, the server uses an emotion analysis engine such as IBM Watson Tone Analyzer to analyze the emotions of users and speakers in real time. This analysis involves recognizing emotions based on voice tone, speech content, and, in some cases, facial expression data captured by a camera, and saving the analysis data.

[1586] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts, allowing users to see not only the changes in the content of comments but also the emotions expressed at the time of the comments at a glance. The server then sends this visualized data to users via email or a notification system.

[1587] The user can review the data and, if necessary, ask their superiors or managers why the content was changed. The user then enters the feedback they received into the system, and the server stores this feedback data. This feedback data can be used to analyze and improve future meetings.

[1588] Specific examples

[1589] Examples of Meeting A and Meeting B

[1590] For example, if a boss says "We'll start a new project next week" during Meeting A, and then the same boss says "We'll start a new project the week after next" during a subsequent Meeting B, the system will operate as follows:

[1591] 1. Recording and Transcription:

[1592] The device records what is said in Conference B and converts it into text in real time. This text data is then sent to the server.

[1593] 2. Comparing minutes:

[1594] The server retrieves the minutes of Meeting A from the database and uses natural language processing technology to compare the content of statements made in Meeting A and Meeting B.

[1595] 3. Extract and visualize the differences:

[1596] The server detects changes in the content of comments and detects changes from "next week" to "the week after next." This is visualized as a graph and sent to the user.

[1597] 4. Emotion analysis:

[1598] The emotion engine analyzes the emotion expressed at the time of speech and identifies emotions such as "anxiety." This is also visualized as a graph and sent to the user.

[1599] 5. Getting feedback:

[1600] The user checks the notification and asks their superior why the comment has changed. The feedback is then entered into the system, and the server stores it in a database.

[1601] Prompt Sentence Examples

[1602] An example of a prompt to input to a generative AI model is as follows:

[1603] "Discuss a system that transcribes speech during a meeting in real time and compares it with past minutes. In this system, the user starts recording, and the data is converted to text through a speech recognition engine and sent to a server. The server compares the speech using a natural language processing algorithm and analyzes the speaker's sentiment using an emotion engine. Finally, the system visualizes the differences and the sentiment analysis results and notifies the user."

[1604] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1605] Step 1: Start recording and real-time transcription

[1606] When a meeting begins, users start recording audio on their device, which then uses a speech recognition engine (Google Cloud Speech-to-Text API) to convert what is being said into text in real time.

[1607] Input: Meeting audio

[1608] Data processing: The voice recognition engine converts the voice data into text data.

[1609] Output: Text data generated in real time

[1610] Specific operation: The user presses the "Start Recording" button in the application. The device collects audio data through the microphone and converts the speech into text using the Google Cloud Speech-to-Text API.

[1611] Step 2: Sending real-time data to the server

[1612] The device sends the generated text data in real time to the server, which temporarily stores the data as provisional minutes.

[1613] Input: Real-time generated text data

[1614] Data processing: Send text data to the server via the network

[1615] Output: Text data saved as provisional minutes

[1616] Specific operation: The terminal sends text data to the server at regular intervals, and the server temporarily stores the received data.

[1617] Step 3: Obtain and compare past minutes

[1618] The server accesses the database and compares the current meeting content with past meeting minutes using natural language processing techniques (such as BERT and GPT).

[1619] Input: Current meeting text data and past meeting minutes data

[1620] Data processing: Analyzing text data using natural language processing techniques to detect variations and inconsistencies in spoken content

[1621] Output: Data showing fluctuations and inconsistencies in speech

[1622] How it works: The server extracts keywords from the stored text data, queries the database to retrieve relevant past meeting transcripts, and then applies natural language processing techniques such as the BERT model to identify variations and inconsistencies in the speech.

[1623] Step 4: Perform sentiment analysis

[1624] The server uses an emotion engine (such as IBM Watson Tone Analyzer) to analyze the emotions of users and speakers in real time.

[1625] Input: Meeting speech text and voice data, and facial expression data if necessary

[1626] Data processing: Using an emotion engine to analyze the speaker's emotions from voice and text

[1627] Output: Parsed emotion data

[1628] Specific operation: The server inputs the acquired voice, text, and facial expression data into the emotion engine, and then quantifies and stores the analyzed emotional state.

[1629] Step 5: Visualize the difference and sentiment analysis data

[1630] The server converts the detected differences and sentiment analysis data into visual formats such as graphs and charts.

[1631] Input: Data on fluctuations in speech content and sentiment analysis data

[1632] Data processing: Generate graphs and charts using visualization tools (e.g., D3.js)

[1633] Output: Visual data (graphs and charts)

[1634] Specific operation: The server uses visualization tools to generate graphs and charts that highlight fluctuations and emotional changes.

[1635] Step 6: Communicate data and get feedback

[1636] The server sends the generated visualization data to the user via email or a notification system. The user checks the notification and, if necessary, asks their superior or manager for the reason for the change in the comment content.

[1637] Input: Visual data

[1638] Data processing: Distributing data via email and notification systems

[1639] Output: Visual data provided to the user

[1640] Specific operation: The server generates the visual data in PDF or web link format and sends a notification to the user, who receives the notification and asks his / her boss for feedback.

[1641] Step 7: Save the feedback and use it in your next meeting

[1642] Users input feedback from their superiors and managers into the system, and the server stores this feedback data and uses it to analyze and improve the next meeting.

[1643] Input: User feedback information

[1644] Data processing: Save the feedback data to a database

[1645] Output: Saved feedback data

[1646] Specific operation: The user enters comments and reasons on the feedback input screen, and the server stores them in the database. The feedback data is then used as a reference for the next analysis.

[1647] (Application example 2)

[1648] 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."

[1649] Conventional autonomous vehicles lack the means to monitor driver and passenger conversations in real time and improve safety based on that information. They also lack the ability to detect risks such as stress and lack of attention while driving in advance. Furthermore, there is a lack of comparative analysis with past driving records, making it difficult to respond appropriately to consistency and fluctuations.

[1650] The identification process by the identification 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 means for transcribing conversation content in real time, means for saving past conversation records, means for comparing transcription data of the current conversation content with past conversation record data, means for extracting fluctuations and inconsistencies in the conversation content, means for visualizing the extracted differences and emotion analysis data as graphs or charts, means for notifying the user, means for collecting feedback from the user, and means for saving the collected feedback data. This makes it possible to monitor conversation content and emotional fluctuations while driving in real time, thereby improving safety.

[1651] A "means for transcribing conversation content in real time" is a means that has the function of converting conversation from voice to text in real time.

[1652] The "means for saving records of past conversations" refers to a means having a function for saving text data of conversations recorded in the past.

[1653] "Means for comparing transcription data of current conversation content with past conversation record data" refers to a means that has the function of comparing data currently transcribed in real time with previously saved conversation record data to verify changes and consistency in content.

[1654] The "means for extracting fluctuations and inconsistencies in conversation content" refers to a means having a function for detecting and extracting changes and inconsistencies in conversation content.

[1655] "Means for visualizing extracted differences and emotion analysis data as graphs or charts" refers to means that has the function of generating graphs or charts to visually represent detected differences and analyzed emotion data.

[1656] The "means for notifying the user" refers to a means having a function for notifying the user of detected information and analysis results.

[1657] The "means for collecting feedback from users" refers to means that have the function of collecting opinions and information provided by users.

[1658] The "means for storing collected feedback data" refers to a means having a function for storing collected user feedback data.

[1659] "Audio recording means" means a means capable of recording conversations within a vehicle and storing the recording for later use.

[1660] A "means for analyzing emotions" is a means that has the function of identifying and analyzing emotions from the content of a conversation and tone of voice.

[1661] "Means for comparing conversation content using natural language processing algorithms" means means that use natural language processing technology to compare current conversation content with past records and verify consistency and variation.

[1662] This invention combines a sentiment analysis engine with a system that transcribes conversation content in real time, compares it with past conversation records, and monitors conversation fluctuations and utterance consistency. To realize this system, the following means are required.

[1663] First, the device records conversations in the car in real time and transcribes them. The device captures audio using a built-in microphone and converts this audio into text data in real time. This process uses voice recognition technology using the speech_recognition library.

[1664] The resulting text data is immediately sent to a server and temporarily stored as a provisional conversation record. The server then accesses a database of past conversation records and compares the current conversation with the past records using a natural language processing algorithm (NLP). This comparison method uses Hugging Face's natural language processing model.

[1665] The server then uses an emotion analysis engine to analyze emotions from the content of the conversation and the tone of voice. Hugging Face's emotion analysis model identifies emotions such as "stress," "impatience," and "fatigue" based on the text data obtained. This allows the server to understand the emotional state of the driver and passengers.

[1666] Furthermore, the server visualizes the comparison results and sentiment analysis data by using the matplotlib library to generate graphs and charts of the detected differences and sentiment data, converting them into a visually easy-to-understand format.

[1667] This visualized data is notified to the user (driver or in-car supervisor) in real time. For example, if the driver repeatedly says "I'm tired," the system analyzes in real time and issues a warning when it detects the emotion of "fatigue." The user can check this notification and take appropriate action if necessary.

[1668] The system also includes a means for users to provide feedback. Users receive notifications from the system and can enter their opinions or additional information into the system. The collected feedback data is stored in a database by the server and used for future conversation analysis.

[1669] As a specific example, "If a driver on their way to the office repeatedly says 'I'm tired,' the system will use its sentiment analysis engine to determine that the driver is 'fatigued' and issue a real-time warning." This prompt would be written as follows:

[1670] "If a driver on the way to the office repeatedly says 'I'm tired,' the system should use its sentiment analysis engine to determine that the driver is 'fatigued,' and devise a program to issue a real-time warning."

[1671] As described above, this invention not only records conversations in real time and compares them with past records, but also analyzes emotions, thereby improving safety and efficiency while driving.

[1672] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1673] Step 1:

[1674] The device uses the in-car microphone to record audio in real time. The input is in-car audio data, and the output is a recorded audio file. Specifically, the device's microphone picks up the conversations between the driver and passengers and records them as audio data.

[1675] Step 2:

[1676] The device transcribes recorded audio in real time. The input is an audio file, and the output is transcribed text data. The device uses the speech_recognition library to analyze the audio and convert it to text.

[1677] Step 3:

[1678] The transcribed text data is sent to a server and temporarily saved as a provisional conversation record. The input is the transcribed data, and the output is the provisional text record data on the server. The device sends the data to the server via network communication.

[1679] Step 4:

[1680] The server accesses a database of past conversation records and compares the current conversation content with past records using a natural language processing algorithm. The input is the current conversation text data and the past conversation record data, and the output is the comparison result data. Specifically, the server uses Hugging Face's NLP model to detect differences and inconsistencies in the text.

[1681] Step 5:

[1682] The server uses an emotion analysis engine to analyze emotions from the content of the conversation and tone of voice. The input is transcribed text data, and the output is analyzed emotional data. The server uses Hugging Face's emotion analysis model to identify emotions such as "stress," "anxiety," and "fatigue."

[1683] Step 6:

[1684] The server visualizes the detected differences and sentiment analysis data as graphs or charts. The input is the comparison results and sentiment data, and the output is a visual graph or chart. The server uses the matplotlib library to visually represent the detected variations and sentiment.

[1685] Step 7:

[1686] The server generates visualization data and notifies the user in real time. The input is the visualized data, and the output is a notification message to the user. For example, if the driver repeatedly says "I'm tired," the server generates a warning message and sends it to the user.

[1687] Step 8:

[1688] The user checks the notification from the system and provides feedback. The input is the user's feedback, and the output is the feedback data. The user checks the notification content and enters their opinions and information corresponding to that content into the system.

[1689] Step 9:

[1690] The server stores the collected feedback data and uses it for future conversation analysis. The input is the feedback data and the output is the stored feedback database. The server stores these data in a database and makes them available for subsequent dialogue analysis.

[1691] The above are the processing steps for realizing the invention.

[1692] 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.

[1693] 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.

[1694] 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.

[1695] 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.

[1696] 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.

[1697] 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.

[1698] 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).

[1699] 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.

[1700] 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."

[1701] 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.

[1702] 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).

[1703] 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.

[1704] 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.

[1705] 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.

[1706] 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.

[1707] 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.

[1708] 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.

[1709] 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.

[1710] 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.

[1711] 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.

[1712] 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.

[1713] The following is further disclosed regarding the above embodiment.

[1714] (Claim 1)

[1715] A means to transcribe what is said during meetings in real time,

[1716] A means of storing past meeting minutes;

[1717] A means to compare current meeting transcripts with past minutes;

[1718] A means of detecting variations and inconsistencies in the content of statements,

[1719] A means to visualize the extracted differences as graphs and charts,

[1720] a means for notifying a user;

[1721] a means for collecting feedback from users;

[1722] a means for storing the collected feedback data;

[1723] A system that includes these.

[1724] (Claim 2)

[1725] 10. The system of claim 1, further comprising means for recording audio.

[1726] (Claim 3)

[1727] 10. The system of claim 1, further comprising means for comparing utterances using a natural language processing algorithm.

[1728] "Example 1"

[1729] (Claim 1)

[1730] A means to transcribe what is said during meetings in real time,

[1731] A means of storing past meeting minutes;

[1732] A means to compare current meeting transcripts with past minutes;

[1733] A means of detecting variations and inconsistencies in the content of statements,

[1734] A means to visualize the extracted differences as graphs and charts,

[1735] a means for notifying a user;

[1736] a means for collecting feedback from users;

[1737] a means for storing the collected feedback data;

[1738] A means for transmitting voice data to a voice recognition engine in real time and acquiring the voice data as text data;

[1739] A method for comparing multiple conference statements using a natural language processing algorithm and analyzing the consistency of the statements with high accuracy.

[1740] A system that includes these.

[1741] (Claim 2)

[1742] 10. The system of claim 1, further comprising means for recording audio.

[1743] (Claim 3)

[1744] 10. The system of claim 1, further comprising means for comparing utterances using a natural language processing algorithm.

[1745] "Application Example 1"

[1746] (Claim 1)

[1747] A means to transcribe what is said during meetings in real time,

[1748] A means of storing past meeting minutes;

[1749] A means to compare current meeting transcripts with past minutes;

[1750] A means of detecting variations and inconsistencies in the content of statements,

[1751] A means to visualize the extracted differences as graphs and charts,

[1752] a means for notifying a user;

[1753] a means for collecting feedback from users;

[1754] a means for storing the collected feedback data;

[1755] a means for obtaining work instructions through real-time voice recognition;

[1756] a means of comparison with a database of past work instructions;

[1757] a means of detecting and visualizing the differences;

[1758] means for displaying the detected difference on a display;

[1759] a means for sending user feedback to an administrator;

[1760] A system that includes these.

[1761] (Claim 2)

[1762] 10. The system of claim 1, further comprising means for recording audio.

[1763] (Claim 3)

[1764] 10. The system of claim 1, further comprising means for comparing utterances using a natural language processing algorithm.

[1765] "Example 2: Combining Emotion Engines"

[1766] (Claim 1)

[1767] A means to transcribe what is said during meetings in real time,

[1768] A means of storing past meeting minutes;

[1769] A means to compare current meeting transcripts with past minutes;

[1770] A means of detecting variations and inconsistencies in the content of statements,

[1771] A means for analyzing the emotions of users and speakers;

[1772] A means to visualize the extracted differential and sentiment analysis data as graphs and charts;

[1773] a means for notifying a user;

[1774] a means for collecting feedback from users;

[1775] a means for storing the collected feedback data;

[1776] A system that includes these.

[1777] (Claim 2)

[1778] 10. The system of claim 1, further comprising means for recording audio and means for converting the audio into text through speech recognition in real time.

[1779] (Claim 3)

[1780] 10. The system of claim 1, further comprising means for comparing utterances using natural language processing techniques and means for analyzing emotions using emotion recognition techniques.

[1781] "Application example 2 when combining emotion engines"

[1782] (Claim 1)

[1783] A means of transcribing conversations in real time,

[1784] A means of storing records of past conversations;

[1785] A means for comparing current conversation transcripts with past conversation transcripts;

[1786] A means of extracting variations and inconsistencies in conversation content,

[1787] A means to visualize the extracted differences and sentiment analysis data as graphs and charts,

[1788] a means for notifying a user;

[1789] a means for collecting feedback from users;

[1790] a means for storing the collected feedback data;

[1791] A system that includes these.

[1792] (Claim 2)

[1793] 10. The system of claim 1, further comprising: means for recording audio; and means for analyzing emotions.

[1794] (Claim 3)

[1795] 10. The system of claim 1, further comprising means for comparing conversation content using a natural language processing algorithm. [Explanation of symbols]

[1796] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A means to transcribe what is said during meetings in real time, A means of storing past meeting minutes; A means to compare current meeting transcripts with past minutes; A means of detecting variations and inconsistencies in the content of statements, A means to visualize the extracted differences as graphs and charts, a means for notifying a user; a means for collecting feedback from users; a means for storing the collected feedback data; A system that includes these.

2. The system of claim 1 further comprising means for recording audio.

3. The system of claim 1 further comprising means for comparing utterances using a natural language processing algorithm.

Images