System

Abstract

A system is provided.SOLUTION: A system comprising: means for collecting speech of a participant as audio data; means for converting the audio data into text data; means for analyzing the text data to generate real-time feedback; and means for providing the feedback to the participant.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 many organizations, superiors or influential members often take the lead, creating an environment in which other participants find it difficult to express their opinions. This can lead to issues such as a lack of fairness in discussions, lowering participants' sense of satisfaction and productivity. Furthermore, overbearing remarks and harassment can occur, further reducing the quality of discussions. The present invention aims to solve these problems and provide a fair and effective meeting environment in which everyone can easily express their opinions. [Means for solving the problem]

[0005] The system according to the invention comprises the following means:

[0006] 1. Methods for collecting participant speech as audio data:

[0007] The speech of each participant during a meeting is collected in real time as audio data.

[0008] 2. How to convert audio data to text data:

[0009] By converting the collected voice data into text data, the content of the speech can be analyzed.

[0010] 3. Means of analyzing text data and generating real-time feedback:

[0011] The text data is analyzed using natural language processing (NLP) technology, the content and sentiment of the comments are evaluated, and real-time feedback is generated, which may be either closed feedback or open feedback.

[0012] 4. How to provide feedback to participants:

[0013] Closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window that can be seen by all participants.

[0014] 5. How to detect and immediately flag bullying and harassment:

[0015] It monitors comments made during meetings in real time, detects comments that are deemed to be overbearing or harassing, and provides a warning message to the relevant participants.

[0016] 6. A means to record the generated feedback and speech data and automatically generate meeting minutes after the meeting:

[0017] All comments and feedback during the meeting are recorded, and minutes are automatically generated and provided to participants after the meeting ends.

[0018] 7. How to start a meeting and manage participant invitations:

[0019] Includes means to schedule meetings, generate join links, and invite participants.

[0020] This makes it easier for everyone to express their opinions, provides a fair and equitable forum for discussion, and allows real-time feedback to improve the quality of discussions, ensuring that everyone feels empowered and productive in their work.

[0021] "Participant" refers to a person who attends a meeting and expresses an opinion.

[0022] "Speech" refers to verbal opinions or comments made by participants during a meeting.

[0023] "Audio Data" means digital audio files of a Participant's speech.

[0024] "Text data" is data that is voice data converted into character information and expressed as text.

[0025] "Collection means" refers to the devices and methods used to record participants' speech as audio data.

[0026] "Conversion means" refers to the process or technology used to convert audio data into text data.

[0027] "Feedback" refers to advice, evaluation, or comments provided in response to a participant's comments or the content of the discussion.

[0028] "Closed feedback" refers to feedback provided in a format that can only be viewed by certain participants.

[0029] "Open feedback" refers to feedback provided in a format that is visible to all meeting participants.

[0030] "High-handed remarks" refers to attitudes and language that intimidate other participants.

[0031] "Harassment" refers to any remarks or actions that make other participants feel uncomfortable or intimidated.

[0032] "Detection methods" refer to processes or technologies used to identify specific patterns, such as coercive language or harassment.

[0033] "Warning measures" refer to methods of providing participants who make problematic statements with a message urging them to immediately correct their comments.

[0034] "Minutes" refers to a document that compiles a record of what was said and feedback given during a meeting.

[0035] "Recording methods" refers to the process or technology used to capture what is said and feedback during a meeting.

[0036] "Automatic generation means" refers to algorithms or technologies for generating minutes without human intervention. [Brief explanation of the drawings]

[0037] [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

[0038] 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.

[0039] First, the terms used in the following description will be explained.

[0040] 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).

[0041] 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.

[0042] 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.

[0043] 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.

[0044] 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."

[0045] [First embodiment]

[0046] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0047] 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.

[0048] 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).

[0049] 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.

[0050] 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.

[0051] 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.

[0052] 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.

[0053] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0054] 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.

[0055] 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.

[0056] 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.

[0057] 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."

[0058] Program Overview

[0059] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. This system includes "means for collecting participants' comments as audio data," "means for converting the audio data into text data," "means for analyzing the text data and generating real-time feedback," "means for providing feedback to participants," and "means for detecting overbearing comments and harassment and issuing an immediate warning."

[0060] Program processing

[0061] The program of the system of the present invention is processed as follows.

[0062] 1. Audio data collection:

[0063] The device collects speech from meeting participants as audio data in real time, specifically including the ability to capture microphone input from each participant.

[0064] 2. Audio to text conversion:

[0065] The server receives the voice data sent from the device and converts it into text data using a voice recognition API, etc. This makes it possible to analyze the content of the speech as text information.

[0066] 3. Text data analysis:

[0067] The server then analyzes the converted text data using natural language processing (NLP) technology, analyzing keywords, emotions, tone, and other elements within the text to assess the specificity and relevance of the comments.

[0068] 4. Generate real-time feedback:

[0069] The server generates real-time feedback based on the analysis results, including closed and open feedback, in an appropriate format.

[0070] 5. Providing Feedback:

[0071] The terminal receives the generated feedback and provides it to the user. Closed feedback is displayed in the chat window of a specific participant, while open feedback is displayed so that all participants can see it.

[0072] 6. Detect and warn against aggressive speech:

[0073] The server performs real-time emotional analysis of comments and detects statements that are deemed to be overbearing or harassing.

[0074] If the server detects such a situation, it immediately generates a warning message and sends it to the terminal of the participant in question.

[0075] The terminal displays a warning message to the user in real time, prompting the user to take appropriate action.

[0076] 7. Generate and share meeting minutes:

[0077] The server records all comments and feedback during the meeting, and automatically generates minutes after the meeting and provides them to all participants.

[0078] The device displays the generated minutes to the user and provides a link that can be downloaded or shared.

[0079] Specific examples

[0080] Meeting Management

[0081] User (Person A): "I have some opinions about the direction this project should take."

[0082] The device captures Mr. A's speech as audio data and sends it to the server.

[0083] The server converts the voice data into text data and analyzes the content using NLP technology.

[0084] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[0085] Person A receives the feedback and makes a statement providing specific data.

[0086] The device recaptures Mr. A's new comments and sends them to the server.

[0087] The server analyzes the new utterance and provides further feedback.

[0088] In this way, the system provides a fair and effective meeting environment in real time, ensuring that all participants feel comfortable speaking up. Furthermore, the system automatically generates meeting minutes, making it easy to review the content of meetings even after they have ended.

[0089] The processing flow will be explained below.

[0090] Program processing flow

[0091] Step 1: Set up a meeting

[0092] 1. A user schedules a meeting.

[0093] Users log into the system and set the meeting date and time, participant list, and meeting goal.

[0094] 2. The server creates a meeting room.

[0095] The server receives the request from the user, generates a unique meeting room ID, and creates a join link.

[0096] The server pre-loads and prepares the AI ​​models to be used in the meeting.

[0097] 3. The user invites participants.

[0098] The user sends the generated meeting link to the participants, who click the link to enter the meeting room.

[0099] Step 2: Collecting audio data

[0100] 4. The device captures the audio input.

[0101] The device captures microphone input each time a participant speaks and transmits it to the server as audio data in real time.

[0102] Step 3: Convert audio data to text

[0103] 5. The server converts the audio data into text data.

[0104] The server converts the received voice data into text data using a speech recognition API, which converts the voice into text information that can be analyzed.

[0105] Step 4: Analyze the text data

[0106] 6. The server analyzes the text data.

[0107] The server then analyzes the converted text data using natural language processing (NLP) techniques, specifically extracting keywords, analyzing sentiment, and evaluating tone.

[0108] Step 5: Generate real-time feedback

[0109] 7. The server generates real-time feedback.

[0110] Based on the analysis results, the server generates either crowded or open feedback. Closed feedback is given to specific participants, while open feedback is given to all participants.

[0111] 8. The device will display feedback.

[0112] The terminal displays closed feedback to a specific user in a chat window, and displays open feedback in a feedback window that can be viewed by all participants.

[0113] Step 6: Detect and flag coercive language

[0114] 9. The server monitors the comments.

[0115] The server monitors all comments in real time and performs sentiment analysis.

[0116] 10. The server detects overbearing comments.

[0117] The server detects overbearing remarks and harassment based on the results of sentiment analysis.

[0118] 11. The server issues a warning.

[0119] If the server detects an overbearing remark, it generates and sends a warning message to the user.

[0120] 12. The terminal displays a warning message.

[0121] The device will display a warning message to the affected user in real time, urging them to reconsider their comments.

[0122] Step 7: Generate and share meeting minutes

[0123] 13. The server records your comments and feedback.

[0124] The server records all comments and feedback during the meeting and stores them in a database.

[0125] 14. The server generates the minutes.

[0126] The server automatically generates minutes based on the data recorded after the meeting ends.

[0127] 15. The server shares the minutes.

[0128] The server provides the generated minutes to all participants via email or link.

[0129] 16. The device will be able to display and download the minutes.

[0130] The terminal provides a function that allows the user to view and download the minutes.

[0131] Example: Feedback process

[0132] User (Person A): "I have some opinions about the direction this project should take."

[0133] The device captures A's remarks and sends them to the server.

[0134] The server converts the voice data into text data and analyzes the content using NLP technology.

[0135] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or specific data."

[0136] Person A receives the feedback and makes a statement providing specific data.

[0137] The device recaptures the new utterance and sends it to the server.

[0138] The server analyzes the new utterance and provides further feedback.

[0139] Through this process, the system provides a fair and effective meeting environment in real time.

[0140] Example 1

[0141] 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."

[0142] In online and offline meetings, it is necessary to create an environment where all participants can express their opinions fairly and easily, and to prevent discussions from being biased towards certain individuals. There is also a need for systems that can detect overbearing remarks and harassment in real time and respond immediately. Furthermore, to efficiently conduct meetings, it is desirable to provide real-time feedback and automatically generate meeting minutes. However, conventional systems have not been able to adequately resolve these issues.

[0143] 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.

[0144] In this invention, the server includes means for collecting participants' comments as audio data, means for converting the audio data into text data, means for analyzing the text data using natural language processing technology, means for generating real-time feedback based on the analysis results, means for providing the feedback to participants, means for detecting overbearing comments or harassment and generating a warning message, and means for recording the comments and feedback of the meeting and generating minutes after the meeting. This provides an environment in which all participants can speak fairly and easily, making it possible to provide effective feedback in real time. Furthermore, overbearing comments and harassment can be addressed immediately, and the automatically generated minutes make it easy to review the content of the discussion.

[0145] "Participant" means an individual who participates in a meeting or conference.

[0146] "Audio data" refers to data that has been digitized from audio recorded as sound waves.

[0147] "Text data" is data that has been digitized as character information.

[0148] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[0149] "Real-time feedback" refers to instant assessment and advice provided during a meeting.

[0150] "Positive speech" is speech that sounds intimidating or aggressive to others.

[0151] "Harassment" refers to behavior or speech that harasses or offends others.

[0152] A "warning message" is a notification that warns you against certain actions or statements.

[0153] "Minutes" are a written record of all statements and discussions made during a meeting or conference.

[0154] "Means of collection" refers to devices or software that capture participants' voices as digital data.

[0155] "Means for converting" refers to a technology or system that converts voice data into text data.

[0156] "Means of analysis" refers to the technology or system used to analyze text data and determine meaning and sentiment.

[0157] "Means for generating" refers to a technology or system that generates feedback or warning messages based on the analysis results.

[0158] "Means for providing" refers to the equipment or software used to deliver generated feedback and warning messages to participants.

[0159] "Means of recording and generating" refers to technology or systems that store comments and feedback during meetings and later compile them into minutes.

[0160] A "system" is a set of devices or software that integrates these means to achieve a set of functions.

[0161] MODE FOR CARRYING OUT THE INVENTION

[0162] This invention is a system that uses AI to provide real-time feedback during online and offline meetings, creating an environment where all participants can easily express their opinions. This system includes "means for collecting participants' comments as audio data," "means for converting the audio data into text data," "means for analyzing the text data using natural language processing technology," "means for generating real-time feedback based on the analysis results," "means for providing feedback to participants," "means for detecting overbearing comments and harassment and generating a warning message," and "means for recording comments and feedback during meetings and generating meeting minutes after the meeting."

[0163] Specifically, the process is as follows:

[0164] Audio data collection

[0165] The device collects speech from each participant during the meeting as real-time audio data using a built-in microphone or an external microphone, and transmits the collected audio data to a server in a compressed format (e.g., MP3 or WAV format).

[0166] Converting audio data to text

[0167] The server receives the voice data sent from the device and converts it into text using a speech recognition API such as Google Cloud Speech-to-Text API or IBM Watson Speech to Text, then temporarily stores the converted text in a database.

[0168] Text data analysis

[0169] The server analyzes the stored text data using natural language processing (NLP) technology. Specifically, it uses Google Cloud Natural Language API and spaCy to evaluate keywords, sentence tone, and sentiment within the text. It evaluates the specificity and relevance of the text data and generates closed-form feedback.

[0170] Generate real-time feedback

[0171] The server generates real-time feedback based on the analysis results, classifies the feedback into closed feedback (for specific participants) and open feedback (for all participants), and sends it to the terminal in the appropriate format.

[0172] Providing Feedback

[0173] The device receives the feedback sent from the server, and displays the closed feedback in the chat window of the specific participant, while the open feedback is displayed on a screen that can be viewed by all participants. Platforms such as Microsoft Teams and Zoom may be used.

[0174] Detecting and warning against coercive speech

[0175] The server analyzes the sentiment of comments in real time and detects any overbearing or harassing content. If any are detected, a warning message is generated and sent to the relevant participant's device.

[0176] The device displays a warning message to the user in real time and prompts the user to take appropriate action, typically using the notification function of the web browser.

[0177] Generate and share meeting minutes

[0178] The server records all comments and feedback during the meeting, automatically generates minutes after the meeting, and generates a link to share the minutes with all participants.

[0179] The device displays the generated minutes to the user and provides download and sharing links, and can output them in Microsoft Word and Google Docs formats.

[0180] Specific examples

[0181] Meeting Management

[0182] User (Person A): "I have some opinions about the direction this project should take."

[0183] The device captures A's speech as audio data and sends it to the server. The captured audio data is saved in WAV format.

[0184] The server uses the Google Cloud Speech-to-Text API to convert Mr. A's voice data into text data in real time. The converted text data is recorded as "I have an opinion about the direction of this project."

[0185] The server analyzes the text and determines that it is "not specific enough" - NLP analysis detects that it does not contain specific keywords.

[0186] The server generates closed feedback such as "It would be more effective if you provided examples or concrete data."

[0187] The device will display this feedback in Mr. A's individual chat window.

[0188] Examples of problematic statements

[0189] User B makes a commanding statement: "Your idea is pointless."

[0190] The device captures Mr. B's speech as audio data and sends it to the server.

[0191] The server converts the voice data into text data and performs sentiment analysis using NLP technology. The analysis results indicate that the remark is "overbearing."

[0192] The server generates a warning message saying, "An overbearing remark has been detected. Please choose appropriate words." and sends it to Mr. B's device.

[0193] The device will notify Mr. B of this warning message in real time, urging him to be careful.

[0194] Example prompts for generative AI models

[0195] Please analyze the following text data and determine whether it constitutes coercive remarks or harassment.

[0196] Data Text: "That suggestion makes absolutely no sense. Anyone could have thought of something like that."

[0197] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0198] Program processing flow

[0199] Step 1: Collecting audio data

[0200] The device collects speech from each participant during a meeting as audio data in real time. The input is the participant's voice, and the output is audio data. Specifically, the device captures audio using a built-in microphone or an external microphone. The collected audio data is temporarily saved in WAV format and sent to a server.

[0201] Step 2: Convert audio data to text

[0202] The server receives the voice data sent from the device. The input is WAV format voice data, and the output is text data. Specifically, the voice data is converted into text data using the Google Cloud Speech-to-Text API or IBM Watson Speech to Text. At this time, the voice data is converted into text data and temporarily stored in a database.

[0203] Step 3: Analyze the text data

[0204] The server analyzes the stored text data using natural language processing (NLP) technology. The input is text data, and the output is the analysis results (keywords, tone, and emotional information). Specifically, it uses Google Cloud Natural Language API and spaCy to extract keywords from the text data, evaluate tone, and analyze emotions. Based on the results of this analysis, the specificity and relevance of the text are evaluated, and feedback candidates are generated.

[0205] Step 4: Generate real-time feedback

[0206] The server generates real-time feedback based on the analysis results. The input is the analysis results, and the output is the feedback message. The server classifies the feedback into closed feedback (for specific participants) and open feedback (for all participants) formats and sends it to the terminal in the appropriate format. The generative AI model is used to generate the feedback text.

[0207] Step 5: Provide feedback

[0208] The device receives feedback sent from the server. The input is the feedback message, and the output is the feedback provided to the user. Specifically, closed feedback is displayed in the chat window of a specific participant, and open feedback is displayed on a shared screen that can be viewed by all participants. The chat functions of Microsoft Teams or Zoom are used.

[0209] Step 6: Detect and flag coercive language

[0210] The server performs real-time sentiment analysis of comments and detects content that is deemed to be overbearing or harassing. The input is text data, and the output is a warning message. If overbearing or harassing comments are detected, a warning message is generated and sent to the relevant participant's device. An appropriate warning message is created using a generative AI model.

[0211] The device displays a warning message to the user in real time, specifically by using the notification function of the web browser, and prompts the user to take appropriate action.

[0212] Step 7: Generate and share meeting minutes

[0213] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting. The input is all comments and feedback during the meeting, and the output is the minutes data.

[0214] The device displays the generated minutes to the user, provides a download link and a sharing link for sharing with all participants, and is configured to output the minutes in Microsoft Word or Google Docs format.

[0215] (Application example 1)

[0216] 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."

[0217] The present invention aims to improve the quality and security of comments made in meetings, provide a fair and safe environment, and provide a system that allows for post-conference review and auditing. In particular, it is necessary to strengthen meeting security by detecting coercive comments and harassment in real time and responding immediately. There is also a need for real-time feedback and automatic generation of meeting minutes so that comments made during meetings can be used as reference material later.

[0218] 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.

[0219] In this invention, the server includes a means for collecting participants' comments as audio data, a means for converting the audio data into text data, a means for analyzing the text data to generate real-time feedback, and a means for detecting comments containing specific keywords and generating a warning. This makes it possible to improve the quality of comments in real time during a meeting, and to detect and warn against overbearing comments and harassment. It is also possible to automatically record the contents of the meeting and generate materials that can be referenced later.

[0220] "Means for collecting participants' comments as audio data" refers to a device or system that captures and records the content of participants' comments in a conference room or on an online platform as audio data in real time.

[0221] The "means for converting voice data into text data" refers to a technology that converts collected voice data into text information using voice recognition technology and outputs it as text data that can be analyzed.

[0222] The "means for analyzing text data and generating real-time feedback" is a system that analyzes the converted text data using natural language processing technology and provides participants with immediate feedback.

[0223] A "means for providing feedback to participants" is a device or system that visually or audibly communicates the generated feedback to participants.

[0224] The "means for detecting statements containing specific keywords and generating warnings" refers to a system that analyzes statements made during a meeting, detects specific keywords (e.g., overbearing statements or harassment) in real time, and generates warning messages based on the detection results.

[0225] MODE FOR CARRYING OUT THE INVENTION

[0226] Program Overview

[0227] This invention provides a system that uses real-time feedback from AI to improve the quality and security of comments made during meetings. The system collects participants' comments as audio data, converts them into text data, and analyzes them. Based on the analysis results, it generates real-time feedback and provides it to participants. The system also has a function to detect comments containing specific keywords and generate a warning. Specific embodiments of the system are described in detail below.

[0228] System configuration

[0229] Hardware and Software

[0230] Microphone: A device that collects speech from conference participants and is connected to a terminal.

[0231] Device (PC, smartphone, smart glasses, head-mounted display, etc.): Receives audio data from the microphone and sends it to the server.

[0232] Server: Converts voice data into text data, analyzes the text data using natural language processing technology, and generates feedback. It also detects utterances containing specific keywords and generates warnings.

[0233] Speech recognition API (such as Google Speech Recognition API): An API for converting voice data into text data.

[0234] Natural language processing libraries (NLTK, VADER, etc.): Libraries for analyzing text data and assessing sentiment and content.

[0235] System Operation

[0236] 1. Collecting and transmitting participant comments:

[0237] The device collects participants' speech in real time through a connected microphone, and the collected voice data is sent to a server.

[0238] 2. Audio to text conversion:

[0239] The server converts the transmitted voice data into text information using a speech recognition API (Google Speech Recognition API).

[0240] 3. Text data analysis:

[0241] The server then analyzes the converted text data using natural language processing (NLTK, VADER) to evaluate the sentiment, tone, and specificity of the comments.

[0242] 4. Generating and Presenting Real-Time Feedback:

[0243] The server generates feedback based on the analysis results and sends it to the device. The device then presents the generated feedback to the participant in real time. For example, feedback such as "It would be more effective if you provided specific data" is displayed on the participant's screen.

[0244] 5. Detecting specific keywords and generating alerts:

[0245] The server detects comments containing specific keywords (e.g., overbearing, harassment), and immediately generates and sends a warning message to the device after the relevant comment is made. The warning is displayed to participants in real time.

[0246] Specific examples

[0247] If participant A says during a meeting, "I have some opinions about the direction this project should take, but I think they're too harsh," the system will collect this comment, convert the audio data into text data, and analyze it. Based on the analysis results, a warning such as "Negative comments have been detected" will be generated and displayed in real time on participant A's device.

[0248] Example prompts for generative AI models

[0249] Analyze the following statements in meetings to detect negative or positive sentiment and generate feedback: "I have some opinions about the direction we should take this project, but I think they're too harsh."

[0250] This system will improve the quality of speech in meetings, detect overbearing remarks and harassment, and provide a safe meeting environment.

[0251] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0252] Step 1:

[0253] Audio data collection

[0254] Subject: Terminal

[0255] Specific operation: The device collects the speech of meeting participants in real time via a microphone.

[0256] Input: Remarks by conference participants (audio data)

[0257] Output: Collected audio data

[0258] Data processing: The device captures the speech as digital audio data and prepares it for transmission to the server.

[0259] Step 2:

[0260] Converting audio data to text

[0261] Subject: Server

[0262] Specific operation: The server receives the voice data sent from the device and converts it into text data using the speech recognition API (Google Speech Recognition API).

[0263] Input: Collected audio data

[0264] Output: Converted text data

[0265] Data processing: Using voice recognition technology, the voice data is analyzed and text data is generated. The server stores the results of this conversion process in a related database.

[0266] Step 3:

[0267] Text data analysis

[0268] Subject: Server

[0269] Specific operation: The server analyzes the converted text data using natural language processing (NLP) techniques (NLTK, VADER).

[0270] Input: Text data

[0271] Output: Analysis results (emotion, tone, specificity, etc.)

[0272] Data processing: The text data is parsed, keywords are extracted, sentiment analysis is performed, and the specificity of the comments is evaluated. The server uses the analysis results to proceed to the next processing step.

[0273] Step 4:

[0274] Generate real-time feedback

[0275] Subject: Server

[0276] Specific behavior: The server generates real-time feedback based on the results of analyzing the text data and sends it to the device. For example, if a comment lacks specificity, the server generates feedback such as, "It would be more effective if you provided specific data."

[0277] Input: Analysis results

[0278] Output: Real-time feedback

[0279] Data processing: Based on the analysis results, feedback messages are generated and sent to the device instructing it to display. The server places particular emphasis on oppressive or negative comments.

[0280] Step 5:

[0281] Providing feedback

[0282] Subject: Terminal

[0283] Specific operation: The device receives the feedback sent from the server and presents it to the participants. Closed feedback is presented to specific participants, and open feedback is presented to all participants.

[0284] Input: Real-time feedback

[0285] Output: Displayed feedback

[0286] Data Calculation: The terminal visually displays feedback messages and prompts specific participants to take appropriate action.

[0287] Step 6:

[0288] Detecting specific keywords and generating alerts

[0289] Subject: Server

[0290] Specific operation: The server detects specific keywords (e.g., overbearing, harassment) in the analyzed text data, and if detected, immediately generates a warning message and sends it to the terminal.

[0291] Input: Analysis results of text data

[0292] Output: Warning message

[0293] Data calculation: The server checks the analysis results against a specific keyword list, and if there is a match, generates and sends a warning message.

[0294] Step 7:

[0295] Present a warning message

[0296] Subject: Terminal

[0297] Specific operation: The terminal receives the warning message sent from the server and displays it to the relevant participant.

[0298] Input: warning message

[0299] Output: The displayed warning message

[0300] Data calculation: The terminal visually displays a warning message and prompts participants who make overbearing remarks to take appropriate action.

[0301] This processing flow enables the system to improve speech quality in real time and enhance meeting security.

[0302] 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.

[0303] Program Overview

[0304] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. Furthermore, this system includes "means for collecting participants' comments as audio data," "means for converting audio data into text data," "means for generating real-time feedback by analyzing the text data," "means for providing feedback to participants," "means for detecting overbearing comments and harassment and issuing an immediate warning," as well as "an emotion engine that recognizes the user's emotions."

[0305] Program processing

[0306] The program of the system of the present invention is processed as follows.

[0307] 1. Audio data collection:

[0308] The device collects speech from meeting participants as audio data in real time, specifically including the ability to capture microphone input from each participant.

[0309] 2. Audio to text conversion:

[0310] The server receives the voice data sent from the device and converts it into text data using a voice recognition API, etc. This converts the voice into text information that can be analyzed.

[0311] 3. Text data analysis:

[0312] The server then analyzes the converted text data using natural language processing (NLP) technology, analyzing keywords, emotions, tone, and other elements within the text to assess the specificity and relevance of the comments.

[0313] 4. Use of Emotion Engine:

[0314] The server then uses an emotion engine to recognize the user's emotions from the text data, extracting emotions from the content and tone of the speech and classifying them into emotion categories such as positive, negative, and neutral.

[0315] 5. Generate real-time feedback:

[0316] The server generates real-time feedback based on the text data and the results of the emotion engine, including closed and open feedback, in an appropriate format.

[0317] 6. Providing Feedback:

[0318] The terminal receives the generated feedback and provides it to the user: closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window that can be viewed by all participants.

[0319] 7. Detect and warn against aggressive speech:

[0320] The server performs real-time emotional analysis of comments and detects statements that are deemed to be overbearing or harassing.

[0321] If the server detects such a situation, it immediately generates a warning message and sends it to the terminal of the participant in question.

[0322] The terminal displays a warning message to the user in real time, prompting the user to take appropriate action.

[0323] 8. Generate and share meeting minutes:

[0324] The server records all comments and feedback during the meeting, and automatically generates minutes after the meeting and provides them to all participants.

[0325] The device displays the generated minutes to the user and provides a link that can be downloaded or shared.

[0326] Specific examples

[0327] Meeting Management

[0328] User (Person A): "I have some opinions about the direction this project should take."

[0329] The device captures Mr. A's speech as audio data and sends it to the server.

[0330] The server converts the voice data into text data and analyzes the content using NLP technology and an emotion engine.

[0331] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[0332] The server also detects negative emotions in Mr. A's comments and provides open feedback such as, "We recommend that you speak with more confidence."

[0333] After receiving the feedback, Person A makes a statement that provides specific data and continues speaking with even more confidence.

[0334] The terminal recaptures the new utterance and transmits it to the server.

[0335] The server again analyzes the new utterance and provides feedback if necessary.

[0336] Through this process, the system provides a fair and effective meeting environment in real time, and by utilizing an emotion engine, it also appropriately manages the emotions of participants. Furthermore, by automatically generating meeting minutes, the content of the meeting can be easily reviewed even after it has ended.

[0337] The processing flow will be explained below.

[0338] MODE FOR CARRYING OUT THE INVENTION

[0339] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. Furthermore, this system includes an "emotion engine" that recognizes the user's emotions and can provide feedback based on their emotions.

[0340] Specific processing steps

[0341] Step 1:

[0342] A user schedules a meeting.

[0343] Users log into the system and set the meeting date and time, participant list, and meeting goal.

[0344] Step 2:

[0345] The server creates a meeting room.

[0346] The server receives the request from the user, generates a unique meeting room ID, and creates a join link.

[0347] The server is pre-loaded and ready with the AI ​​models and emotion engines that will be used for the meeting.

[0348] Step 3:

[0349] The user invites participants.

[0350] The user sends the generated meeting link to prospective participants, who then click the link to enter the meeting room.

[0351] Step 4:

[0352] The device captures audio input.

[0353] The device captures microphone input each time a participant speaks and transmits it to the server as audio data in real time.

[0354] Step 5:

[0355] The server converts the voice data into text data.

[0356] The server converts the received voice data into text data using a speech recognition API, which makes it possible to analyze the spoken content as text information.

[0357] Step 6:

[0358] The server analyzes the text data.

[0359] The server then analyzes the converted text data using natural language processing (NLP) techniques, specifically extracting keywords, analyzing sentiment, and evaluating tone.

[0360] Step 7:

[0361] The server uses an emotion engine to recognize the user's emotion.

[0362] The server then uses an emotion engine to recognize the user's emotions from the text data, extracting emotions from the content and tone of the speech and classifying them into emotion categories such as positive, negative, and neutral.

[0363] Step 8:

[0364] The server generates real-time feedback.

[0365] The server generates real-time feedback based on the analysis of the text data and the emotion engine. The feedback can be either closed or open, and is provided in an appropriate format.

[0366] Step 9:

[0367] The device displays feedback.

[0368] The terminal displays the generated feedback in a chat window for a specific user (closed feedback), and also displays it in a feedback window for all participants (open feedback).

[0369] Step 10:

[0370] The server monitors and detects coercive remarks.

[0371] The server monitors all comments in real time and performs sentiment analysis.

[0372] Detects statements that are deemed to be overbearing or harassment.

[0373] Step 11:

[0374] The server will warn you against overbearing comments.

[0375] When an overbearing remark is detected, the server generates a warning message for the user and sends it to the terminal.

[0376] Step 12:

[0377] The terminal displays a warning message.

[0378] The device will display a warning message to the affected user in real time, urging them to reconsider their comments.

[0379] Step 13:

[0380] The server records your comments and feedback.

[0381] The server records all comments and feedback during the meeting and stores them in a database.

[0382] Step 14:

[0383] The server generates the minutes.

[0384] After the meeting ends, the server automatically generates minutes based on the recorded data.

[0385] Step 15:

[0386] The server shares the minutes.

[0387] The server provides the generated minutes to all participants via email or link.

[0388] Step 16:

[0389] The device will allow you to view and download the minutes.

[0390] The terminal provides a link to allow the user to view and download the minutes.

[0391] Example 2

[0392] 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."

[0393] In traditional online and offline meetings, some participants find it difficult to speak up, and the environment is prone to overbearing remarks and harassment. Furthermore, there is a lack of means to properly record what is being said and provide feedback in real time. As a result, it is difficult to achieve fair discussions.

[0394] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0395] In this invention, the server includes means for collecting participants' comments as voice data, means for converting the voice data into text data, means for analyzing the text data to generate real-time feedback, means for providing the feedback to participants, and means for recognizing user emotions through sentiment analysis, thereby providing a fair and effective meeting environment, appropriately managing user emotions through sentiment analysis, and improving the quality of discussions.

[0396] "Means for collecting participants' speech as audio data" refers to a device or system that records the speech of meeting participants as audio data in real time. A specific example is a device with a microphone.

[0397] "Means for converting voice data into text data" refers to software or algorithms for analyzing voice data and converting it into text data. A specific example is a voice recognition API.

[0398] "Means for analyzing text data and generating real-time feedback" refers to natural language processing technologies and algorithms for evaluating participants' comments based on text data and generating instant feedback.

[0399] "Means for providing feedback to participants" refers to a device or system that displays or notifies participants of generated feedback in real time. Examples include a chat window or a feedback window.

[0400] "Means for recognizing user emotions through sentiment analysis" refers to software or algorithms that extract participants' emotions from text data and classify and recognize those emotions. A specific example is an emotion engine.

[0401] This invention is a system that uses AI to provide real-time feedback in online and offline meetings. This system creates an environment where all participants can easily express their opinions and ensure fair discussions. Furthermore, this system includes a function to recognize user emotions through sentiment analysis.

[0402] The specific configuration of this system is as follows.

[0403] Hardware and software used

[0404] This system is realized by the following hardware and software.

[0405] Hardware: A device with a microphone (e.g., a computer, smartphone, etc.) used to collect the speech of meeting participants.

[0406] software:

[0407] A speech recognition API for converting voice data into text data (e.g., Google Cloud Speech-to-Text API).

[0408] Natural language processing techniques (e.g., TensorFlow, Natural Language Toolkit (NLTK)) for analyzing text data.

[0409] Software for performing sentiment analysis (e.g., IBM Watson Tone Analyzer).

[0410] Processing flow

[0411] 1. Audio data collection:

[0412] The terminal collects speech from meeting participants as audio data in real time. Specifically, a microphone-equipped device captures speech from participants and transmits it to a server.

[0413] 2. Audio to text conversion:

[0414] The server receives the voice data sent from the device and converts it into text data using the Google Cloud Speech-to-Text API.

[0415] 3. Text data analysis:

[0416] The server analyzes the converted text data using natural language processing technology (TensorFlow and Natural Language Toolkit (NLTK)). This analyzes keywords, emotions, tone, etc. in the text and evaluates the specificity and relevance of the statements.

[0417] 4. Use of Emotion Engine:

[0418] The server also uses an emotion engine (IBM Watson Tone Analyzer) to recognize the user's emotions and classify them into emotion categories such as positive, negative, and neutral.

[0419] 5. Generate real-time feedback:

[0420] The server generates real-time feedback based on the collected data and sentiment analysis results, including closed and open feedback.

[0421] 6. Providing Feedback:

[0422] The terminal receives the generated feedback and provides it to the user. Closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window that can be viewed by all participants.

[0423] 7. Detect and warn against aggressive speech:

[0424] The server analyzes the sentiment of comments in real time and detects any statements that are deemed to be overbearing or harassing. If any are detected, a warning message is immediately generated and sent to the relevant participant's device.

[0425] 8. Generate and share meeting minutes:

[0426] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. The device displays the minutes to the user and provides a link that can be downloaded or shared.

[0427] Specific examples

[0428] As a concrete example, we will show how a meeting proceeds.

[0429] User (Person A): "I have some opinions about the direction this project should take."

[0430] The device captures Mr. A's speech as audio data and sends it to the server.

[0431] The server converts the voice data into text data and analyzes the content using natural language processing technology and an emotion engine.

[0432] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[0433] The server also detects negative emotions in Mr. A's comments and provides open feedback such as, "We recommend that you speak with more confidence."

[0434] Person A receives the feedback, makes statements that provide specific data, and continues to speak with confidence.

[0435] The terminal recaptures the new utterance and transmits it to the server.

[0436] The server then analyzes new comments and provides feedback as necessary. Through this process, the system provides a fair and effective meeting environment in real time, and by utilizing an emotion engine, it also appropriately manages participants' emotions. Furthermore, by automatically generating meeting minutes, the content can be easily reviewed even after the meeting has ended.

[0437] Prompt Sentence Examples

[0438] Prompt: Type what you want to say next in the meeting. The system will provide real-time feedback.

[0439] User says: "I have an opinion on the direction of the project."

[0440] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0441] Step 1:

[0442] Audio data collection

[0443] Input: User's speech (audio)

[0444] Specific operation: Capture participants' speech through a microphone connected to the device.

[0445] Data processing: The captured audio data is converted into a digital signal.

[0446] Output: Digital audio data

[0447] Detailed explanation: The device collects speech from meeting participants as audio data in real time and sends it to a server. For example, a microphone on a laptop installed in a conference room captures each participant's speech and sends the data to the server via Wi-Fi.

[0448] Step 2:

[0449] Converting audio data to text

[0450] Input: Digital audio data

[0451] Specific operation: The server passes the received voice data to the voice recognition API.

[0452] Data processing: Use the Google Cloud Speech-to-Text API to convert the audio data into text data.

[0453] Output: Text data

[0454] Detailed explanation: The server passes the digital voice data sent from the device to the Google Cloud Speech-to-Text API, which converts the voice data into text data. For example, a speech saying "I have an opinion about the direction of the project" is converted into text data saying "I have an opinion about the direction of the project."

[0455] Step 3:

[0456] Text data analysis

[0457] Input: Text data

[0458] Specific operation: The server passes the converted text data to natural language processing technology.

[0459] Data processing: Analyze text for keywords, tone, sentiment, etc. using TensorFlow and the Natural Language Toolkit (NLTK).

[0460] Output: Analysis result data (keywords, tone, sentiment, etc.)

[0461] Detailed explanation: The server uses natural language processing (NLP) technology to analyze text data. For example, keywords are extracted from the text "I have an opinion on the direction of the project" and evaluated as "lacking specificity."

[0462] Step 4:

[0463] Using the Emotion Engine

[0464] Input: Analysis result data (keywords, tone, emotion, etc.)

[0465] Specific operation: The server uses IBM Watson Tone Analyzer to extract and classify emotions.

[0466] Data Processing: The sentiment engine extracts sentiment categories such as positive, negative, and neutral from the analyzed data.

[0467] Output: Sentiment analysis result data

[0468] Detailed explanation: The server further analyzes the parsed text data with a sentiment engine. For example, the phrase "I have an opinion" is classified as negative.

[0469] Step 5:

[0470] Generate real-time feedback

[0471] Input: Analysis result data and sentiment analysis result data

[0472] Specific operation: The server generates real-time feedback based on the collected data.

[0473] Data processing: Generating feedback in the form of closed and open feedback.

[0474] Output: Feedback data

[0475] Detailed explanation: The server generates specific feedback based on the analysis results and sentiment analysis results. For example, closed feedback such as "It would be effective if you provided specific data" is generated, and open feedback such as "We recommend that you speak more confidently."

[0476] Step 6:

[0477] Providing Feedback

[0478] Input: Feedback data

[0479] Specific operation: The terminal displays the generated feedback to the user.

[0480] Data processing: The feedback data is converted into data for display on the user interface.

[0481] Output: Display of feedback message

[0482] Detailed description: The terminal provides the user with real-time feedback sent from the server. For example, closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window visible to all participants.

[0483] Step 7:

[0484] Detecting and warning against coercive speech

[0485] Input: Text data and sentiment analysis result data

[0486] What it does: The server applies rules to detect intrusive and harassing comments.

[0487] Data processing: Detects overbearing remarks and harassment based on sentiment analysis data and generates warning messages.

[0488] Output: Warning message

[0489] Detailed explanation: The server detects overbearing remarks and harassment in real time based on the results of sentiment analysis. For example, if a remark such as "That's impossible!" is judged to be overbearing, a warning message "Please avoid overbearing remarks" is generated and sent to the relevant participant's device.

[0490] Step 8:

[0491] Generate and share meeting minutes

[0492] Input: Speech and feedback data for the entire meeting

[0493] Specific operation: The server records all data during the meeting and generates minutes.

[0494] Data processing: Automatically generate minutes from recorded data and create a link for sharing.

[0495] Output: Meeting minutes data and shared link

[0496] Detailed explanation: The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. For example, after the meeting ends, participants receive an email saying, "All the contents of the meeting have been summarized," and the minutes are displayed on their devices. A link that can be downloaded or shared is also provided.

[0497] This system provides a fair and effective meeting environment, manages participants' emotions appropriately, and improves the quality of discussions. Automatically generating meeting minutes also makes it easy to review important information.

[0498] (Application example 2)

[0499] 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."

[0500] In physical stores, staff meetings and training sessions often lack the fairness and specificity of statements and proper feedback. It's also difficult to detect coercive remarks or harassment in real time and respond immediately, and there's no system for automatically generating and providing meeting minutes to participants. As a result, fair and effective discussions and organizational management can be difficult.

[0501] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0502] In this invention, the server includes means for collecting participants' comments as voice data, means for converting the voice data into text data, means for analyzing the text data to generate real-time feedback, means for providing the feedback to participants, means for recognizing emotions in the text data, means for generating feedback based on the recognized emotions, means for detecting coercive remarks and harassment and issuing immediate warnings, and means for automatically generating meeting minutes and providing them to participants. This makes meetings and training sessions in brick-and-mortar stores fair and effective, and enables real-time feedback, detection and warning of coercive remarks, and automatic generation of meeting minutes.

[0503] "Means for collecting participants' speech as audio data" refers to devices or systems that record the speech of participants in conferences or meetings in real time and store it as digital audio data.

[0504] "Means for converting voice data into text data" refers to a device or system that converts collected voice data into text information using natural language processing technology or voice recognition technology.

[0505] The "means for analyzing text data and generating real-time feedback" refers to a device or system for analyzing the converted text data and generating instantaneous feedback based on the content and sentiment.

[0506] The "means for providing feedback to participants" refers to a device or system for displaying or notifying the participants of the conference or meeting of the generated feedback in an appropriate format.

[0507] A "means for recognizing emotions in text data" is a device or system that uses natural language processing or machine learning techniques to extract and classify emotions from the content and tone of text data.

[0508] The "means for generating feedback based on recognized emotions" refers to a device or system that generates feedback appropriate to the emotions in real time based on analyzed emotional data.

[0509] "Means for detecting coercive remarks and harassment and issuing immediate warnings" refers to devices or systems that automatically detect parts of a participant's remarks that are deemed to be coercive or harassing and immediately display a warning.

[0510] "Means for automatically generating meeting minutes and providing them to participants" refers to a device or system that records all comments and feedback made during a meeting, and automatically creates and distributes minutes to participants after the meeting ends.

[0511] This invention aims to realize a system that uses smart devices and a server to provide real-time feedback for staff meetings and training sessions in brick-and-mortar stores. Specific embodiments of this system are described below.

[0512] The system uses a microphone built into a smart device (e.g., a smartphone or tablet) to collect participants' speech as audio data. After the speech is collected, the audio data is transferred from the smart device to a server.

[0513] The server uses speech recognition software to convert the voice data into text data. This can be effectively done using existing speech recognition technology, such as Google's speech recognition API. The server then analyzes the text data using natural language processing (NLP) techniques. For this purpose, the Transformers library provided by Hugging Face can be used.

[0514] The parsed text data is then analyzed using an emotion engine (e.g., nlptown / bert-base-multilingual-uncased-sentiment model) to recognize emotions. The server uses this emotion data to generate real-time feedback and provides it to participants. The feedback can be displayed on the screen of a smart device or announced via audio.

[0515] Furthermore, the server uses text data and emotional data to detect overbearing remarks and harassment. If detected, it generates feedback to immediately warn the relevant participants and notifies them in real time. All remarks and feedback are also recorded, and minutes are automatically created and shared with all participants after the meeting ends.

[0516] For example, if someone says during a meeting, "I have an opinion on the direction of this project. Next time, let's present more concrete data. Also, speaking based on facts will be more persuasive," the system will capture this statement, convert it from audio data to text data, and generate real-time feedback along with emotion recognition.

[0517] As described above, this system can be used to ensure that meetings and training sessions in physical stores are conducted fairly and effectively, providing an environment in which participants' opinions are properly reflected.

[0518] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0519] Step 1:

[0520] When the device detects the start of a meeting, it collects each participant's speech as audio data. At this time, it uses the smart device's microphone to capture the audio in real time. The input is the participant's voice, and the output is digital audio data. This data is immediately sent to the server.

[0521] Step 2:

[0522] The server converts the received voice data into text data using Google's speech recognition API. The input is digital voice data, and the output is text data converted from this voice data into character string information. This conversion turns the voice information into analyzable text information.

[0523] Step 3:

[0524] The server analyzes the text data and uses natural language processing (NLP) technology to analyze the content. Specifically, it extracts keywords, tone, and utterance relationships within the text. The input is text data generated by a speech recognition API, and the output is the analysis results. Emotion recognition of the text data is also performed based on the analysis results.

[0525] Step 4:

[0526] The server uses an emotion engine (e.g., nlptown / bert-base-multilingual-uncased-sentiment model) to recognize the emotion of text data. The input is text data, and the output is recognized emotion data. This emotion data is classified into categories such as positive, negative, and neutral.

[0527] Step 5:

[0528] The server generates real-time feedback based on the emotion data and text analysis results. The feedback can be in the form of closed feedback (only for specific participants) or open feedback (for all participants). The input is emotion data and text analysis results, and the output is a feedback message.

[0529] Step 6:

[0530] The device receives the generated feedback and provides it to the participants. Closed feedback is displayed on a specific participant's device, and open feedback is displayed on all participants' devices. The input is the feedback message, and the output is a display of the feedback.

[0531] Step 7:

[0532] The server monitors participants' comments in real time and detects overbearing comments or harassment. If any are detected, a warning message is immediately generated and sent to the relevant participant. The input is text data and emotional data, and the output is a warning message.

[0533] Step 8:

[0534] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. The generated minutes are shared with all participants. The input is all comment data and feedback data, and the output is the generation and sharing of meeting minutes.

[0535] 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.

[0536] 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.

[0537] 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.

[0538] [Second embodiment]

[0539] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0540] 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.

[0541] 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).

[0542] 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.

[0543] 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.

[0544] 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).

[0545] 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.

[0546] 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.

[0547] 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.

[0548] 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.

[0549] 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.

[0550] 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."

[0551] Program Overview

[0552] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. This system includes "means for collecting participants' comments as audio data," "means for converting the audio data into text data," "means for analyzing the text data and generating real-time feedback," "means for providing feedback to participants," and "means for detecting overbearing comments and harassment and issuing an immediate warning."

[0553] Program processing

[0554] The program of the system of the present invention is processed as follows.

[0555] 1. Audio data collection:

[0556] The device collects speech from meeting participants as audio data in real time, specifically including the ability to capture microphone input from each participant.

[0557] 2. Audio to text conversion:

[0558] The server receives the voice data sent from the device and converts it into text data using a voice recognition API, etc. This makes it possible to analyze the content of the speech as text information.

[0559] 3. Text data analysis:

[0560] The server then analyzes the converted text data using natural language processing (NLP) technology, analyzing keywords, emotions, tone, and other elements within the text to assess the specificity and relevance of the comments.

[0561] 4. Generate real-time feedback:

[0562] The server generates real-time feedback based on the analysis results, including closed and open feedback, in an appropriate format.

[0563] 5. Providing Feedback:

[0564] The terminal receives the generated feedback and provides it to the user. Closed feedback is displayed in the chat window of a specific participant, while open feedback is displayed so that all participants can see it.

[0565] 6. Detect and warn against aggressive speech:

[0566] The server performs real-time emotional analysis of comments and detects statements that are deemed to be overbearing or harassing.

[0567] If the server detects such a situation, it immediately generates a warning message and sends it to the terminal of the participant in question.

[0568] The terminal displays a warning message to the user in real time, prompting the user to take appropriate action.

[0569] 7. Generate and share meeting minutes:

[0570] The server records all comments and feedback during the meeting, and automatically generates minutes after the meeting and provides them to all participants.

[0571] The device displays the generated minutes to the user and provides a link that can be downloaded or shared.

[0572] Specific examples

[0573] Meeting Management

[0574] User (Person A): "I have some opinions about the direction this project should take."

[0575] The device captures Mr. A's speech as audio data and sends it to the server.

[0576] The server converts the voice data into text data and analyzes the content using NLP technology.

[0577] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[0578] Person A receives the feedback and makes a statement providing specific data.

[0579] The device recaptures Mr. A's new comments and sends them to the server.

[0580] The server analyzes the new utterance and provides further feedback.

[0581] In this way, the system provides a fair and effective meeting environment in real time, ensuring that all participants feel comfortable speaking up. Furthermore, the system automatically generates meeting minutes, making it easy to review the content of meetings even after they have ended.

[0582] The processing flow will be explained below.

[0583] Program processing flow

[0584] Step 1: Set up a meeting

[0585] 1. A user schedules a meeting.

[0586] Users log into the system and set the meeting date and time, participant list, and meeting goal.

[0587] 2. The server creates a meeting room.

[0588] The server receives the request from the user, generates a unique meeting room ID, and creates a join link.

[0589] The server pre-loads and prepares the AI ​​models to be used in the meeting.

[0590] 3. The user invites participants.

[0591] The user sends the generated meeting link to the participants, who click the link to enter the meeting room.

[0592] Step 2: Collecting audio data

[0593] 4. The device captures the audio input.

[0594] The device captures microphone input each time a participant speaks and transmits it to the server as audio data in real time.

[0595] Step 3: Convert audio data to text

[0596] 5. The server converts the audio data into text data.

[0597] The server converts the received voice data into text data using a speech recognition API, which converts the voice into text information that can be analyzed.

[0598] Step 4: Analyze the text data

[0599] 6. The server analyzes the text data.

[0600] The server then analyzes the converted text data using natural language processing (NLP) techniques, specifically extracting keywords, analyzing sentiment, and evaluating tone.

[0601] Step 5: Generate real-time feedback

[0602] 7. The server generates real-time feedback.

[0603] Based on the analysis results, the server generates either crowded or open feedback. Closed feedback is given to specific participants, while open feedback is given to all participants.

[0604] 8. The device will display feedback.

[0605] The terminal displays closed feedback to a specific user in a chat window, and displays open feedback in a feedback window that can be viewed by all participants.

[0606] Step 6: Detect and flag coercive language

[0607] 9. The server monitors the comments.

[0608] The server monitors all comments in real time and performs sentiment analysis.

[0609] 10. The server detects overbearing comments.

[0610] The server detects overbearing remarks and harassment based on the results of sentiment analysis.

[0611] 11. The server issues a warning.

[0612] If the server detects an overbearing remark, it generates and sends a warning message to the user.

[0613] 12. The terminal displays a warning message.

[0614] The device will display a warning message to the affected user in real time, urging them to reconsider their comments.

[0615] Step 7: Generate and share meeting minutes

[0616] 13. The server records your comments and feedback.

[0617] The server records all comments and feedback during the meeting and stores them in a database.

[0618] 14. The server generates the minutes.

[0619] The server automatically generates minutes based on the data recorded after the meeting ends.

[0620] 15. The server shares the minutes.

[0621] The server provides the generated minutes to all participants via email or link.

[0622] 16. The device will be able to display and download the minutes.

[0623] The terminal provides a function that allows the user to view and download the minutes.

[0624] Example: Feedback process

[0625] User (Person A): "I have some opinions about the direction this project should take."

[0626] The device captures A's remarks and sends them to the server.

[0627] The server converts the voice data into text data and analyzes the content using NLP technology.

[0628] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or specific data."

[0629] Person A receives the feedback and makes a statement providing specific data.

[0630] The device recaptures the new utterance and sends it to the server.

[0631] The server analyzes the new utterance and provides further feedback.

[0632] Through this process, the system provides a fair and effective meeting environment in real time.

[0633] Example 1

[0634] 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."

[0635] In online and offline meetings, it is necessary to create an environment where all participants can express their opinions fairly and easily, and to prevent discussions from being biased towards certain individuals. There is also a need for systems that can detect overbearing remarks and harassment in real time and respond immediately. Furthermore, to efficiently conduct meetings, it is desirable to provide real-time feedback and automatically generate meeting minutes. However, conventional systems have not been able to adequately resolve these issues.

[0636] 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.

[0637] In this invention, the server includes means for collecting participants' comments as audio data, means for converting the audio data into text data, means for analyzing the text data using natural language processing technology, means for generating real-time feedback based on the analysis results, means for providing the feedback to participants, means for detecting overbearing comments or harassment and generating a warning message, and means for recording the comments and feedback of the meeting and generating minutes after the meeting. This provides an environment in which all participants can speak fairly and easily, making it possible to provide effective feedback in real time. Furthermore, overbearing comments and harassment can be addressed immediately, and the automatically generated minutes make it easy to review the content of the discussion.

[0638] "Participant" means an individual who participates in a meeting or conference.

[0639] "Audio data" refers to data that has been digitized from audio recorded as sound waves.

[0640] "Text data" is data that has been digitized as character information.

[0641] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[0642] "Real-time feedback" refers to instant assessment and advice provided during a meeting.

[0643] "Positive speech" is speech that sounds intimidating or aggressive to others.

[0644] "Harassment" refers to behavior or speech that harasses or offends others.

[0645] A "warning message" is a notification that warns you against certain actions or statements.

[0646] "Minutes" are a written record of all statements and discussions made during a meeting or conference.

[0647] "Means of collection" refers to devices or software that capture participants' voices as digital data.

[0648] "Means for converting" refers to a technology or system that converts voice data into text data.

[0649] "Means of analysis" refers to the technology or system used to analyze text data and determine meaning and sentiment.

[0650] "Means for generating" refers to a technology or system that generates feedback or warning messages based on the analysis results.

[0651] "Means for providing" refers to the equipment or software used to deliver generated feedback and warning messages to participants.

[0652] "Means of recording and generating" refers to technology or systems that store comments and feedback during meetings and later compile them into minutes.

[0653] A "system" is a set of devices or software that integrates these means to achieve a set of functions.

[0654] MODE FOR CARRYING OUT THE INVENTION

[0655] This invention is a system that uses AI to provide real-time feedback during online and offline meetings, creating an environment where all participants can easily express their opinions. This system includes "means for collecting participants' comments as audio data," "means for converting the audio data into text data," "means for analyzing the text data using natural language processing technology," "means for generating real-time feedback based on the analysis results," "means for providing feedback to participants," "means for detecting overbearing comments and harassment and generating a warning message," and "means for recording comments and feedback during meetings and generating meeting minutes after the meeting."

[0656] Specifically, the process is as follows:

[0657] Audio data collection

[0658] The device collects speech from each participant during the meeting as real-time audio data using a built-in microphone or an external microphone, and transmits the collected audio data to a server in a compressed format (e.g., MP3 or WAV format).

[0659] Converting audio data to text

[0660] The server receives the voice data sent from the device and converts it into text using a speech recognition API such as Google Cloud Speech-to-Text API or IBM Watson Speech to Text, then temporarily stores the converted text in a database.

[0661] Text data analysis

[0662] The server analyzes the stored text data using natural language processing (NLP) technology. Specifically, it uses Google Cloud Natural Language API and spaCy to evaluate keywords, sentence tone, and sentiment within the text. It evaluates the specificity and relevance of the text data and generates closed-form feedback.

[0663] Generate real-time feedback

[0664] The server generates real-time feedback based on the analysis results, classifies the feedback into closed feedback (for specific participants) and open feedback (for all participants), and sends it to the terminal in the appropriate format.

[0665] Providing Feedback

[0666] The device receives the feedback sent from the server, and displays the closed feedback in the chat window of the specific participant, while the open feedback is displayed on a screen that can be viewed by all participants. Platforms such as Microsoft Teams and Zoom may be used.

[0667] Detecting and warning against coercive speech

[0668] The server analyzes the sentiment of comments in real time and detects any overbearing or harassing content. If any are detected, a warning message is generated and sent to the relevant participant's device.

[0669] The device displays a warning message to the user in real time and prompts the user to take appropriate action, typically using the notification function of the web browser.

[0670] Generate and share meeting minutes

[0671] The server records all comments and feedback during the meeting, automatically generates minutes after the meeting, and generates a link to share the minutes with all participants.

[0672] The device displays the generated minutes to the user and provides download and sharing links, and can output them in Microsoft Word and Google Docs formats.

[0673] Specific examples

[0674] Meeting Management

[0675] User (Person A): "I have some opinions about the direction this project should take."

[0676] The device captures A's speech as audio data and sends it to the server. The captured audio data is saved in WAV format.

[0677] The server uses the Google Cloud Speech-to-Text API to convert Mr. A's voice data into text data in real time. The converted text data is recorded as "I have an opinion about the direction of this project."

[0678] The server analyzes the text and determines that it is "not specific enough" - NLP analysis detects that it does not contain specific keywords.

[0679] The server generates closed feedback such as "It would be more effective if you provided examples or concrete data."

[0680] The device will display this feedback in Mr. A's individual chat window.

[0681] Examples of problematic statements

[0682] User B makes a commanding statement: "Your idea is pointless."

[0683] The device captures Mr. B's speech as audio data and sends it to the server.

[0684] The server converts the voice data into text data and performs sentiment analysis using NLP technology. The analysis results indicate that the remark is "overbearing."

[0685] The server generates a warning message saying, "An overbearing remark has been detected. Please choose appropriate words." and sends it to Mr. B's device.

[0686] The device will notify Mr. B of this warning message in real time, urging him to be careful.

[0687] Example prompts for generative AI models

[0688] Please analyze the following text data and determine whether it constitutes coercive remarks or harassment.

[0689] Data Text: "That suggestion makes absolutely no sense. Anyone could have thought of something like that."

[0690] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0691] Program processing flow

[0692] Step 1: Collecting audio data

[0693] The device collects speech from each participant during a meeting as audio data in real time. The input is the participant's voice, and the output is audio data. Specifically, the device captures audio using a built-in microphone or an external microphone. The collected audio data is temporarily saved in WAV format and sent to a server.

[0694] Step 2: Convert audio data to text

[0695] The server receives the voice data sent from the device. The input is WAV format voice data, and the output is text data. Specifically, the voice data is converted into text data using the Google Cloud Speech-to-Text API or IBM Watson Speech to Text. At this time, the voice data is converted into text data and temporarily stored in a database.

[0696] Step 3: Analyze the text data

[0697] The server analyzes the stored text data using natural language processing (NLP) technology. The input is text data, and the output is the analysis results (keywords, tone, and emotional information). Specifically, it uses Google Cloud Natural Language API and spaCy to extract keywords from the text data, evaluate tone, and analyze emotions. Based on the results of this analysis, the specificity and relevance of the text are evaluated, and feedback candidates are generated.

[0698] Step 4: Generate real-time feedback

[0699] The server generates real-time feedback based on the analysis results. The input is the analysis results, and the output is the feedback message. The server classifies the feedback into closed feedback (for specific participants) and open feedback (for all participants) formats and sends it to the terminal in the appropriate format. The generative AI model is used to generate the feedback text.

[0700] Step 5: Provide feedback

[0701] The device receives feedback sent from the server. The input is the feedback message, and the output is the feedback provided to the user. Specifically, closed feedback is displayed in the chat window of a specific participant, and open feedback is displayed on a shared screen that can be viewed by all participants. The chat functions of Microsoft Teams or Zoom are used.

[0702] Step 6: Detect and flag coercive language

[0703] The server performs real-time sentiment analysis of comments and detects content that is deemed to be overbearing or harassing. The input is text data, and the output is a warning message. If overbearing or harassing comments are detected, a warning message is generated and sent to the relevant participant's device. An appropriate warning message is created using a generative AI model.

[0704] The device displays a warning message to the user in real time, specifically by using the notification function of the web browser, and prompts the user to take appropriate action.

[0705] Step 7: Generate and share meeting minutes

[0706] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting. The input is all comments and feedback during the meeting, and the output is the minutes data.

[0707] The device displays the generated minutes to the user, provides a download link and a sharing link for sharing with all participants, and is configured to output the minutes in Microsoft Word or Google Docs format.

[0708] (Application example 1)

[0709] 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."

[0710] The present invention aims to improve the quality and security of comments made in meetings, provide a fair and safe environment, and provide a system that allows for post-conference review and auditing. In particular, it is necessary to strengthen meeting security by detecting coercive comments and harassment in real time and responding immediately. There is also a need for real-time feedback and automatic generation of meeting minutes so that comments made during meetings can be used as reference material later.

[0711] 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.

[0712] In this invention, the server includes a means for collecting participants' comments as audio data, a means for converting the audio data into text data, a means for analyzing the text data to generate real-time feedback, and a means for detecting comments containing specific keywords and generating a warning. This makes it possible to improve the quality of comments in real time during a meeting, and to detect and warn against overbearing comments and harassment. It is also possible to automatically record the contents of the meeting and generate materials that can be referenced later.

[0713] "Means for collecting participants' comments as audio data" refers to a device or system that captures and records the content of participants' comments in a conference room or on an online platform as audio data in real time.

[0714] The "means for converting voice data into text data" refers to a technology that converts collected voice data into text information using voice recognition technology and outputs it as text data that can be analyzed.

[0715] The "means for analyzing text data and generating real-time feedback" is a system that analyzes the converted text data using natural language processing technology and provides participants with immediate feedback.

[0716] A "means for providing feedback to participants" is a device or system that visually or audibly communicates the generated feedback to participants.

[0717] The "means for detecting statements containing specific keywords and generating warnings" refers to a system that analyzes statements made during a meeting, detects specific keywords (e.g., overbearing statements or harassment) in real time, and generates warning messages based on the detection results.

[0718] MODE FOR CARRYING OUT THE INVENTION

[0719] Program Overview

[0720] This invention provides a system that uses real-time feedback from AI to improve the quality and security of comments made during meetings. The system collects participants' comments as audio data, converts them into text data, and analyzes them. Based on the analysis results, it generates real-time feedback and provides it to participants. The system also has a function to detect comments containing specific keywords and generate a warning. Specific embodiments of the system are described in detail below.

[0721] System configuration

[0722] Hardware and Software

[0723] Microphone: A device that collects speech from conference participants and is connected to a terminal.

[0724] Device (PC, smartphone, smart glasses, head-mounted display, etc.): Receives audio data from the microphone and sends it to the server.

[0725] Server: Converts voice data into text data, analyzes the text data using natural language processing technology, and generates feedback. It also detects utterances containing specific keywords and generates warnings.

[0726] Speech recognition API (such as Google Speech Recognition API): An API for converting voice data into text data.

[0727] Natural language processing libraries (NLTK, VADER, etc.): Libraries for analyzing text data and assessing sentiment and content.

[0728] System Operation

[0729] 1. Collecting and transmitting participant comments:

[0730] The device collects participants' speech in real time through a connected microphone, and the collected voice data is sent to a server.

[0731] 2. Audio to text conversion:

[0732] The server converts the transmitted voice data into text information using a speech recognition API (Google Speech Recognition API).

[0733] 3. Text data analysis:

[0734] The server then analyzes the converted text data using natural language processing (NLTK, VADER) to evaluate the sentiment, tone, and specificity of the comments.

[0735] 4. Generating and Presenting Real-Time Feedback:

[0736] The server generates feedback based on the analysis results and sends it to the device. The device then presents the generated feedback to the participant in real time. For example, feedback such as "It would be more effective if you provided specific data" is displayed on the participant's screen.

[0737] 5. Detecting specific keywords and generating alerts:

[0738] The server detects comments containing specific keywords (e.g., overbearing, harassment), and immediately generates and sends a warning message to the device after the relevant comment is made. The warning is displayed to participants in real time.

[0739] Specific examples

[0740] If participant A says during a meeting, "I have some opinions about the direction this project should take, but I think they're too harsh," the system will collect this comment, convert the audio data into text data, and analyze it. Based on the analysis results, a warning such as "Negative comments have been detected" will be generated and displayed in real time on participant A's device.

[0741] Example prompts for generative AI models

[0742] Analyze the following statements in meetings to detect negative or positive sentiment and generate feedback: "I have some opinions about the direction we should take this project, but I think they're too harsh."

[0743] This system will improve the quality of speech in meetings, detect overbearing remarks and harassment, and provide a safe meeting environment.

[0744] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0745] Step 1:

[0746] Audio data collection

[0747] Subject: Terminal

[0748] Specific operation: The device collects the speech of meeting participants in real time via a microphone.

[0749] Input: Remarks by conference participants (audio data)

[0750] Output: Collected audio data

[0751] Data processing: The device captures the speech as digital audio data and prepares it for transmission to the server.

[0752] Step 2:

[0753] Converting audio data to text

[0754] Subject: Server

[0755] Specific operation: The server receives the voice data sent from the device and converts it into text data using the speech recognition API (Google Speech Recognition API).

[0756] Input: Collected audio data

[0757] Output: Converted text data

[0758] Data processing: Using voice recognition technology, the voice data is analyzed and text data is generated. The server stores the results of this conversion process in a related database.

[0759] Step 3:

[0760] Text data analysis

[0761] Subject: Server

[0762] Specific operation: The server analyzes the converted text data using natural language processing (NLP) techniques (NLTK, VADER).

[0763] Input: Text data

[0764] Output: Analysis results (emotion, tone, specificity, etc.)

[0765] Data processing: The text data is parsed, keywords are extracted, sentiment analysis is performed, and the specificity of the comments is evaluated. The server uses the analysis results to proceed to the next processing step.

[0766] Step 4:

[0767] Generate real-time feedback

[0768] Subject: Server

[0769] Specific behavior: The server generates real-time feedback based on the results of analyzing the text data and sends it to the device. For example, if a comment lacks specificity, the server generates feedback such as, "It would be more effective if you provided specific data."

[0770] Input: Analysis results

[0771] Output: Real-time feedback

[0772] Data processing: Based on the analysis results, feedback messages are generated and sent to the device instructing it to display. The server places particular emphasis on oppressive or negative comments.

[0773] Step 5:

[0774] Providing feedback

[0775] Subject: Terminal

[0776] Specific operation: The device receives the feedback sent from the server and presents it to the participants. Closed feedback is presented to specific participants, and open feedback is presented to all participants.

[0777] Input: Real-time feedback

[0778] Output: Displayed feedback

[0779] Data Calculation: The terminal visually displays feedback messages and prompts specific participants to take appropriate action.

[0780] Step 6:

[0781] Detecting specific keywords and generating alerts

[0782] Subject: Server

[0783] Specific operation: The server detects specific keywords (e.g., overbearing, harassment) in the analyzed text data, and if detected, immediately generates a warning message and sends it to the terminal.

[0784] Input: Analysis results of text data

[0785] Output: Warning message

[0786] Data calculation: The server checks the analysis results against a specific keyword list, and if there is a match, generates and sends a warning message.

[0787] Step 7:

[0788] Present a warning message

[0789] Subject: Terminal

[0790] Specific operation: The terminal receives the warning message sent from the server and displays it to the relevant participant.

[0791] Input: warning message

[0792] Output: The displayed warning message

[0793] Data calculation: The terminal visually displays a warning message and prompts participants who make overbearing remarks to take appropriate action.

[0794] This processing flow enables the system to improve speech quality in real time and enhance meeting security.

[0795] 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.

[0796] Program Overview

[0797] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. Furthermore, this system includes "means for collecting participants' comments as audio data," "means for converting audio data into text data," "means for generating real-time feedback by analyzing the text data," "means for providing feedback to participants," "means for detecting overbearing comments and harassment and issuing an immediate warning," as well as "an emotion engine that recognizes the user's emotions."

[0798] Program processing

[0799] The program of the system of the present invention is processed as follows.

[0800] 1. Audio data collection:

[0801] The device collects speech from meeting participants as audio data in real time, specifically including the ability to capture microphone input from each participant.

[0802] 2. Audio to text conversion:

[0803] The server receives the voice data sent from the device and converts it into text data using a voice recognition API, etc. This converts the voice into text information that can be analyzed.

[0804] 3. Text data analysis:

[0805] The server then analyzes the converted text data using natural language processing (NLP) technology, analyzing keywords, emotions, tone, and other elements within the text to assess the specificity and relevance of the comments.

[0806] 4. Use of Emotion Engine:

[0807] The server then uses an emotion engine to recognize the user's emotions from the text data, extracting emotions from the content and tone of the speech and classifying them into emotion categories such as positive, negative, and neutral.

[0808] 5. Generate real-time feedback:

[0809] The server generates real-time feedback based on the text data and the results of the emotion engine, including closed and open feedback, in an appropriate format.

[0810] 6. Providing Feedback:

[0811] The terminal receives the generated feedback and provides it to the user: closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window that can be viewed by all participants.

[0812] 7. Detect and warn against aggressive speech:

[0813] The server performs real-time emotional analysis of comments and detects statements that are deemed to be overbearing or harassing.

[0814] If the server detects such a situation, it immediately generates a warning message and sends it to the terminal of the participant in question.

[0815] The terminal displays a warning message to the user in real time, prompting the user to take appropriate action.

[0816] 8. Generate and share meeting minutes:

[0817] The server records all comments and feedback during the meeting, and automatically generates minutes after the meeting and provides them to all participants.

[0818] The device displays the generated minutes to the user and provides a link that can be downloaded or shared.

[0819] Specific examples

[0820] Meeting Management

[0821] User (Person A): "I have some opinions about the direction this project should take."

[0822] The device captures Mr. A's speech as audio data and sends it to the server.

[0823] The server converts the voice data into text data and analyzes the content using NLP technology and an emotion engine.

[0824] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[0825] The server also detects negative emotions in Mr. A's comments and provides open feedback such as, "We recommend that you speak with more confidence."

[0826] After receiving the feedback, Person A makes a statement that provides specific data and continues speaking with even more confidence.

[0827] The terminal recaptures the new utterance and transmits it to the server.

[0828] The server again analyzes the new utterance and provides feedback if necessary.

[0829] Through this process, the system provides a fair and effective meeting environment in real time, and by utilizing an emotion engine, it also appropriately manages the emotions of participants. Furthermore, by automatically generating meeting minutes, the content of the meeting can be easily reviewed even after it has ended.

[0830] The processing flow will be explained below.

[0831] MODE FOR CARRYING OUT THE INVENTION

[0832] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. Furthermore, this system includes an "emotion engine" that recognizes the user's emotions and can provide feedback based on their emotions.

[0833] Specific processing steps

[0834] Step 1:

[0835] A user schedules a meeting.

[0836] Users log into the system and set the meeting date and time, participant list, and meeting goal.

[0837] Step 2:

[0838] The server creates a meeting room.

[0839] The server receives the request from the user, generates a unique meeting room ID, and creates a join link.

[0840] The server is pre-loaded and ready with the AI ​​models and emotion engines that will be used for the meeting.

[0841] Step 3:

[0842] The user invites participants.

[0843] The user sends the generated meeting link to prospective participants, who then click the link to enter the meeting room.

[0844] Step 4:

[0845] The device captures audio input.

[0846] The device captures microphone input each time a participant speaks and transmits it to the server as audio data in real time.

[0847] Step 5:

[0848] The server converts the voice data into text data.

[0849] The server converts the received voice data into text data using a speech recognition API, which makes it possible to analyze the spoken content as text information.

[0850] Step 6:

[0851] The server analyzes the text data.

[0852] The server then analyzes the converted text data using natural language processing (NLP) techniques, specifically extracting keywords, analyzing sentiment, and evaluating tone.

[0853] Step 7:

[0854] The server uses an emotion engine to recognize the user's emotion.

[0855] The server then uses an emotion engine to recognize the user's emotions from the text data, extracting emotions from the content and tone of the speech and classifying them into emotion categories such as positive, negative, and neutral.

[0856] Step 8:

[0857] The server generates real-time feedback.

[0858] The server generates real-time feedback based on the analysis of the text data and the emotion engine. The feedback can be either closed or open, and is provided in an appropriate format.

[0859] Step 9:

[0860] The device displays feedback.

[0861] The terminal displays the generated feedback in a chat window for a specific user (closed feedback), and also displays it in a feedback window for all participants (open feedback).

[0862] Step 10:

[0863] The server monitors and detects coercive remarks.

[0864] The server monitors all comments in real time and performs sentiment analysis.

[0865] Detects statements that are deemed to be overbearing or harassment.

[0866] Step 11:

[0867] The server will warn you against overbearing comments.

[0868] When an overbearing remark is detected, the server generates a warning message for the user and sends it to the terminal.

[0869] Step 12:

[0870] The terminal displays a warning message.

[0871] The device will display a warning message to the affected user in real time, urging them to reconsider their comments.

[0872] Step 13:

[0873] The server records your comments and feedback.

[0874] The server records all comments and feedback during the meeting and stores them in a database.

[0875] Step 14:

[0876] The server generates the minutes.

[0877] After the meeting ends, the server automatically generates minutes based on the recorded data.

[0878] Step 15:

[0879] The server shares the minutes.

[0880] The server provides the generated minutes to all participants via email or link.

[0881] Step 16:

[0882] The device will allow you to view and download the minutes.

[0883] The terminal provides a link to allow the user to view and download the minutes.

[0884] Example 2

[0885] 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."

[0886] In traditional online and offline meetings, some participants find it difficult to speak up, and the environment is prone to overbearing remarks and harassment. Furthermore, there is a lack of means to properly record what is being said and provide feedback in real time. As a result, it is difficult to achieve fair discussions.

[0887] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0888] In this invention, the server includes means for collecting participants' comments as voice data, means for converting the voice data into text data, means for analyzing the text data to generate real-time feedback, means for providing the feedback to participants, and means for recognizing user emotions through sentiment analysis, thereby providing a fair and effective meeting environment, appropriately managing user emotions through sentiment analysis, and improving the quality of discussions.

[0889] "Means for collecting participants' speech as audio data" refers to a device or system that records the speech of meeting participants as audio data in real time. A specific example is a device with a microphone.

[0890] "Means for converting voice data into text data" refers to software or algorithms for analyzing voice data and converting it into text data. A specific example is a voice recognition API.

[0891] "Means for analyzing text data and generating real-time feedback" refers to natural language processing technologies and algorithms for evaluating participants' comments based on text data and generating instant feedback.

[0892] "Means for providing feedback to participants" refers to a device or system that displays or notifies participants of generated feedback in real time. Examples include a chat window or a feedback window.

[0893] "Means for recognizing user emotions through sentiment analysis" refers to software or algorithms that extract participants' emotions from text data and classify and recognize those emotions. A specific example is an emotion engine.

[0894] This invention is a system that uses AI to provide real-time feedback in online and offline meetings. This system creates an environment where all participants can easily express their opinions and ensure fair discussions. Furthermore, this system includes a function to recognize user emotions through sentiment analysis.

[0895] The specific configuration of this system is as follows.

[0896] Hardware and software used

[0897] This system is realized by the following hardware and software.

[0898] Hardware: A device with a microphone (e.g., a computer, smartphone, etc.) used to collect the speech of meeting participants.

[0899] software:

[0900] A speech recognition API for converting voice data into text data (e.g., Google Cloud Speech-to-Text API).

[0901] Natural language processing techniques (e.g., TensorFlow, Natural Language Toolkit (NLTK)) for analyzing text data.

[0902] Software for performing sentiment analysis (e.g., IBM Watson Tone Analyzer).

[0903] Processing flow

[0904] 1. Audio data collection:

[0905] The terminal collects speech from meeting participants as audio data in real time. Specifically, a microphone-equipped device captures speech from participants and transmits it to a server.

[0906] 2. Audio to text conversion:

[0907] The server receives the voice data sent from the device and converts it into text data using the Google Cloud Speech-to-Text API.

[0908] 3. Text data analysis:

[0909] The server analyzes the converted text data using natural language processing technology (TensorFlow and Natural Language Toolkit (NLTK)). This analyzes keywords, emotions, tone, etc. in the text and evaluates the specificity and relevance of the statements.

[0910] 4. Use of Emotion Engine:

[0911] The server also uses an emotion engine (IBM Watson Tone Analyzer) to recognize the user's emotions and classify them into emotion categories such as positive, negative, and neutral.

[0912] 5. Generate real-time feedback:

[0913] The server generates real-time feedback based on the collected data and sentiment analysis results, including closed and open feedback.

[0914] 6. Providing Feedback:

[0915] The terminal receives the generated feedback and provides it to the user. Closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window that can be viewed by all participants.

[0916] 7. Detect and warn against aggressive speech:

[0917] The server analyzes the sentiment of comments in real time and detects any statements that are deemed to be overbearing or harassing. If any are detected, a warning message is immediately generated and sent to the relevant participant's device.

[0918] 8. Generate and share meeting minutes:

[0919] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. The device displays the minutes to the user and provides a link that can be downloaded or shared.

[0920] Specific examples

[0921] As a concrete example, we will show how a meeting proceeds.

[0922] User (Person A): "I have some opinions about the direction this project should take."

[0923] The device captures Mr. A's speech as audio data and sends it to the server.

[0924] The server converts the voice data into text data and analyzes the content using natural language processing technology and an emotion engine.

[0925] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[0926] The server also detects negative emotions in Mr. A's comments and provides open feedback such as, "We recommend that you speak with more confidence."

[0927] Person A receives the feedback, makes statements that provide specific data, and continues to speak with confidence.

[0928] The terminal recaptures the new utterance and transmits it to the server.

[0929] The server then analyzes new comments and provides feedback as necessary. Through this process, the system provides a fair and effective meeting environment in real time, and by utilizing an emotion engine, it also appropriately manages participants' emotions. Furthermore, by automatically generating meeting minutes, the content can be easily reviewed even after the meeting has ended.

[0930] Prompt Sentence Examples

[0931] Prompt: Type what you want to say next in the meeting. The system will provide real-time feedback.

[0932] User says: "I have an opinion on the direction of the project."

[0933] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0934] Step 1:

[0935] Audio data collection

[0936] Input: User's speech (audio)

[0937] Specific operation: Capture participants' speech through a microphone connected to the device.

[0938] Data processing: The captured audio data is converted into a digital signal.

[0939] Output: Digital audio data

[0940] Detailed explanation: The device collects speech from meeting participants as audio data in real time and sends it to a server. For example, a microphone on a laptop installed in a conference room captures each participant's speech and sends the data to the server via Wi-Fi.

[0941] Step 2:

[0942] Converting audio data to text

[0943] Input: Digital audio data

[0944] Specific operation: The server passes the received voice data to the voice recognition API.

[0945] Data processing: Use the Google Cloud Speech-to-Text API to convert the audio data into text data.

[0946] Output: Text data

[0947] Detailed explanation: The server passes the digital voice data sent from the device to the Google Cloud Speech-to-Text API, which converts the voice data into text data. For example, a speech saying "I have an opinion about the direction of the project" is converted into text data saying "I have an opinion about the direction of the project."

[0948] Step 3:

[0949] Text data analysis

[0950] Input: Text data

[0951] Specific operation: The server passes the converted text data to natural language processing technology.

[0952] Data processing: Analyze text for keywords, tone, sentiment, etc. using TensorFlow and the Natural Language Toolkit (NLTK).

[0953] Output: Analysis result data (keywords, tone, sentiment, etc.)

[0954] Detailed explanation: The server uses natural language processing (NLP) technology to analyze text data. For example, keywords are extracted from the text "I have an opinion on the direction of the project" and evaluated as "lacking specificity."

[0955] Step 4:

[0956] Using the Emotion Engine

[0957] Input: Analysis result data (keywords, tone, emotion, etc.)

[0958] Specific operation: The server uses IBM Watson Tone Analyzer to extract and classify emotions.

[0959] Data Processing: The sentiment engine extracts sentiment categories such as positive, negative, and neutral from the analyzed data.

[0960] Output: Sentiment analysis result data

[0961] Detailed explanation: The server further analyzes the parsed text data with a sentiment engine. For example, the phrase "I have an opinion" is classified as negative.

[0962] Step 5:

[0963] Generate real-time feedback

[0964] Input: Analysis result data and sentiment analysis result data

[0965] Specific operation: The server generates real-time feedback based on the collected data.

[0966] Data processing: Generating feedback in the form of closed and open feedback.

[0967] Output: Feedback data

[0968] Detailed explanation: The server generates specific feedback based on the analysis results and sentiment analysis results. For example, closed feedback such as "It would be effective if you provided specific data" is generated, and open feedback such as "We recommend that you speak more confidently."

[0969] Step 6:

[0970] Providing Feedback

[0971] Input: Feedback data

[0972] Specific operation: The terminal displays the generated feedback to the user.

[0973] Data processing: The feedback data is converted into data for display on the user interface.

[0974] Output: Display of feedback message

[0975] Detailed description: The terminal provides the user with real-time feedback sent from the server. For example, closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window visible to all participants.

[0976] Step 7:

[0977] Detecting and warning against coercive speech

[0978] Input: Text data and sentiment analysis result data

[0979] What it does: The server applies rules to detect intrusive and harassing comments.

[0980] Data processing: Detects overbearing remarks and harassment based on sentiment analysis data and generates warning messages.

[0981] Output: Warning message

[0982] Detailed explanation: The server detects overbearing remarks and harassment in real time based on the results of sentiment analysis. For example, if a remark such as "That's impossible!" is judged to be overbearing, a warning message "Please avoid overbearing remarks" is generated and sent to the relevant participant's device.

[0983] Step 8:

[0984] Generate and share meeting minutes

[0985] Input: Speech and feedback data for the entire meeting

[0986] Specific operation: The server records all data during the meeting and generates minutes.

[0987] Data processing: Automatically generate minutes from recorded data and create a link for sharing.

[0988] Output: Meeting minutes data and shared link

[0989] Detailed explanation: The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. For example, after the meeting ends, participants receive an email saying, "All the contents of the meeting have been summarized," and the minutes are displayed on their devices. A link that can be downloaded or shared is also provided.

[0990] This system provides a fair and effective meeting environment, manages participants' emotions appropriately, and improves the quality of discussions. Automatically generating meeting minutes also makes it easy to review important information.

[0991] (Application example 2)

[0992] 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."

[0993] In physical stores, staff meetings and training sessions often lack the fairness and specificity of statements and proper feedback. It's also difficult to detect coercive remarks or harassment in real time and respond immediately, and there's no system for automatically generating and providing meeting minutes to participants. As a result, fair and effective discussions and organizational management can be difficult.

[0994] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0995] In this invention, the server includes means for collecting participants' comments as voice data, means for converting the voice data into text data, means for analyzing the text data to generate real-time feedback, means for providing the feedback to participants, means for recognizing emotions in the text data, means for generating feedback based on the recognized emotions, means for detecting coercive remarks and harassment and issuing immediate warnings, and means for automatically generating meeting minutes and providing them to participants. This makes meetings and training sessions in brick-and-mortar stores fair and effective, and enables real-time feedback, detection and warning of coercive remarks, and automatic generation of meeting minutes.

[0996] "Means for collecting participants' speech as audio data" refers to devices or systems that record the speech of participants in conferences or meetings in real time and store it as digital audio data.

[0997] "Means for converting voice data into text data" refers to a device or system that converts collected voice data into text information using natural language processing technology or voice recognition technology.

[0998] The "means for analyzing text data and generating real-time feedback" refers to a device or system for analyzing the converted text data and generating instantaneous feedback based on the content and sentiment.

[0999] The "means for providing feedback to participants" refers to a device or system for displaying or notifying the participants of the conference or meeting of the generated feedback in an appropriate format.

[1000] A "means for recognizing emotions in text data" is a device or system that uses natural language processing or machine learning techniques to extract and classify emotions from the content and tone of text data.

[1001] The "means for generating feedback based on recognized emotions" refers to a device or system that generates feedback appropriate to the emotions in real time based on analyzed emotional data.

[1002] "Means for detecting coercive remarks and harassment and issuing immediate warnings" refers to devices or systems that automatically detect parts of a participant's remarks that are deemed to be coercive or harassing and immediately display a warning.

[1003] "Means for automatically generating meeting minutes and providing them to participants" refers to a device or system that records all comments and feedback made during a meeting, and automatically creates and distributes minutes to participants after the meeting ends.

[1004] This invention aims to realize a system that uses smart devices and a server to provide real-time feedback for staff meetings and training sessions in brick-and-mortar stores. Specific embodiments of this system are described below.

[1005] The system uses a microphone built into a smart device (e.g., a smartphone or tablet) to collect participants' speech as audio data. After the speech is collected, the audio data is transferred from the smart device to a server.

[1006] The server uses speech recognition software to convert the voice data into text data. This can be effectively done using existing speech recognition technology, such as Google's speech recognition API. The server then analyzes the text data using natural language processing (NLP) techniques. For this purpose, the Transformers library provided by Hugging Face can be used.

[1007] The parsed text data is then analyzed using an emotion engine (e.g., nlptown / bert-base-multilingual-uncased-sentiment model) to recognize emotions. The server uses this emotion data to generate real-time feedback and provides it to participants. The feedback can be displayed on the screen of a smart device or announced via audio.

[1008] Furthermore, the server uses text data and emotional data to detect overbearing remarks and harassment. If detected, it generates feedback to immediately warn the relevant participants and notifies them in real time. All remarks and feedback are also recorded, and minutes are automatically created and shared with all participants after the meeting ends.

[1009] For example, if someone says during a meeting, "I have an opinion on the direction of this project. Next time, let's present more concrete data. Also, speaking based on facts will be more persuasive," the system will capture this statement, convert it from audio data to text data, and generate real-time feedback along with emotion recognition.

[1010] As described above, this system can be used to ensure that meetings and training sessions in physical stores are conducted fairly and effectively, providing an environment in which participants' opinions are properly reflected.

[1011] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1012] Step 1:

[1013] When the device detects the start of a meeting, it collects each participant's speech as audio data. At this time, it uses the smart device's microphone to capture the audio in real time. The input is the participant's voice, and the output is digital audio data. This data is immediately sent to the server.

[1014] Step 2:

[1015] The server converts the received voice data into text data using Google's speech recognition API. The input is digital voice data, and the output is text data converted from this voice data into character string information. This conversion turns the voice information into analyzable text information.

[1016] Step 3:

[1017] The server analyzes the text data and uses natural language processing (NLP) technology to analyze the content. Specifically, it extracts keywords, tone, and utterance relationships within the text. The input is text data generated by a speech recognition API, and the output is the analysis results. Emotion recognition of the text data is also performed based on the analysis results.

[1018] Step 4:

[1019] The server uses an emotion engine (e.g., nlptown / bert-base-multilingual-uncased-sentiment model) to recognize the emotion of text data. The input is text data, and the output is recognized emotion data. This emotion data is classified into categories such as positive, negative, and neutral.

[1020] Step 5:

[1021] The server generates real-time feedback based on the emotion data and text analysis results. The feedback can be in the form of closed feedback (only for specific participants) or open feedback (for all participants). The input is emotion data and text analysis results, and the output is a feedback message.

[1022] Step 6:

[1023] The device receives the generated feedback and provides it to the participants. Closed feedback is displayed on a specific participant's device, and open feedback is displayed on all participants' devices. The input is the feedback message, and the output is a display of the feedback.

[1024] Step 7:

[1025] The server monitors participants' comments in real time and detects overbearing comments or harassment. If any are detected, a warning message is immediately generated and sent to the relevant participant. The input is text data and emotional data, and the output is a warning message.

[1026] Step 8:

[1027] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. The generated minutes are shared with all participants. The input is all comment data and feedback data, and the output is the generation and sharing of meeting minutes.

[1028] 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.

[1029] 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.

[1030] 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.

[1031] [Third embodiment]

[1032] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1033] 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.

[1034] 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).

[1035] 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.

[1036] 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.

[1037] 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).

[1038] 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.

[1039] 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.

[1040] 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.

[1041] 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.

[1042] 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.

[1043] 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."

[1044] Program Overview

[1045] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. This system includes "means for collecting participants' comments as audio data," "means for converting the audio data into text data," "means for analyzing the text data and generating real-time feedback," "means for providing feedback to participants," and "means for detecting overbearing comments and harassment and issuing an immediate warning."

[1046] Program processing

[1047] The program of the system of the present invention is processed as follows.

[1048] 1. Audio data collection:

[1049] The device collects speech from meeting participants as audio data in real time, specifically including the ability to capture microphone input from each participant.

[1050] 2. Audio to text conversion:

[1051] The server receives the voice data sent from the device and converts it into text data using a voice recognition API, etc. This makes it possible to analyze the content of the speech as text information.

[1052] 3. Text data analysis:

[1053] The server then analyzes the converted text data using natural language processing (NLP) technology, analyzing keywords, emotions, tone, and other elements within the text to assess the specificity and relevance of the comments.

[1054] 4. Generate real-time feedback:

[1055] The server generates real-time feedback based on the analysis results, including closed and open feedback, in an appropriate format.

[1056] 5. Providing Feedback:

[1057] The terminal receives the generated feedback and provides it to the user. Closed feedback is displayed in the chat window of a specific participant, while open feedback is displayed so that all participants can see it.

[1058] 6. Detect and warn against aggressive speech:

[1059] The server performs real-time emotional analysis of comments and detects statements that are deemed to be overbearing or harassing.

[1060] If the server detects such a situation, it immediately generates a warning message and sends it to the terminal of the participant in question.

[1061] The terminal displays a warning message to the user in real time, prompting the user to take appropriate action.

[1062] 7. Generate and share meeting minutes:

[1063] The server records all comments and feedback during the meeting, and automatically generates minutes after the meeting and provides them to all participants.

[1064] The device displays the generated minutes to the user and provides a link that can be downloaded or shared.

[1065] Specific examples

[1066] Meeting Management

[1067] User (Person A): "I have some opinions about the direction this project should take."

[1068] The device captures Mr. A's speech as audio data and sends it to the server.

[1069] The server converts the voice data into text data and analyzes the content using NLP technology.

[1070] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[1071] Person A receives the feedback and makes a statement providing specific data.

[1072] The device recaptures Mr. A's new comments and sends them to the server.

[1073] The server analyzes the new utterance and provides further feedback.

[1074] In this way, the system provides a fair and effective meeting environment in real time, ensuring that all participants feel comfortable speaking up. Furthermore, the system automatically generates meeting minutes, making it easy to review the content of meetings even after they have ended.

[1075] The processing flow will be explained below.

[1076] Program processing flow

[1077] Step 1: Set up a meeting

[1078] 1. A user schedules a meeting.

[1079] Users log into the system and set the meeting date and time, participant list, and meeting goal.

[1080] 2. The server creates a meeting room.

[1081] The server receives the request from the user, generates a unique meeting room ID, and creates a join link.

[1082] The server pre-loads and prepares the AI ​​models to be used in the meeting.

[1083] 3. The user invites participants.

[1084] The user sends the generated meeting link to the participants, who click the link to enter the meeting room.

[1085] Step 2: Collecting audio data

[1086] 4. The device captures the audio input.

[1087] The device captures microphone input each time a participant speaks and transmits it to the server as audio data in real time.

[1088] Step 3: Convert audio data to text

[1089] 5. The server converts the audio data into text data.

[1090] The server converts the received voice data into text data using a speech recognition API, which converts the voice into text information that can be analyzed.

[1091] Step 4: Analyze the text data

[1092] 6. The server analyzes the text data.

[1093] The server then analyzes the converted text data using natural language processing (NLP) techniques, specifically extracting keywords, analyzing sentiment, and evaluating tone.

[1094] Step 5: Generate real-time feedback

[1095] 7. The server generates real-time feedback.

[1096] Based on the analysis results, the server generates either crowded or open feedback. Closed feedback is given to specific participants, while open feedback is given to all participants.

[1097] 8. The device will display feedback.

[1098] The terminal displays closed feedback to a specific user in a chat window, and displays open feedback in a feedback window that can be viewed by all participants.

[1099] Step 6: Detect and flag coercive language

[1100] 9. The server monitors the comments.

[1101] The server monitors all comments in real time and performs sentiment analysis.

[1102] 10. The server detects overbearing comments.

[1103] The server detects overbearing remarks and harassment based on the results of sentiment analysis.

[1104] 11. The server issues a warning.

[1105] If the server detects an overbearing remark, it generates and sends a warning message to the user.

[1106] 12. The terminal displays a warning message.

[1107] The device will display a warning message to the affected user in real time, urging them to reconsider their comments.

[1108] Step 7: Generate and share meeting minutes

[1109] 13. The server records your comments and feedback.

[1110] The server records all comments and feedback during the meeting and stores them in a database.

[1111] 14. The server generates the minutes.

[1112] The server automatically generates minutes based on the data recorded after the meeting ends.

[1113] 15. The server shares the minutes.

[1114] The server provides the generated minutes to all participants via email or link.

[1115] 16. The device will be able to display and download the minutes.

[1116] The terminal provides a function that allows the user to view and download the minutes.

[1117] Example: Feedback process

[1118] User (Person A): "I have some opinions about the direction this project should take."

[1119] The device captures A's remarks and sends them to the server.

[1120] The server converts the voice data into text data and analyzes the content using NLP technology.

[1121] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or specific data."

[1122] Person A receives the feedback and makes a statement providing specific data.

[1123] The device recaptures the new utterance and sends it to the server.

[1124] The server analyzes the new utterance and provides further feedback.

[1125] Through this process, the system provides a fair and effective meeting environment in real time.

[1126] Example 1

[1127] 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."

[1128] In online and offline meetings, it is necessary to create an environment where all participants can express their opinions fairly and easily, and to prevent discussions from being biased towards certain individuals. There is also a need for systems that can detect overbearing remarks and harassment in real time and respond immediately. Furthermore, to efficiently conduct meetings, it is desirable to provide real-time feedback and automatically generate meeting minutes. However, conventional systems have not been able to adequately resolve these issues.

[1129] 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.

[1130] In this invention, the server includes means for collecting participants' comments as audio data, means for converting the audio data into text data, means for analyzing the text data using natural language processing technology, means for generating real-time feedback based on the analysis results, means for providing the feedback to participants, means for detecting overbearing comments or harassment and generating a warning message, and means for recording the comments and feedback of the meeting and generating minutes after the meeting. This provides an environment in which all participants can speak fairly and easily, making it possible to provide effective feedback in real time. Furthermore, overbearing comments and harassment can be addressed immediately, and the automatically generated minutes make it easy to review the content of the discussion.

[1131] "Participant" means an individual who participates in a meeting or conference.

[1132] "Audio data" refers to data that has been digitized from audio recorded as sound waves.

[1133] "Text data" is data that has been digitized as character information.

[1134] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[1135] "Real-time feedback" refers to instant assessment and advice provided during a meeting.

[1136] "Positive speech" is speech that sounds intimidating or aggressive to others.

[1137] "Harassment" refers to behavior or speech that harasses or offends others.

[1138] A "warning message" is a notification that warns you against certain actions or statements.

[1139] "Minutes" are a written record of all statements and discussions made during a meeting or conference.

[1140] "Means of collection" refers to devices or software that capture participants' voices as digital data.

[1141] "Means for converting" refers to a technology or system that converts voice data into text data.

[1142] "Means of analysis" refers to the technology or system used to analyze text data and determine meaning and sentiment.

[1143] "Means for generating" refers to a technology or system that generates feedback or warning messages based on the analysis results.

[1144] "Means for providing" refers to the equipment or software used to deliver generated feedback and warning messages to participants.

[1145] "Means of recording and generating" refers to technology or systems that store comments and feedback during meetings and later compile them into minutes.

[1146] A "system" is a set of devices or software that integrates these means to achieve a set of functions.

[1147] MODE FOR CARRYING OUT THE INVENTION

[1148] This invention is a system that uses AI to provide real-time feedback during online and offline meetings, creating an environment where all participants can easily express their opinions. This system includes "means for collecting participants' comments as audio data," "means for converting the audio data into text data," "means for analyzing the text data using natural language processing technology," "means for generating real-time feedback based on the analysis results," "means for providing feedback to participants," "means for detecting overbearing comments and harassment and generating a warning message," and "means for recording comments and feedback during meetings and generating meeting minutes after the meeting."

[1149] Specifically, the process is as follows:

[1150] Audio data collection

[1151] The device collects speech from each participant during the meeting as real-time audio data using a built-in microphone or an external microphone, and transmits the collected audio data to a server in a compressed format (e.g., MP3 or WAV format).

[1152] Converting audio data to text

[1153] The server receives the voice data sent from the device and converts it into text using a speech recognition API such as Google Cloud Speech-to-Text API or IBM Watson Speech to Text, then temporarily stores the converted text in a database.

[1154] Text data analysis

[1155] The server analyzes the stored text data using natural language processing (NLP) technology. Specifically, it uses Google Cloud Natural Language API and spaCy to evaluate keywords, sentence tone, and sentiment within the text. It evaluates the specificity and relevance of the text data and generates closed-form feedback.

[1156] Generate real-time feedback

[1157] The server generates real-time feedback based on the analysis results, classifies the feedback into closed feedback (for specific participants) and open feedback (for all participants), and sends it to the terminal in the appropriate format.

[1158] Providing Feedback

[1159] The device receives the feedback sent from the server, and displays the closed feedback in the chat window of the specific participant, while the open feedback is displayed on a screen that can be viewed by all participants. Platforms such as Microsoft Teams and Zoom may be used.

[1160] Detecting and warning against coercive speech

[1161] The server analyzes the sentiment of comments in real time and detects any overbearing or harassing content. If any are detected, a warning message is generated and sent to the relevant participant's device.

[1162] The device displays a warning message to the user in real time and prompts the user to take appropriate action, typically using the notification function of the web browser.

[1163] Generate and share meeting minutes

[1164] The server records all comments and feedback during the meeting, automatically generates minutes after the meeting, and generates a link to share the minutes with all participants.

[1165] The device displays the generated minutes to the user and provides download and sharing links, and can output them in Microsoft Word and Google Docs formats.

[1166] Specific examples

[1167] Meeting Management

[1168] User (Person A): "I have some opinions about the direction this project should take."

[1169] The device captures A's speech as audio data and sends it to the server. The captured audio data is saved in WAV format.

[1170] The server uses the Google Cloud Speech-to-Text API to convert Mr. A's voice data into text data in real time. The converted text data is recorded as "I have an opinion about the direction of this project."

[1171] The server analyzes the text and determines that it is "not specific enough" - NLP analysis detects that it does not contain specific keywords.

[1172] The server generates closed feedback such as "It would be more effective if you provided examples or concrete data."

[1173] The device will display this feedback in Mr. A's individual chat window.

[1174] Examples of problematic statements

[1175] User B makes a commanding statement: "Your idea is pointless."

[1176] The device captures Mr. B's speech as audio data and sends it to the server.

[1177] The server converts the voice data into text data and performs sentiment analysis using NLP technology. The analysis results indicate that the remark is "overbearing."

[1178] The server generates a warning message saying, "An overbearing remark has been detected. Please choose appropriate words." and sends it to Mr. B's device.

[1179] The device will notify Mr. B of this warning message in real time, urging him to be careful.

[1180] Example prompts for generative AI models

[1181] Please analyze the following text data and determine whether it constitutes coercive remarks or harassment.

[1182] Data Text: "That suggestion makes absolutely no sense. Anyone could have thought of something like that."

[1183] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1184] Program processing flow

[1185] Step 1: Collecting audio data

[1186] The device collects speech from each participant during a meeting as audio data in real time. The input is the participant's voice, and the output is audio data. Specifically, the device captures audio using a built-in microphone or an external microphone. The collected audio data is temporarily saved in WAV format and sent to a server.

[1187] Step 2: Convert audio data to text

[1188] The server receives the voice data sent from the device. The input is WAV format voice data, and the output is text data. Specifically, the voice data is converted into text data using the Google Cloud Speech-to-Text API or IBM Watson Speech to Text. At this time, the voice data is converted into text data and temporarily stored in a database.

[1189] Step 3: Analyze the text data

[1190] The server analyzes the stored text data using natural language processing (NLP) technology. The input is text data, and the output is the analysis results (keywords, tone, and emotional information). Specifically, it uses Google Cloud Natural Language API and spaCy to extract keywords from the text data, evaluate tone, and analyze emotions. Based on the results of this analysis, the specificity and relevance of the text are evaluated, and feedback candidates are generated.

[1191] Step 4: Generate real-time feedback

[1192] The server generates real-time feedback based on the analysis results. The input is the analysis results, and the output is the feedback message. The server classifies the feedback into closed feedback (for specific participants) and open feedback (for all participants) formats and sends it to the terminal in the appropriate format. The generative AI model is used to generate the feedback text.

[1193] Step 5: Provide feedback

[1194] The device receives feedback sent from the server. The input is the feedback message, and the output is the feedback provided to the user. Specifically, closed feedback is displayed in the chat window of a specific participant, and open feedback is displayed on a shared screen that can be viewed by all participants. The chat functions of Microsoft Teams or Zoom are used.

[1195] Step 6: Detect and flag coercive language

[1196] The server performs real-time sentiment analysis of comments and detects content that is deemed to be overbearing or harassing. The input is text data, and the output is a warning message. If overbearing or harassing comments are detected, a warning message is generated and sent to the relevant participant's device. An appropriate warning message is created using a generative AI model.

[1197] The device displays a warning message to the user in real time, specifically by using the notification function of the web browser, and prompts the user to take appropriate action.

[1198] Step 7: Generate and share meeting minutes

[1199] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting. The input is all comments and feedback during the meeting, and the output is the minutes data.

[1200] The device displays the generated minutes to the user, provides a download link and a sharing link for sharing with all participants, and is configured to output the minutes in Microsoft Word or Google Docs format.

[1201] (Application example 1)

[1202] 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."

[1203] The present invention aims to improve the quality and security of comments made in meetings, provide a fair and safe environment, and provide a system that allows for post-conference review and auditing. In particular, it is necessary to strengthen meeting security by detecting coercive comments and harassment in real time and responding immediately. There is also a need for real-time feedback and automatic generation of meeting minutes so that comments made during meetings can be used as reference material later.

[1204] 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.

[1205] In this invention, the server includes a means for collecting participants' comments as audio data, a means for converting the audio data into text data, a means for analyzing the text data to generate real-time feedback, and a means for detecting comments containing specific keywords and generating a warning. This makes it possible to improve the quality of comments in real time during a meeting, and to detect and warn against overbearing comments and harassment. It is also possible to automatically record the contents of the meeting and generate materials that can be referenced later.

[1206] "Means for collecting participants' comments as audio data" refers to a device or system that captures and records the content of participants' comments in a conference room or on an online platform as audio data in real time.

[1207] The "means for converting voice data into text data" refers to a technology that converts collected voice data into text information using voice recognition technology and outputs it as text data that can be analyzed.

[1208] The "means for analyzing text data and generating real-time feedback" is a system that analyzes the converted text data using natural language processing technology and provides participants with immediate feedback.

[1209] A "means for providing feedback to participants" is a device or system that visually or audibly communicates the generated feedback to participants.

[1210] The "means for detecting statements containing specific keywords and generating warnings" refers to a system that analyzes statements made during a meeting, detects specific keywords (e.g., overbearing statements or harassment) in real time, and generates warning messages based on the detection results.

[1211] MODE FOR CARRYING OUT THE INVENTION

[1212] Program Overview

[1213] This invention provides a system that uses real-time feedback from AI to improve the quality and security of comments made during meetings. The system collects participants' comments as audio data, converts them into text data, and analyzes them. Based on the analysis results, it generates real-time feedback and provides it to participants. The system also has a function to detect comments containing specific keywords and generate a warning. Specific embodiments of the system are described in detail below.

[1214] System configuration

[1215] Hardware and Software

[1216] Microphone: A device that collects speech from conference participants and is connected to a terminal.

[1217] Device (PC, smartphone, smart glasses, head-mounted display, etc.): Receives audio data from the microphone and sends it to the server.

[1218] Server: Converts voice data into text data, analyzes the text data using natural language processing technology, and generates feedback. It also detects utterances containing specific keywords and generates warnings.

[1219] Speech recognition API (such as Google Speech Recognition API): An API for converting voice data into text data.

[1220] Natural language processing libraries (NLTK, VADER, etc.): Libraries for analyzing text data and assessing sentiment and content.

[1221] System Operation

[1222] 1. Collecting and transmitting participant comments:

[1223] The device collects participants' speech in real time through a connected microphone, and the collected voice data is sent to a server.

[1224] 2. Audio to text conversion:

[1225] The server converts the transmitted voice data into text information using a speech recognition API (Google Speech Recognition API).

[1226] 3. Text data analysis:

[1227] The server then analyzes the converted text data using natural language processing (NLTK, VADER) to evaluate the sentiment, tone, and specificity of the comments.

[1228] 4. Generating and Presenting Real-Time Feedback:

[1229] The server generates feedback based on the analysis results and sends it to the device. The device then presents the generated feedback to the participant in real time. For example, feedback such as "It would be more effective if you provided specific data" is displayed on the participant's screen.

[1230] 5. Detecting specific keywords and generating alerts:

[1231] The server detects comments containing specific keywords (e.g., overbearing, harassment), and immediately generates and sends a warning message to the device after the relevant comment is made. The warning is displayed to participants in real time.

[1232] Specific examples

[1233] If participant A says during a meeting, "I have some opinions about the direction this project should take, but I think they're too harsh," the system will collect this comment, convert the audio data into text data, and analyze it. Based on the analysis results, a warning such as "Negative comments have been detected" will be generated and displayed in real time on participant A's device.

[1234] Example prompts for generative AI models

[1235] Analyze the following statements in meetings to detect negative or positive sentiment and generate feedback: "I have some opinions about the direction we should take this project, but I think they're too harsh."

[1236] This system will improve the quality of speech in meetings, detect overbearing remarks and harassment, and provide a safe meeting environment.

[1237] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1238] Step 1:

[1239] Audio data collection

[1240] Subject: Terminal

[1241] Specific operation: The device collects the speech of meeting participants in real time via a microphone.

[1242] Input: Remarks by conference participants (audio data)

[1243] Output: Collected audio data

[1244] Data processing: The device captures the speech as digital audio data and prepares it for transmission to the server.

[1245] Step 2:

[1246] Converting audio data to text

[1247] Subject: Server

[1248] Specific operation: The server receives the voice data sent from the device and converts it into text data using the speech recognition API (Google Speech Recognition API).

[1249] Input: Collected audio data

[1250] Output: Converted text data

[1251] Data processing: Using voice recognition technology, the voice data is analyzed and text data is generated. The server stores the results of this conversion process in a related database.

[1252] Step 3:

[1253] Text data analysis

[1254] Subject: Server

[1255] Specific operation: The server analyzes the converted text data using natural language processing (NLP) techniques (NLTK, VADER).

[1256] Input: Text data

[1257] Output: Analysis results (emotion, tone, specificity, etc.)

[1258] Data processing: The text data is parsed, keywords are extracted, sentiment analysis is performed, and the specificity of the comments is evaluated. The server uses the analysis results to proceed to the next processing step.

[1259] Step 4:

[1260] Generate real-time feedback

[1261] Subject: Server

[1262] Specific behavior: The server generates real-time feedback based on the results of analyzing the text data and sends it to the device. For example, if a comment lacks specificity, the server generates feedback such as, "It would be more effective if you provided specific data."

[1263] Input: Analysis results

[1264] Output: Real-time feedback

[1265] Data processing: Based on the analysis results, feedback messages are generated and sent to the device instructing it to display. The server places particular emphasis on oppressive or negative comments.

[1266] Step 5:

[1267] Providing feedback

[1268] Subject: Terminal

[1269] Specific operation: The device receives the feedback sent from the server and presents it to the participants. Closed feedback is presented to specific participants, and open feedback is presented to all participants.

[1270] Input: Real-time feedback

[1271] Output: Displayed feedback

[1272] Data Calculation: The terminal visually displays feedback messages and prompts specific participants to take appropriate action.

[1273] Step 6:

[1274] Detecting specific keywords and generating alerts

[1275] Subject: Server

[1276] Specific operation: The server detects specific keywords (e.g., overbearing, harassment) in the analyzed text data, and if detected, immediately generates a warning message and sends it to the terminal.

[1277] Input: Analysis results of text data

[1278] Output: Warning message

[1279] Data calculation: The server checks the analysis results against a specific keyword list, and if there is a match, generates and sends a warning message.

[1280] Step 7:

[1281] Present a warning message

[1282] Subject: Terminal

[1283] Specific operation: The terminal receives the warning message sent from the server and displays it to the relevant participant.

[1284] Input: warning message

[1285] Output: The displayed warning message

[1286] Data calculation: The terminal visually displays a warning message and prompts participants who make overbearing remarks to take appropriate action.

[1287] This processing flow enables the system to improve speech quality in real time and enhance meeting security.

[1288] 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.

[1289] Program Overview

[1290] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. Furthermore, this system includes "means for collecting participants' comments as audio data," "means for converting audio data into text data," "means for generating real-time feedback by analyzing the text data," "means for providing feedback to participants," "means for detecting overbearing comments and harassment and issuing an immediate warning," as well as "an emotion engine that recognizes the user's emotions."

[1291] Program processing

[1292] The program of the system of the present invention is processed as follows.

[1293] 1. Audio data collection:

[1294] The device collects speech from meeting participants as audio data in real time, specifically including the ability to capture microphone input from each participant.

[1295] 2. Audio to text conversion:

[1296] The server receives the voice data sent from the device and converts it into text data using a voice recognition API, etc. This converts the voice into text information that can be analyzed.

[1297] 3. Text data analysis:

[1298] The server then analyzes the converted text data using natural language processing (NLP) technology, analyzing keywords, emotions, tone, and other elements within the text to assess the specificity and relevance of the comments.

[1299] 4. Use of Emotion Engine:

[1300] The server then uses an emotion engine to recognize the user's emotions from the text data, extracting emotions from the content and tone of the speech and classifying them into emotion categories such as positive, negative, and neutral.

[1301] 5. Generate real-time feedback:

[1302] The server generates real-time feedback based on the text data and the results of the emotion engine, including closed and open feedback, in an appropriate format.

[1303] 6. Providing Feedback:

[1304] The terminal receives the generated feedback and provides it to the user: closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window that can be viewed by all participants.

[1305] 7. Detect and warn against aggressive speech:

[1306] The server performs real-time emotional analysis of comments and detects statements that are deemed to be overbearing or harassing.

[1307] If the server detects such a situation, it immediately generates a warning message and sends it to the terminal of the participant in question.

[1308] The terminal displays a warning message to the user in real time, prompting the user to take appropriate action.

[1309] 8. Generate and share meeting minutes:

[1310] The server records all comments and feedback during the meeting, and automatically generates minutes after the meeting and provides them to all participants.

[1311] The device displays the generated minutes to the user and provides a link that can be downloaded or shared.

[1312] Specific examples

[1313] Meeting Management

[1314] User (Person A): "I have some opinions about the direction this project should take."

[1315] The device captures Mr. A's speech as audio data and sends it to the server.

[1316] The server converts the voice data into text data and analyzes the content using NLP technology and an emotion engine.

[1317] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[1318] The server also detects negative emotions in Mr. A's comments and provides open feedback such as, "We recommend that you speak with more confidence."

[1319] After receiving the feedback, Person A makes a statement that provides specific data and continues speaking with even more confidence.

[1320] The terminal recaptures the new utterance and transmits it to the server.

[1321] The server again analyzes the new utterance and provides feedback if necessary.

[1322] Through this process, the system provides a fair and effective meeting environment in real time, and by utilizing an emotion engine, it also appropriately manages the emotions of participants. Furthermore, by automatically generating meeting minutes, the content of the meeting can be easily reviewed even after it has ended.

[1323] The processing flow will be explained below.

[1324] MODE FOR CARRYING OUT THE INVENTION

[1325] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. Furthermore, this system includes an "emotion engine" that recognizes the user's emotions and can provide feedback based on their emotions.

[1326] Specific processing steps

[1327] Step 1:

[1328] A user schedules a meeting.

[1329] Users log into the system and set the meeting date and time, participant list, and meeting goal.

[1330] Step 2:

[1331] The server creates a meeting room.

[1332] The server receives the request from the user, generates a unique meeting room ID, and creates a join link.

[1333] The server is pre-loaded and ready with the AI ​​models and emotion engines that will be used for the meeting.

[1334] Step 3:

[1335] The user invites participants.

[1336] The user sends the generated meeting link to prospective participants, who then click the link to enter the meeting room.

[1337] Step 4:

[1338] The device captures audio input.

[1339] The device captures microphone input each time a participant speaks and transmits it to the server as audio data in real time.

[1340] Step 5:

[1341] The server converts the voice data into text data.

[1342] The server converts the received voice data into text data using a speech recognition API, which makes it possible to analyze the spoken content as text information.

[1343] Step 6:

[1344] The server analyzes the text data.

[1345] The server then analyzes the converted text data using natural language processing (NLP) techniques, specifically extracting keywords, analyzing sentiment, and evaluating tone.

[1346] Step 7:

[1347] The server uses an emotion engine to recognize the user's emotion.

[1348] The server then uses an emotion engine to recognize the user's emotions from the text data, extracting emotions from the content and tone of the speech and classifying them into emotion categories such as positive, negative, and neutral.

[1349] Step 8:

[1350] The server generates real-time feedback.

[1351] The server generates real-time feedback based on the analysis of the text data and the emotion engine. The feedback can be either closed or open, and is provided in an appropriate format.

[1352] Step 9:

[1353] The device displays feedback.

[1354] The terminal displays the generated feedback in a chat window for a specific user (closed feedback), and also displays it in a feedback window for all participants (open feedback).

[1355] Step 10:

[1356] The server monitors and detects coercive remarks.

[1357] The server monitors all comments in real time and performs sentiment analysis.

[1358] Detects statements that are deemed to be overbearing or harassment.

[1359] Step 11:

[1360] The server will warn you against overbearing comments.

[1361] When an overbearing remark is detected, the server generates a warning message for the user and sends it to the terminal.

[1362] Step 12:

[1363] The terminal displays a warning message.

[1364] The device will display a warning message to the affected user in real time, urging them to reconsider their comments.

[1365] Step 13:

[1366] The server records your comments and feedback.

[1367] The server records all comments and feedback during the meeting and stores them in a database.

[1368] Step 14:

[1369] The server generates the minutes.

[1370] After the meeting ends, the server automatically generates minutes based on the recorded data.

[1371] Step 15:

[1372] The server shares the minutes.

[1373] The server provides the generated minutes to all participants via email or link.

[1374] Step 16:

[1375] The device will allow you to view and download the minutes.

[1376] The terminal provides a link to allow the user to view and download the minutes.

[1377] Example 2

[1378] 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."

[1379] In traditional online and offline meetings, some participants find it difficult to speak up, and the environment is prone to overbearing remarks and harassment. Furthermore, there is a lack of means to properly record what is being said and provide feedback in real time. As a result, it is difficult to achieve fair discussions.

[1380] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1381] In this invention, the server includes means for collecting participants' comments as voice data, means for converting the voice data into text data, means for analyzing the text data to generate real-time feedback, means for providing the feedback to participants, and means for recognizing user emotions through sentiment analysis, thereby providing a fair and effective meeting environment, appropriately managing user emotions through sentiment analysis, and improving the quality of discussions.

[1382] "Means for collecting participants' speech as audio data" refers to a device or system that records the speech of meeting participants as audio data in real time. A specific example is a device with a microphone.

[1383] "Means for converting voice data into text data" refers to software or algorithms for analyzing voice data and converting it into text data. A specific example is a voice recognition API.

[1384] "Means for analyzing text data and generating real-time feedback" refers to natural language processing technologies and algorithms for evaluating participants' comments based on text data and generating instant feedback.

[1385] "Means for providing feedback to participants" refers to a device or system that displays or notifies participants of generated feedback in real time. Examples include a chat window or a feedback window.

[1386] "Means for recognizing user emotions through sentiment analysis" refers to software or algorithms that extract participants' emotions from text data and classify and recognize those emotions. A specific example is an emotion engine.

[1387] This invention is a system that uses AI to provide real-time feedback in online and offline meetings. This system creates an environment where all participants can easily express their opinions and ensure fair discussions. Furthermore, this system includes a function to recognize user emotions through sentiment analysis.

[1388] The specific configuration of this system is as follows.

[1389] Hardware and software used

[1390] This system is realized by the following hardware and software.

[1391] Hardware: A device with a microphone (e.g., a computer, smartphone, etc.) used to collect the speech of meeting participants.

[1392] software:

[1393] A speech recognition API for converting voice data into text data (e.g., Google Cloud Speech-to-Text API).

[1394] Natural language processing techniques (e.g., TensorFlow, Natural Language Toolkit (NLTK)) for analyzing text data.

[1395] Software for performing sentiment analysis (e.g., IBM Watson Tone Analyzer).

[1396] Processing flow

[1397] 1. Audio data collection:

[1398] The terminal collects speech from meeting participants as audio data in real time. Specifically, a microphone-equipped device captures speech from participants and transmits it to a server.

[1399] 2. Audio to text conversion:

[1400] The server receives the voice data sent from the device and converts it into text data using the Google Cloud Speech-to-Text API.

[1401] 3. Text data analysis:

[1402] The server analyzes the converted text data using natural language processing technology (TensorFlow and Natural Language Toolkit (NLTK)). This analyzes keywords, emotions, tone, etc. in the text and evaluates the specificity and relevance of the statements.

[1403] 4. Use of Emotion Engine:

[1404] The server also uses an emotion engine (IBM Watson Tone Analyzer) to recognize the user's emotions and classify them into emotion categories such as positive, negative, and neutral.

[1405] 5. Generate real-time feedback:

[1406] The server generates real-time feedback based on the collected data and sentiment analysis results, including closed and open feedback.

[1407] 6. Providing Feedback:

[1408] The terminal receives the generated feedback and provides it to the user. Closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window that can be viewed by all participants.

[1409] 7. Detect and warn against aggressive speech:

[1410] The server analyzes the sentiment of comments in real time and detects any statements that are deemed to be overbearing or harassing. If any are detected, a warning message is immediately generated and sent to the relevant participant's device.

[1411] 8. Generate and share meeting minutes:

[1412] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. The device displays the minutes to the user and provides a link that can be downloaded or shared.

[1413] Specific examples

[1414] As a concrete example, we will show how a meeting proceeds.

[1415] User (Person A): "I have some opinions about the direction this project should take."

[1416] The device captures Mr. A's speech as audio data and sends it to the server.

[1417] The server converts the voice data into text data and analyzes the content using natural language processing technology and an emotion engine.

[1418] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[1419] The server also detects negative emotions in Mr. A's comments and provides open feedback such as, "We recommend that you speak with more confidence."

[1420] Person A receives the feedback, makes statements that provide specific data, and continues to speak with confidence.

[1421] The terminal recaptures the new utterance and transmits it to the server.

[1422] The server then analyzes new comments and provides feedback as necessary. Through this process, the system provides a fair and effective meeting environment in real time, and by utilizing an emotion engine, it also appropriately manages participants' emotions. Furthermore, by automatically generating meeting minutes, the content can be easily reviewed even after the meeting has ended.

[1423] Prompt Sentence Examples

[1424] Prompt: Type what you want to say next in the meeting. The system will provide real-time feedback.

[1425] User says: "I have an opinion on the direction of the project."

[1426] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1427] Step 1:

[1428] Audio data collection

[1429] Input: User's speech (audio)

[1430] Specific operation: Capture participants' speech through a microphone connected to the device.

[1431] Data processing: The captured audio data is converted into a digital signal.

[1432] Output: Digital audio data

[1433] Detailed explanation: The device collects speech from meeting participants as audio data in real time and sends it to a server. For example, a microphone on a laptop installed in a conference room captures each participant's speech and sends the data to the server via Wi-Fi.

[1434] Step 2:

[1435] Converting audio data to text

[1436] Input: Digital audio data

[1437] Specific operation: The server passes the received voice data to the voice recognition API.

[1438] Data processing: Use the Google Cloud Speech-to-Text API to convert the audio data into text data.

[1439] Output: Text data

[1440] Detailed explanation: The server passes the digital voice data sent from the device to the Google Cloud Speech-to-Text API, which converts the voice data into text data. For example, a speech saying "I have an opinion about the direction of the project" is converted into text data saying "I have an opinion about the direction of the project."

[1441] Step 3:

[1442] Text data analysis

[1443] Input: Text data

[1444] Specific operation: The server passes the converted text data to natural language processing technology.

[1445] Data processing: Analyze text for keywords, tone, sentiment, etc. using TensorFlow and the Natural Language Toolkit (NLTK).

[1446] Output: Analysis result data (keywords, tone, sentiment, etc.)

[1447] Detailed explanation: The server uses natural language processing (NLP) technology to analyze text data. For example, keywords are extracted from the text "I have an opinion on the direction of the project" and evaluated as "lacking specificity."

[1448] Step 4:

[1449] Using the Emotion Engine

[1450] Input: Analysis result data (keywords, tone, emotion, etc.)

[1451] Specific operation: The server uses IBM Watson Tone Analyzer to extract and classify emotions.

[1452] Data Processing: The sentiment engine extracts sentiment categories such as positive, negative, and neutral from the analyzed data.

[1453] Output: Sentiment analysis result data

[1454] Detailed explanation: The server further analyzes the parsed text data with a sentiment engine. For example, the phrase "I have an opinion" is classified as negative.

[1455] Step 5:

[1456] Generate real-time feedback

[1457] Input: Analysis result data and sentiment analysis result data

[1458] Specific operation: The server generates real-time feedback based on the collected data.

[1459] Data processing: Generating feedback in the form of closed and open feedback.

[1460] Output: Feedback data

[1461] Detailed explanation: The server generates specific feedback based on the analysis results and sentiment analysis results. For example, closed feedback such as "It would be effective if you provided specific data" is generated, and open feedback such as "We recommend that you speak more confidently."

[1462] Step 6:

[1463] Providing Feedback

[1464] Input: Feedback data

[1465] Specific operation: The terminal displays the generated feedback to the user.

[1466] Data processing: The feedback data is converted into data for display on the user interface.

[1467] Output: Display of feedback message

[1468] Detailed description: The terminal provides the user with real-time feedback sent from the server. For example, closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window visible to all participants.

[1469] Step 7:

[1470] Detecting and warning against coercive speech

[1471] Input: Text data and sentiment analysis result data

[1472] What it does: The server applies rules to detect intrusive and harassing comments.

[1473] Data processing: Detects overbearing remarks and harassment based on sentiment analysis data and generates warning messages.

[1474] Output: Warning message

[1475] Detailed explanation: The server detects overbearing remarks and harassment in real time based on the results of sentiment analysis. For example, if a remark such as "That's impossible!" is judged to be overbearing, a warning message "Please avoid overbearing remarks" is generated and sent to the relevant participant's device.

[1476] Step 8:

[1477] Generate and share meeting minutes

[1478] Input: Speech and feedback data for the entire meeting

[1479] Specific operation: The server records all data during the meeting and generates minutes.

[1480] Data processing: Automatically generate minutes from recorded data and create a link for sharing.

[1481] Output: Meeting minutes data and shared link

[1482] Detailed explanation: The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. For example, after the meeting ends, participants receive an email saying, "All the contents of the meeting have been summarized," and the minutes are displayed on their devices. A link that can be downloaded or shared is also provided.

[1483] This system provides a fair and effective meeting environment, manages participants' emotions appropriately, and improves the quality of discussions. Automatically generating meeting minutes also makes it easy to review important information.

[1484] (Application example 2)

[1485] 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."

[1486] In physical stores, staff meetings and training sessions often lack the fairness and specificity of statements and proper feedback. It's also difficult to detect coercive remarks or harassment in real time and respond immediately, and there's no system for automatically generating and providing meeting minutes to participants. As a result, fair and effective discussions and organizational management can be difficult.

[1487] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1488] In this invention, the server includes means for collecting participants' comments as voice data, means for converting the voice data into text data, means for analyzing the text data to generate real-time feedback, means for providing the feedback to participants, means for recognizing emotions in the text data, means for generating feedback based on the recognized emotions, means for detecting coercive remarks and harassment and issuing immediate warnings, and means for automatically generating meeting minutes and providing them to participants. This makes meetings and training sessions in brick-and-mortar stores fair and effective, and enables real-time feedback, detection and warning of coercive remarks, and automatic generation of meeting minutes.

[1489] "Means for collecting participants' speech as audio data" refers to devices or systems that record the speech of participants in conferences or meetings in real time and store it as digital audio data.

[1490] "Means for converting voice data into text data" refers to a device or system that converts collected voice data into text information using natural language processing technology or voice recognition technology.

[1491] The "means for analyzing text data and generating real-time feedback" refers to a device or system for analyzing the converted text data and generating instantaneous feedback based on the content and sentiment.

[1492] The "means for providing feedback to participants" refers to a device or system for displaying or notifying the participants of the conference or meeting of the generated feedback in an appropriate format.

[1493] A "means for recognizing emotions in text data" is a device or system that uses natural language processing or machine learning techniques to extract and classify emotions from the content and tone of text data.

[1494] The "means for generating feedback based on recognized emotions" refers to a device or system that generates feedback appropriate to the emotions in real time based on analyzed emotional data.

[1495] "Means for detecting coercive remarks and harassment and issuing immediate warnings" refers to devices or systems that automatically detect parts of a participant's remarks that are deemed to be coercive or harassing and immediately display a warning.

[1496] "Means for automatically generating meeting minutes and providing them to participants" refers to a device or system that records all comments and feedback made during a meeting, and automatically creates and distributes minutes to participants after the meeting ends.

[1497] This invention aims to realize a system that uses smart devices and a server to provide real-time feedback for staff meetings and training sessions in brick-and-mortar stores. Specific embodiments of this system are described below.

[1498] The system uses a microphone built into a smart device (e.g., a smartphone or tablet) to collect participants' speech as audio data. After the speech is collected, the audio data is transferred from the smart device to a server.

[1499] The server uses speech recognition software to convert the voice data into text data. This can be effectively done using existing speech recognition technology, such as Google's speech recognition API. The server then analyzes the text data using natural language processing (NLP) techniques. For this purpose, the Transformers library provided by Hugging Face can be used.

[1500] The parsed text data is then analyzed using an emotion engine (e.g., nlptown / bert-base-multilingual-uncased-sentiment model) to recognize emotions. The server uses this emotion data to generate real-time feedback and provides it to participants. The feedback can be displayed on the screen of a smart device or announced via audio.

[1501] Furthermore, the server uses text data and emotional data to detect overbearing remarks and harassment. If detected, it generates feedback to immediately warn the relevant participants and notifies them in real time. All remarks and feedback are also recorded, and minutes are automatically created and shared with all participants after the meeting ends.

[1502] For example, if someone says during a meeting, "I have an opinion on the direction of this project. Next time, let's present more concrete data. Also, speaking based on facts will be more persuasive," the system will capture this statement, convert it from audio data to text data, and generate real-time feedback along with emotion recognition.

[1503] As described above, this system can be used to ensure that meetings and training sessions in physical stores are conducted fairly and effectively, providing an environment in which participants' opinions are properly reflected.

[1504] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1505] Step 1:

[1506] When the device detects the start of a meeting, it collects each participant's speech as audio data. At this time, it uses the smart device's microphone to capture the audio in real time. The input is the participant's voice, and the output is digital audio data. This data is immediately sent to the server.

[1507] Step 2:

[1508] The server converts the received voice data into text data using Google's speech recognition API. The input is digital voice data, and the output is text data converted from this voice data into character string information. This conversion turns the voice information into analyzable text information.

[1509] Step 3:

[1510] The server analyzes the text data and uses natural language processing (NLP) technology to analyze the content. Specifically, it extracts keywords, tone, and utterance relationships within the text. The input is text data generated by a speech recognition API, and the output is the analysis results. Emotion recognition of the text data is also performed based on the analysis results.

[1511] Step 4:

[1512] The server uses an emotion engine (e.g., nlptown / bert-base-multilingual-uncased-sentiment model) to recognize the emotion of text data. The input is text data, and the output is recognized emotion data. This emotion data is classified into categories such as positive, negative, and neutral.

[1513] Step 5:

[1514] The server generates real-time feedback based on the emotion data and text analysis results. The feedback can be in the form of closed feedback (only for specific participants) or open feedback (for all participants). The input is emotion data and text analysis results, and the output is a feedback message.

[1515] Step 6:

[1516] The device receives the generated feedback and provides it to the participants. Closed feedback is displayed on a specific participant's device, and open feedback is displayed on all participants' devices. The input is the feedback message, and the output is a display of the feedback.

[1517] Step 7:

[1518] The server monitors participants' comments in real time and detects overbearing comments or harassment. If any are detected, a warning message is immediately generated and sent to the relevant participant. The input is text data and emotional data, and the output is a warning message.

[1519] Step 8:

[1520] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. The generated minutes are shared with all participants. The input is all comment data and feedback data, and the output is the generation and sharing of meeting minutes.

[1521] 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.

[1522] 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.

[1523] 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.

[1524] [Fourth embodiment]

[1525] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1526] 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.

[1527] 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).

[1528] 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.

[1529] 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.

[1530] 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).

[1531] 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.

[1532] 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.

[1533] 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.

[1534] 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.

[1535] 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.

[1536] 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.

[1537] 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."

[1538] Program Overview

[1539] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. This system includes "means for collecting participants' comments as audio data," "means for converting the audio data into text data," "means for analyzing the text data and generating real-time feedback," "means for providing feedback to participants," and "means for detecting overbearing comments and harassment and issuing an immediate warning."

[1540] Program processing

[1541] The program of the system of the present invention is processed as follows.

[1542] 1. Audio data collection:

[1543] The device collects speech from meeting participants as audio data in real time, specifically including the ability to capture microphone input from each participant.

[1544] 2. Audio to text conversion:

[1545] The server receives the voice data sent from the device and converts it into text data using a voice recognition API, etc. This makes it possible to analyze the content of the speech as text information.

[1546] 3. Text data analysis:

[1547] The server then analyzes the converted text data using natural language processing (NLP) technology, analyzing keywords, emotions, tone, and other elements within the text to assess the specificity and relevance of the comments.

[1548] 4. Generate real-time feedback:

[1549] The server generates real-time feedback based on the analysis results, including closed and open feedback, in an appropriate format.

[1550] 5. Providing Feedback:

[1551] The terminal receives the generated feedback and provides it to the user. Closed feedback is displayed in the chat window of a specific participant, while open feedback is displayed so that all participants can see it.

[1552] 6. Detect and warn against aggressive speech:

[1553] The server performs real-time emotional analysis of comments and detects statements that are deemed to be overbearing or harassing.

[1554] If the server detects such a situation, it immediately generates a warning message and sends it to the terminal of the participant in question.

[1555] The terminal displays a warning message to the user in real time, prompting the user to take appropriate action.

[1556] 7. Generate and share meeting minutes:

[1557] The server records all comments and feedback during the meeting, and automatically generates minutes after the meeting and provides them to all participants.

[1558] The device displays the generated minutes to the user and provides a link that can be downloaded or shared.

[1559] Specific examples

[1560] Meeting Management

[1561] User (Person A): "I have some opinions about the direction this project should take."

[1562] The device captures Mr. A's speech as audio data and sends it to the server.

[1563] The server converts the voice data into text data and analyzes the content using NLP technology.

[1564] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[1565] Person A receives the feedback and makes a statement providing specific data.

[1566] The device recaptures Mr. A's new comments and sends them to the server.

[1567] The server analyzes the new utterance and provides further feedback.

[1568] In this way, the system provides a fair and effective meeting environment in real time, ensuring that all participants feel comfortable speaking up. Furthermore, the system automatically generates meeting minutes, making it easy to review the content of meetings even after they have ended.

[1569] The processing flow will be explained below.

[1570] Program processing flow

[1571] Step 1: Set up a meeting

[1572] 1. A user schedules a meeting.

[1573] Users log into the system and set the meeting date and time, participant list, and meeting goal.

[1574] 2. The server creates a meeting room.

[1575] The server receives the request from the user, generates a unique meeting room ID, and creates a join link.

[1576] The server pre-loads and prepares the AI ​​models to be used in the meeting.

[1577] 3. The user invites participants.

[1578] The user sends the generated meeting link to the participants, who click the link to enter the meeting room.

[1579] Step 2: Collecting audio data

[1580] 4. The device captures the audio input.

[1581] The device captures microphone input each time a participant speaks and transmits it to the server as audio data in real time.

[1582] Step 3: Convert audio data to text

[1583] 5. The server converts the audio data into text data.

[1584] The server converts the received voice data into text data using a speech recognition API, which converts the voice into text information that can be analyzed.

[1585] Step 4: Analyze the text data

[1586] 6. The server analyzes the text data.

[1587] The server then analyzes the converted text data using natural language processing (NLP) techniques, specifically extracting keywords, analyzing sentiment, and evaluating tone.

[1588] Step 5: Generate real-time feedback

[1589] 7. The server generates real-time feedback.

[1590] Based on the analysis results, the server generates either crowded or open feedback. Closed feedback is given to specific participants, while open feedback is given to all participants.

[1591] 8. The device will display feedback.

[1592] The terminal displays closed feedback to a specific user in a chat window, and displays open feedback in a feedback window that can be viewed by all participants.

[1593] Step 6: Detect and flag coercive language

[1594] 9. The server monitors the comments.

[1595] The server monitors all comments in real time and performs sentiment analysis.

[1596] 10. The server detects overbearing comments.

[1597] The server detects overbearing remarks and harassment based on the results of sentiment analysis.

[1598] 11. The server issues a warning.

[1599] If the server detects an overbearing remark, it generates and sends a warning message to the user.

[1600] 12. The terminal displays a warning message.

[1601] The device will display a warning message to the affected user in real time, urging them to reconsider their comments.

[1602] Step 7: Generate and share meeting minutes

[1603] 13. The server records your comments and feedback.

[1604] The server records all comments and feedback during the meeting and stores them in a database.

[1605] 14. The server generates the minutes.

[1606] The server automatically generates minutes based on the data recorded after the meeting ends.

[1607] 15. The server shares the minutes.

[1608] The server provides the generated minutes to all participants via email or link.

[1609] 16. The device will be able to display and download the minutes.

[1610] The terminal provides a function that allows the user to view and download the minutes.

[1611] Example: Feedback process

[1612] User (Person A): "I have some opinions about the direction this project should take."

[1613] The device captures A's remarks and sends them to the server.

[1614] The server converts the voice data into text data and analyzes the content using NLP technology.

[1615] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or specific data."

[1616] Person A receives the feedback and makes a statement providing specific data.

[1617] The device recaptures the new utterance and sends it to the server.

[1618] The server analyzes the new utterance and provides further feedback.

[1619] Through this process, the system provides a fair and effective meeting environment in real time.

[1620] Example 1

[1621] 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."

[1622] In online and offline meetings, it is necessary to create an environment where all participants can express their opinions fairly and easily, and to prevent discussions from being biased towards certain individuals. There is also a need for systems that can detect overbearing remarks and harassment in real time and respond immediately. Furthermore, to efficiently conduct meetings, it is desirable to provide real-time feedback and automatically generate meeting minutes. However, conventional systems have not been able to adequately resolve these issues.

[1623] 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.

[1624] In this invention, the server includes means for collecting participants' comments as audio data, means for converting the audio data into text data, means for analyzing the text data using natural language processing technology, means for generating real-time feedback based on the analysis results, means for providing the feedback to participants, means for detecting overbearing comments or harassment and generating a warning message, and means for recording the comments and feedback of the meeting and generating minutes after the meeting. This provides an environment in which all participants can speak fairly and easily, making it possible to provide effective feedback in real time. Furthermore, overbearing comments and harassment can be addressed immediately, and the automatically generated minutes make it easy to review the content of the discussion.

[1625] "Participant" means an individual who participates in a meeting or conference.

[1626] "Audio data" refers to data that has been digitized from audio recorded as sound waves.

[1627] "Text data" is data that has been digitized as character information.

[1628] "Natural language processing technology" is a technology for processing and analyzing human language using a computer.

[1629] "Real-time feedback" refers to instant assessment and advice provided during a meeting.

[1630] "Positive speech" is speech that sounds intimidating or aggressive to others.

[1631] "Harassment" refers to behavior or speech that harasses or offends others.

[1632] A "warning message" is a notification that warns you against certain actions or statements.

[1633] "Minutes" are a written record of all statements and discussions made during a meeting or conference.

[1634] "Means of collection" refers to devices or software that capture participants' voices as digital data.

[1635] "Means for converting" refers to a technology or system that converts voice data into text data.

[1636] "Means of analysis" refers to the technology or system used to analyze text data and determine meaning and sentiment.

[1637] "Means for generating" refers to a technology or system that generates feedback or warning messages based on the analysis results.

[1638] "Means for providing" refers to the equipment or software used to deliver generated feedback and warning messages to participants.

[1639] "Means of recording and generating" refers to technology or systems that store comments and feedback during meetings and later compile them into minutes.

[1640] A "system" is a set of devices or software that integrates these means to achieve a set of functions.

[1641] MODE FOR CARRYING OUT THE INVENTION

[1642] This invention is a system that uses AI to provide real-time feedback during online and offline meetings, creating an environment where all participants can easily express their opinions. This system includes "means for collecting participants' comments as audio data," "means for converting the audio data into text data," "means for analyzing the text data using natural language processing technology," "means for generating real-time feedback based on the analysis results," "means for providing feedback to participants," "means for detecting overbearing comments and harassment and generating a warning message," and "means for recording comments and feedback during meetings and generating meeting minutes after the meeting."

[1643] Specifically, the process is as follows:

[1644] Audio data collection

[1645] The device collects speech from each participant during the meeting as real-time audio data using a built-in microphone or an external microphone, and transmits the collected audio data to a server in a compressed format (e.g., MP3 or WAV format).

[1646] Converting audio data to text

[1647] The server receives the voice data sent from the device and converts it into text using a speech recognition API such as Google Cloud Speech-to-Text API or IBM Watson Speech to Text, then temporarily stores the converted text in a database.

[1648] Text data analysis

[1649] The server analyzes the stored text data using natural language processing (NLP) technology. Specifically, it uses Google Cloud Natural Language API and spaCy to evaluate keywords, sentence tone, and sentiment within the text. It evaluates the specificity and relevance of the text data and generates closed-form feedback.

[1650] Generate real-time feedback

[1651] The server generates real-time feedback based on the analysis results, classifies the feedback into closed feedback (for specific participants) and open feedback (for all participants), and sends it to the terminal in the appropriate format.

[1652] Providing Feedback

[1653] The device receives the feedback sent from the server, and displays the closed feedback in the chat window of the specific participant, while the open feedback is displayed on a screen that can be viewed by all participants. Platforms such as Microsoft Teams and Zoom may be used.

[1654] Detecting and warning against coercive speech

[1655] The server analyzes the sentiment of comments in real time and detects any overbearing or harassing content. If any are detected, a warning message is generated and sent to the relevant participant's device.

[1656] The device displays a warning message to the user in real time and prompts the user to take appropriate action, typically using the notification function of the web browser.

[1657] Generate and share meeting minutes

[1658] The server records all comments and feedback during the meeting, automatically generates minutes after the meeting, and generates a link to share the minutes with all participants.

[1659] The device displays the generated minutes to the user and provides download and sharing links, and can output them in Microsoft Word and Google Docs formats.

[1660] Specific examples

[1661] Meeting Management

[1662] User (Person A): "I have some opinions about the direction this project should take."

[1663] The device captures A's speech as audio data and sends it to the server. The captured audio data is saved in WAV format.

[1664] The server uses the Google Cloud Speech-to-Text API to convert Mr. A's voice data into text data in real time. The converted text data is recorded as "I have an opinion about the direction of this project."

[1665] The server analyzes the text and determines that it is "not specific enough" - NLP analysis detects that it does not contain specific keywords.

[1666] The server generates closed feedback such as "It would be more effective if you provided examples or concrete data."

[1667] The device will display this feedback in Mr. A's individual chat window.

[1668] Examples of problematic statements

[1669] User B makes a commanding statement: "Your idea is pointless."

[1670] The device captures Mr. B's speech as audio data and sends it to the server.

[1671] The server converts the voice data into text data and performs sentiment analysis using NLP technology. The analysis results indicate that the remark is "overbearing."

[1672] The server generates a warning message saying, "An overbearing remark has been detected. Please choose appropriate words." and sends it to Mr. B's device.

[1673] The device will notify Mr. B of this warning message in real time, urging him to be careful.

[1674] Example prompts for generative AI models

[1675] Please analyze the following text data and determine whether it constitutes coercive remarks or harassment.

[1676] Data Text: "That suggestion makes absolutely no sense. Anyone could have thought of something like that."

[1677] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1678] Program processing flow

[1679] Step 1: Collecting audio data

[1680] The device collects speech from each participant during a meeting as audio data in real time. The input is the participant's voice, and the output is audio data. Specifically, the device captures audio using a built-in microphone or an external microphone. The collected audio data is temporarily saved in WAV format and sent to a server.

[1681] Step 2: Convert audio data to text

[1682] The server receives the voice data sent from the device. The input is WAV format voice data, and the output is text data. Specifically, the voice data is converted into text data using the Google Cloud Speech-to-Text API or IBM Watson Speech to Text. At this time, the voice data is converted into text data and temporarily stored in a database.

[1683] Step 3: Analyze the text data

[1684] The server analyzes the stored text data using natural language processing (NLP) technology. The input is text data, and the output is the analysis results (keywords, tone, and emotional information). Specifically, it uses Google Cloud Natural Language API and spaCy to extract keywords from the text data, evaluate tone, and analyze emotions. Based on the results of this analysis, the specificity and relevance of the text are evaluated, and feedback candidates are generated.

[1685] Step 4: Generate real-time feedback

[1686] The server generates real-time feedback based on the analysis results. The input is the analysis results, and the output is the feedback message. The server classifies the feedback into closed feedback (for specific participants) and open feedback (for all participants) formats and sends it to the terminal in the appropriate format. The generative AI model is used to generate the feedback text.

[1687] Step 5: Provide feedback

[1688] The device receives feedback sent from the server. The input is the feedback message, and the output is the feedback provided to the user. Specifically, closed feedback is displayed in the chat window of a specific participant, and open feedback is displayed on a shared screen that can be viewed by all participants. The chat functions of Microsoft Teams or Zoom are used.

[1689] Step 6: Detect and flag coercive language

[1690] The server performs real-time sentiment analysis of comments and detects content that is deemed to be overbearing or harassing. The input is text data, and the output is a warning message. If overbearing or harassing comments are detected, a warning message is generated and sent to the relevant participant's device. An appropriate warning message is created using a generative AI model.

[1691] The device displays a warning message to the user in real time, specifically by using the notification function of the web browser, and prompts the user to take appropriate action.

[1692] Step 7: Generate and share meeting minutes

[1693] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting. The input is all comments and feedback during the meeting, and the output is the minutes data.

[1694] The device displays the generated minutes to the user, provides a download link and a sharing link for sharing with all participants, and is configured to output the minutes in Microsoft Word or Google Docs format.

[1695] (Application example 1)

[1696] 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."

[1697] The present invention aims to improve the quality and security of comments made in meetings, provide a fair and safe environment, and provide a system that allows for post-conference review and auditing. In particular, it is necessary to strengthen meeting security by detecting coercive comments and harassment in real time and responding immediately. There is also a need for real-time feedback and automatic generation of meeting minutes so that comments made during meetings can be used as reference material later.

[1698] 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.

[1699] In this invention, the server includes a means for collecting participants' comments as audio data, a means for converting the audio data into text data, a means for analyzing the text data to generate real-time feedback, and a means for detecting comments containing specific keywords and generating a warning. This makes it possible to improve the quality of comments in real time during a meeting, and to detect and warn against overbearing comments and harassment. It is also possible to automatically record the contents of the meeting and generate materials that can be referenced later.

[1700] "Means for collecting participants' comments as audio data" refers to a device or system that captures and records the content of participants' comments in a conference room or on an online platform as audio data in real time.

[1701] The "means for converting voice data into text data" refers to a technology that converts collected voice data into text information using voice recognition technology and outputs it as text data that can be analyzed.

[1702] The "means for analyzing text data and generating real-time feedback" is a system that analyzes the converted text data using natural language processing technology and provides participants with immediate feedback.

[1703] A "means for providing feedback to participants" is a device or system that visually or audibly communicates the generated feedback to participants.

[1704] The "means for detecting statements containing specific keywords and generating warnings" refers to a system that analyzes statements made during a meeting, detects specific keywords (e.g., overbearing statements or harassment) in real time, and generates warning messages based on the detection results.

[1705] MODE FOR CARRYING OUT THE INVENTION

[1706] Program Overview

[1707] This invention provides a system that uses real-time feedback from AI to improve the quality and security of comments made during meetings. The system collects participants' comments as audio data, converts them into text data, and analyzes them. Based on the analysis results, it generates real-time feedback and provides it to participants. The system also has a function to detect comments containing specific keywords and generate a warning. Specific embodiments of the system are described in detail below.

[1708] System configuration

[1709] Hardware and Software

[1710] Microphone: A device that collects speech from conference participants and is connected to a terminal.

[1711] Device (PC, smartphone, smart glasses, head-mounted display, etc.): Receives audio data from the microphone and sends it to the server.

[1712] Server: Converts voice data into text data, analyzes the text data using natural language processing technology, and generates feedback. It also detects utterances containing specific keywords and generates warnings.

[1713] Speech recognition API (such as Google Speech Recognition API): An API for converting voice data into text data.

[1714] Natural language processing libraries (NLTK, VADER, etc.): Libraries for analyzing text data and assessing sentiment and content.

[1715] System Operation

[1716] 1. Collecting and transmitting participant comments:

[1717] The device collects participants' speech in real time through a connected microphone, and the collected voice data is sent to a server.

[1718] 2. Audio to text conversion:

[1719] The server converts the transmitted voice data into text information using a speech recognition API (Google Speech Recognition API).

[1720] 3. Text data analysis:

[1721] The server then analyzes the converted text data using natural language processing (NLTK, VADER) to evaluate the sentiment, tone, and specificity of the comments.

[1722] 4. Generating and Presenting Real-Time Feedback:

[1723] The server generates feedback based on the analysis results and sends it to the device. The device then presents the generated feedback to the participant in real time. For example, feedback such as "It would be more effective if you provided specific data" is displayed on the participant's screen.

[1724] 5. Detecting specific keywords and generating alerts:

[1725] The server detects comments containing specific keywords (e.g., overbearing, harassment), and immediately generates and sends a warning message to the device after the relevant comment is made. The warning is displayed to participants in real time.

[1726] Specific examples

[1727] If participant A says during a meeting, "I have some opinions about the direction this project should take, but I think they're too harsh," the system will collect this comment, convert the audio data into text data, and analyze it. Based on the analysis results, a warning such as "Negative comments have been detected" will be generated and displayed in real time on participant A's device.

[1728] Example prompts for generative AI models

[1729] Analyze the following statements in meetings to detect negative or positive sentiment and generate feedback: "I have some opinions about the direction we should take this project, but I think they're too harsh."

[1730] This system will improve the quality of speech in meetings, detect overbearing remarks and harassment, and provide a safe meeting environment.

[1731] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1732] Step 1:

[1733] Audio data collection

[1734] Subject: Terminal

[1735] Specific operation: The device collects the speech of meeting participants in real time via a microphone.

[1736] Input: Remarks by conference participants (audio data)

[1737] Output: Collected audio data

[1738] Data processing: The device captures the speech as digital audio data and prepares it for transmission to the server.

[1739] Step 2:

[1740] Converting audio data to text

[1741] Subject: Server

[1742] Specific operation: The server receives the voice data sent from the device and converts it into text data using the speech recognition API (Google Speech Recognition API).

[1743] Input: Collected audio data

[1744] Output: Converted text data

[1745] Data processing: Using voice recognition technology, the voice data is analyzed and text data is generated. The server stores the results of this conversion process in a related database.

[1746] Step 3:

[1747] Text data analysis

[1748] Subject: Server

[1749] Specific operation: The server analyzes the converted text data using natural language processing (NLP) techniques (NLTK, VADER).

[1750] Input: Text data

[1751] Output: Analysis results (emotion, tone, specificity, etc.)

[1752] Data processing: The text data is parsed, keywords are extracted, sentiment analysis is performed, and the specificity of the comments is evaluated. The server uses the analysis results to proceed to the next processing step.

[1753] Step 4:

[1754] Generate real-time feedback

[1755] Subject: Server

[1756] Specific behavior: The server generates real-time feedback based on the results of analyzing the text data and sends it to the device. For example, if a comment lacks specificity, the server generates feedback such as, "It would be more effective if you provided specific data."

[1757] Input: Analysis results

[1758] Output: Real-time feedback

[1759] Data processing: Based on the analysis results, feedback messages are generated and sent to the device instructing it to display. The server places particular emphasis on oppressive or negative comments.

[1760] Step 5:

[1761] Providing feedback

[1762] Subject: Terminal

[1763] Specific operation: The device receives the feedback sent from the server and presents it to the participants. Closed feedback is presented to specific participants, and open feedback is presented to all participants.

[1764] Input: Real-time feedback

[1765] Output: Displayed feedback

[1766] Data Calculation: The terminal visually displays feedback messages and prompts specific participants to take appropriate action.

[1767] Step 6:

[1768] Detecting specific keywords and generating alerts

[1769] Subject: Server

[1770] Specific operation: The server detects specific keywords (e.g., overbearing, harassment) in the analyzed text data, and if detected, immediately generates a warning message and sends it to the terminal.

[1771] Input: Analysis results of text data

[1772] Output: Warning message

[1773] Data calculation: The server checks the analysis results against a specific keyword list, and if there is a match, generates and sends a warning message.

[1774] Step 7:

[1775] Present a warning message

[1776] Subject: Terminal

[1777] Specific operation: The terminal receives the warning message sent from the server and displays it to the relevant participant.

[1778] Input: warning message

[1779] Output: The displayed warning message

[1780] Data calculation: The terminal visually displays a warning message and prompts participants who make overbearing remarks to take appropriate action.

[1781] This processing flow enables the system to improve speech quality in real time and enhance meeting security.

[1782] 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.

[1783] Program Overview

[1784] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. Furthermore, this system includes "means for collecting participants' comments as audio data," "means for converting audio data into text data," "means for generating real-time feedback by analyzing the text data," "means for providing feedback to participants," "means for detecting overbearing comments and harassment and issuing an immediate warning," as well as "an emotion engine that recognizes the user's emotions."

[1785] Program processing

[1786] The program of the system of the present invention is processed as follows.

[1787] 1. Audio data collection:

[1788] The device collects speech from meeting participants as audio data in real time, specifically including the ability to capture microphone input from each participant.

[1789] 2. Audio to text conversion:

[1790] The server receives the voice data sent from the device and converts it into text data using a voice recognition API, etc. This converts the voice into text information that can be analyzed.

[1791] 3. Text data analysis:

[1792] The server then analyzes the converted text data using natural language processing (NLP) technology, analyzing keywords, emotions, tone, and other elements within the text to assess the specificity and relevance of the comments.

[1793] 4. Use of Emotion Engine:

[1794] The server then uses an emotion engine to recognize the user's emotions from the text data, extracting emotions from the content and tone of the speech and classifying them into emotion categories such as positive, negative, and neutral.

[1795] 5. Generate real-time feedback:

[1796] The server generates real-time feedback based on the text data and the results of the emotion engine, including closed and open feedback, in an appropriate format.

[1797] 6. Providing Feedback:

[1798] The terminal receives the generated feedback and provides it to the user: closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window that can be viewed by all participants.

[1799] 7. Detect and warn against aggressive speech:

[1800] The server performs real-time emotional analysis of comments and detects statements that are deemed to be overbearing or harassing.

[1801] If the server detects such a situation, it immediately generates a warning message and sends it to the terminal of the participant in question.

[1802] The terminal displays a warning message to the user in real time, prompting the user to take appropriate action.

[1803] 8. Generate and share meeting minutes:

[1804] The server records all comments and feedback during the meeting, and automatically generates minutes after the meeting and provides them to all participants.

[1805] The device displays the generated minutes to the user and provides a link that can be downloaded or shared.

[1806] Specific examples

[1807] Meeting Management

[1808] User (Person A): "I have some opinions about the direction this project should take."

[1809] The device captures Mr. A's speech as audio data and sends it to the server.

[1810] The server converts the voice data into text data and analyzes the content using NLP technology and an emotion engine.

[1811] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[1812] The server also detects negative emotions in Mr. A's comments and provides open feedback such as, "We recommend that you speak with more confidence."

[1813] After receiving the feedback, Person A makes a statement that provides specific data and continues speaking with even more confidence.

[1814] The terminal recaptures the new utterance and transmits it to the server.

[1815] The server again analyzes the new utterance and provides feedback if necessary.

[1816] Through this process, the system provides a fair and effective meeting environment in real time, and by utilizing an emotion engine, it also appropriately manages the emotions of participants. Furthermore, by automatically generating meeting minutes, the content of the meeting can be easily reviewed even after it has ended.

[1817] The processing flow will be explained below.

[1818] MODE FOR CARRYING OUT THE INVENTION

[1819] This invention is a system that enables fair discussions by providing real-time feedback from AI during online and offline meetings, creating an environment where everyone can easily express their opinions. Furthermore, this system includes an "emotion engine" that recognizes the user's emotions and can provide feedback based on their emotions.

[1820] Specific processing steps

[1821] Step 1:

[1822] A user schedules a meeting.

[1823] Users log into the system and set the meeting date and time, participant list, and meeting goal.

[1824] Step 2:

[1825] The server creates a meeting room.

[1826] The server receives the request from the user, generates a unique meeting room ID, and creates a join link.

[1827] The server is pre-loaded and ready with the AI ​​models and emotion engines that will be used for the meeting.

[1828] Step 3:

[1829] The user invites participants.

[1830] The user sends the generated meeting link to prospective participants, who then click the link to enter the meeting room.

[1831] Step 4:

[1832] The device captures audio input.

[1833] The device captures microphone input each time a participant speaks and transmits it to the server as audio data in real time.

[1834] Step 5:

[1835] The server converts the voice data into text data.

[1836] The server converts the received voice data into text data using a speech recognition API, which makes it possible to analyze the spoken content as text information.

[1837] Step 6:

[1838] The server analyzes the text data.

[1839] The server then analyzes the converted text data using natural language processing (NLP) techniques, specifically extracting keywords, analyzing sentiment, and evaluating tone.

[1840] Step 7:

[1841] The server uses an emotion engine to recognize the user's emotion.

[1842] The server then uses an emotion engine to recognize the user's emotions from the text data, extracting emotions from the content and tone of the speech and classifying them into emotion categories such as positive, negative, and neutral.

[1843] Step 8:

[1844] The server generates real-time feedback.

[1845] The server generates real-time feedback based on the analysis of the text data and the emotion engine. The feedback can be either closed or open, and is provided in an appropriate format.

[1846] Step 9:

[1847] The device displays feedback.

[1848] The terminal displays the generated feedback in a chat window for a specific user (closed feedback), and also displays it in a feedback window for all participants (open feedback).

[1849] Step 10:

[1850] The server monitors and detects coercive remarks.

[1851] The server monitors all comments in real time and performs sentiment analysis.

[1852] Detects statements that are deemed to be overbearing or harassment.

[1853] Step 11:

[1854] The server will warn you against overbearing comments.

[1855] When an overbearing remark is detected, the server generates a warning message for the user and sends it to the terminal.

[1856] Step 12:

[1857] The terminal displays a warning message.

[1858] The device will display a warning message to the affected user in real time, urging them to reconsider their comments.

[1859] Step 13:

[1860] The server records your comments and feedback.

[1861] The server records all comments and feedback during the meeting and stores them in a database.

[1862] Step 14:

[1863] The server generates the minutes.

[1864] After the meeting ends, the server automatically generates minutes based on the recorded data.

[1865] Step 15:

[1866] The server shares the minutes.

[1867] The server provides the generated minutes to all participants via email or link.

[1868] Step 16:

[1869] The device will allow you to view and download the minutes.

[1870] The terminal provides a link to allow the user to view and download the minutes.

[1871] Example 2

[1872] 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."

[1873] In traditional online and offline meetings, some participants find it difficult to speak up, and the environment is prone to overbearing remarks and harassment. Furthermore, there is a lack of means to properly record what is being said and provide feedback in real time. As a result, it is difficult to achieve fair discussions.

[1874] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1875] In this invention, the server includes means for collecting participants' comments as voice data, means for converting the voice data into text data, means for analyzing the text data to generate real-time feedback, means for providing the feedback to participants, and means for recognizing user emotions through sentiment analysis, thereby providing a fair and effective meeting environment, appropriately managing user emotions through sentiment analysis, and improving the quality of discussions.

[1876] "Means for collecting participants' speech as audio data" refers to a device or system that records the speech of meeting participants as audio data in real time. A specific example is a device with a microphone.

[1877] "Means for converting voice data into text data" refers to software or algorithms for analyzing voice data and converting it into text data. A specific example is a voice recognition API.

[1878] "Means for analyzing text data and generating real-time feedback" refers to natural language processing technologies and algorithms for evaluating participants' comments based on text data and generating instant feedback.

[1879] "Means for providing feedback to participants" refers to a device or system that displays or notifies participants of generated feedback in real time. Examples include a chat window or a feedback window.

[1880] "Means for recognizing user emotions through sentiment analysis" refers to software or algorithms that extract participants' emotions from text data and classify and recognize those emotions. A specific example is an emotion engine.

[1881] This invention is a system that uses AI to provide real-time feedback in online and offline meetings. This system creates an environment where all participants can easily express their opinions and ensure fair discussions. Furthermore, this system includes a function to recognize user emotions through sentiment analysis.

[1882] The specific configuration of this system is as follows.

[1883] Hardware and software used

[1884] This system is realized by the following hardware and software.

[1885] Hardware: A device with a microphone (e.g., a computer, smartphone, etc.) used to collect the speech of meeting participants.

[1886] software:

[1887] A speech recognition API for converting voice data into text data (e.g., Google Cloud Speech-to-Text API).

[1888] Natural language processing techniques (e.g., TensorFlow, Natural Language Toolkit (NLTK)) for analyzing text data.

[1889] Software for performing sentiment analysis (e.g., IBM Watson Tone Analyzer).

[1890] Processing flow

[1891] 1. Audio data collection:

[1892] The terminal collects speech from meeting participants as audio data in real time. Specifically, a microphone-equipped device captures speech from participants and transmits it to a server.

[1893] 2. Audio to text conversion:

[1894] The server receives the voice data sent from the device and converts it into text data using the Google Cloud Speech-to-Text API.

[1895] 3. Text data analysis:

[1896] The server analyzes the converted text data using natural language processing technology (TensorFlow and Natural Language Toolkit (NLTK)). This analyzes keywords, emotions, tone, etc. in the text and evaluates the specificity and relevance of the statements.

[1897] 4. Use of Emotion Engine:

[1898] The server also uses an emotion engine (IBM Watson Tone Analyzer) to recognize the user's emotions and classify them into emotion categories such as positive, negative, and neutral.

[1899] 5. Generate real-time feedback:

[1900] The server generates real-time feedback based on the collected data and sentiment analysis results, including closed and open feedback.

[1901] 6. Providing Feedback:

[1902] The terminal receives the generated feedback and provides it to the user. Closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window that can be viewed by all participants.

[1903] 7. Detect and warn against aggressive speech:

[1904] The server analyzes the sentiment of comments in real time and detects any statements that are deemed to be overbearing or harassing. If any are detected, a warning message is immediately generated and sent to the relevant participant's device.

[1905] 8. Generate and share meeting minutes:

[1906] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. The device displays the minutes to the user and provides a link that can be downloaded or shared.

[1907] Specific examples

[1908] As a concrete example, we will show how a meeting proceeds.

[1909] User (Person A): "I have some opinions about the direction this project should take."

[1910] The device captures Mr. A's speech as audio data and sends it to the server.

[1911] The server converts the voice data into text data and analyzes the content using natural language processing technology and an emotion engine.

[1912] The server determines that the content is not specific enough and provides closed feedback such as, "It would be more effective if you provided examples or concrete data."

[1913] The server also detects negative emotions in Mr. A's comments and provides open feedback such as, "We recommend that you speak with more confidence."

[1914] Person A receives the feedback, makes statements that provide specific data, and continues to speak with confidence.

[1915] The terminal recaptures the new utterance and transmits it to the server.

[1916] The server then analyzes new comments and provides feedback as necessary. Through this process, the system provides a fair and effective meeting environment in real time, and by utilizing an emotion engine, it also appropriately manages participants' emotions. Furthermore, by automatically generating meeting minutes, the content can be easily reviewed even after the meeting has ended.

[1917] Prompt Sentence Examples

[1918] Prompt: Type what you want to say next in the meeting. The system will provide real-time feedback.

[1919] User says: "I have an opinion on the direction of the project."

[1920] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1921] Step 1:

[1922] Audio data collection

[1923] Input: User's speech (audio)

[1924] Specific operation: Capture participants' speech through a microphone connected to the device.

[1925] Data processing: The captured audio data is converted into a digital signal.

[1926] Output: Digital audio data

[1927] Detailed explanation: The device collects speech from meeting participants as audio data in real time and sends it to a server. For example, a microphone on a laptop installed in a conference room captures each participant's speech and sends the data to the server via Wi-Fi.

[1928] Step 2:

[1929] Converting audio data to text

[1930] Input: Digital audio data

[1931] Specific operation: The server passes the received voice data to the voice recognition API.

[1932] Data processing: Use the Google Cloud Speech-to-Text API to convert the audio data into text data.

[1933] Output: Text data

[1934] Detailed explanation: The server passes the digital voice data sent from the device to the Google Cloud Speech-to-Text API, which converts the voice data into text data. For example, a speech saying "I have an opinion about the direction of the project" is converted into text data saying "I have an opinion about the direction of the project."

[1935] Step 3:

[1936] Text data analysis

[1937] Input: Text data

[1938] Specific operation: The server passes the converted text data to natural language processing technology.

[1939] Data processing: Analyze text for keywords, tone, sentiment, etc. using TensorFlow and the Natural Language Toolkit (NLTK).

[1940] Output: Analysis result data (keywords, tone, sentiment, etc.)

[1941] Detailed explanation: The server uses natural language processing (NLP) technology to analyze text data. For example, keywords are extracted from the text "I have an opinion on the direction of the project" and evaluated as "lacking specificity."

[1942] Step 4:

[1943] Using the Emotion Engine

[1944] Input: Analysis result data (keywords, tone, emotion, etc.)

[1945] Specific operation: The server uses IBM Watson Tone Analyzer to extract and classify emotions.

[1946] Data Processing: The sentiment engine extracts sentiment categories such as positive, negative, and neutral from the analyzed data.

[1947] Output: Sentiment analysis result data

[1948] Detailed explanation: The server further analyzes the parsed text data with a sentiment engine. For example, the phrase "I have an opinion" is classified as negative.

[1949] Step 5:

[1950] Generate real-time feedback

[1951] Input: Analysis result data and sentiment analysis result data

[1952] Specific operation: The server generates real-time feedback based on the collected data.

[1953] Data processing: Generating feedback in the form of closed and open feedback.

[1954] Output: Feedback data

[1955] Detailed explanation: The server generates specific feedback based on the analysis results and sentiment analysis results. For example, closed feedback such as "It would be effective if you provided specific data" is generated, and open feedback such as "We recommend that you speak more confidently."

[1956] Step 6:

[1957] Providing Feedback

[1958] Input: Feedback data

[1959] Specific operation: The terminal displays the generated feedback to the user.

[1960] Data processing: The feedback data is converted into data for display on the user interface.

[1961] Output: Display of feedback message

[1962] Detailed description: The terminal provides the user with real-time feedback sent from the server. For example, closed feedback is displayed in a chat window for a specific participant, and open feedback is displayed in a feedback window visible to all participants.

[1963] Step 7:

[1964] Detecting and warning against coercive speech

[1965] Input: Text data and sentiment analysis result data

[1966] What it does: The server applies rules to detect intrusive and harassing comments.

[1967] Data processing: Detects overbearing remarks and harassment based on sentiment analysis data and generates warning messages.

[1968] Output: Warning message

[1969] Detailed explanation: The server detects overbearing remarks and harassment in real time based on the results of sentiment analysis. For example, if a remark such as "That's impossible!" is judged to be overbearing, a warning message "Please avoid overbearing remarks" is generated and sent to the relevant participant's device.

[1970] Step 8:

[1971] Generate and share meeting minutes

[1972] Input: Speech and feedback data for the entire meeting

[1973] Specific operation: The server records all data during the meeting and generates minutes.

[1974] Data processing: Automatically generate minutes from recorded data and create a link for sharing.

[1975] Output: Meeting minutes data and shared link

[1976] Detailed explanation: The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. For example, after the meeting ends, participants receive an email saying, "All the contents of the meeting have been summarized," and the minutes are displayed on their devices. A link that can be downloaded or shared is also provided.

[1977] This system provides a fair and effective meeting environment, manages participants' emotions appropriately, and improves the quality of discussions. Automatically generating meeting minutes also makes it easy to review important information.

[1978] (Application example 2)

[1979] 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."

[1980] In physical stores, staff meetings and training sessions often lack the fairness and specificity of statements and proper feedback. It's also difficult to detect coercive remarks or harassment in real time and respond immediately, and there's no system for automatically generating and providing meeting minutes to participants. As a result, fair and effective discussions and organizational management can be difficult.

[1981] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1982] In this invention, the server includes means for collecting participants' comments as voice data, means for converting the voice data into text data, means for analyzing the text data to generate real-time feedback, means for providing the feedback to participants, means for recognizing emotions in the text data, means for generating feedback based on the recognized emotions, means for detecting coercive remarks and harassment and issuing immediate warnings, and means for automatically generating meeting minutes and providing them to participants. This makes meetings and training sessions in brick-and-mortar stores fair and effective, and enables real-time feedback, detection and warning of coercive remarks, and automatic generation of meeting minutes.

[1983] "Means for collecting participants' speech as audio data" refers to devices or systems that record the speech of participants in conferences or meetings in real time and store it as digital audio data.

[1984] "Means for converting voice data into text data" refers to a device or system that converts collected voice data into text information using natural language processing technology or voice recognition technology.

[1985] The "means for analyzing text data and generating real-time feedback" refers to a device or system for analyzing the converted text data and generating instantaneous feedback based on the content and sentiment.

[1986] The "means for providing feedback to participants" refers to a device or system for displaying or notifying the participants of the conference or meeting of the generated feedback in an appropriate format.

[1987] A "means for recognizing emotions in text data" is a device or system that uses natural language processing or machine learning techniques to extract and classify emotions from the content and tone of text data.

[1988] The "means for generating feedback based on recognized emotions" refers to a device or system that generates feedback appropriate to the emotions in real time based on analyzed emotional data.

[1989] "Means for detecting coercive remarks and harassment and issuing immediate warnings" refers to devices or systems that automatically detect parts of a participant's remarks that are deemed to be coercive or harassing and immediately display a warning.

[1990] "Means for automatically generating meeting minutes and providing them to participants" refers to a device or system that records all comments and feedback made during a meeting, and automatically creates and distributes minutes to participants after the meeting ends.

[1991] This invention aims to realize a system that uses smart devices and a server to provide real-time feedback for staff meetings and training sessions in brick-and-mortar stores. Specific embodiments of this system are described below.

[1992] The system uses a microphone built into a smart device (e.g., a smartphone or tablet) to collect participants' speech as audio data. After the speech is collected, the audio data is transferred from the smart device to a server.

[1993] The server uses speech recognition software to convert the voice data into text data. This can be effectively done using existing speech recognition technology, such as Google's speech recognition API. The server then analyzes the text data using natural language processing (NLP) techniques. For this purpose, the Transformers library provided by Hugging Face can be used.

[1994] The parsed text data is then analyzed using an emotion engine (e.g., nlptown / bert-base-multilingual-uncased-sentiment model) to recognize emotions. The server uses this emotion data to generate real-time feedback and provides it to participants. The feedback can be displayed on the screen of a smart device or announced via audio.

[1995] Furthermore, the server uses text data and emotional data to detect overbearing remarks and harassment. If detected, it generates feedback to immediately warn the relevant participants and notifies them in real time. All remarks and feedback are also recorded, and minutes are automatically created and shared with all participants after the meeting ends.

[1996] For example, if someone says during a meeting, "I have an opinion on the direction of this project. Next time, let's present more concrete data. Also, speaking based on facts will be more persuasive," the system will capture this statement, convert it from audio data to text data, and generate real-time feedback along with emotion recognition.

[1997] As described above, this system can be used to ensure that meetings and training sessions in physical stores are conducted fairly and effectively, providing an environment in which participants' opinions are properly reflected.

[1998] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1999] Step 1:

[2000] When the device detects the start of a meeting, it collects each participant's speech as audio data. At this time, it uses the smart device's microphone to capture the audio in real time. The input is the participant's voice, and the output is digital audio data. This data is immediately sent to the server.

[2001] Step 2:

[2002] The server converts the received voice data into text data using Google's speech recognition API. The input is digital voice data, and the output is text data converted from this voice data into character string information. This conversion turns the voice information into analyzable text information.

[2003] Step 3:

[2004] The server analyzes the text data and uses natural language processing (NLP) technology to analyze the content. Specifically, it extracts keywords, tone, and utterance relationships within the text. The input is text data generated by a speech recognition API, and the output is the analysis results. Emotion recognition of the text data is also performed based on the analysis results.

[2005] Step 4:

[2006] The server uses an emotion engine (e.g., nlptown / bert-base-multilingual-uncased-sentiment model) to recognize the emotion of text data. The input is text data, and the output is recognized emotion data. This emotion data is classified into categories such as positive, negative, and neutral.

[2007] Step 5:

[2008] The server generates real-time feedback based on the emotion data and text analysis results. The feedback can be in the form of closed feedback (only for specific participants) or open feedback (for all participants). The input is emotion data and text analysis results, and the output is a feedback message.

[2009] Step 6:

[2010] The device receives the generated feedback and provides it to the participants. Closed feedback is displayed on a specific participant's device, and open feedback is displayed on all participants' devices. The input is the feedback message, and the output is a display of the feedback.

[2011] Step 7:

[2012] The server monitors participants' comments in real time and detects overbearing comments or harassment. If any are detected, a warning message is immediately generated and sent to the relevant participant. The input is text data and emotional data, and the output is a warning message.

[2013] Step 8:

[2014] The server records all comments and feedback during the meeting and automatically generates minutes after the meeting ends. The generated minutes are shared with all participants. The input is all comment data and feedback data, and the output is the generation and sharing of meeting minutes.

[2015] 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.

[2016] 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.

[2017] 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.

[2018] 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.

[2019] 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.

[2020] 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.

[2021] 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).

[2022] 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.

[2023] 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."

[2024] 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.

[2025] 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).

[2026] 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.

[2027] 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.

[2028] 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.

[2029] 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.

[2030] 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.

[2031] 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.

[2032] 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.

[2033] 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.

[2034] 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.

[2035] 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.

[2036] The following is further disclosed regarding the above embodiment.

[2037] (Claim 1)

[2038] A means of collecting participants' comments as audio data;

[2039] means for converting voice data into text data;

[2040] a means for analyzing the text data to generate real-time feedback;

[2041] a means of providing feedback to participants;

[2042] A system including:

[2043] (Claim 2)

[2044] 10. The system of claim 1, wherein the system generates at least one of closed feedback and open feedback.

[2045] (Claim 3)

[2046] 10. The system of claim 1, further comprising means for detecting and immediately issuing a warning against intrusive remarks or harassment.

[2047] (Claim 4)

[2048] 10. The system of claim 1, further comprising means for recording the generated feedback and speech data and automatically generating minutes after the meeting is over.

[2049] (Claim 5)

[2050] 10. The system of claim 1, further comprising means for managing the initiation of meetings and the invitation of participants.

[2051] "Example 1"

[2052] (Claim 1)

[2053] A means of collecting participants' comments as audio data;

[2054] means for converting voice data into text data;

[2055] A means for analyzing text data using natural language processing technology;

[2056] means for generating real-time feedback based on the analysis results;

[2057] a means of providing feedback to participants;

[2058] A means of detecting coercive remarks and harassment and generating a warning message;

[2059] A way to record comments and feedback during meetings and generate minutes after the meeting has ended.

[2060] A system including:

[2061] (Claim 2)

[2062] 10. The system of claim 1, wherein the system generates at least one of closed feedback and open feedback.

[2063] (Claim 3)

[2064] 10. The system of claim 1, further comprising means for detecting and immediately issuing a warning against intrusive remarks or harassment.

[2065] "Application Example 1"

[2066] (Claim 1)

[2067] A means of collecting participants' comments as audio data;

[2068] means for converting voice data into text data;

[2069] a means for analyzing the text data to generate real-time feedback;

[2070] a means of providing feedback to participants;

[2071] A means for detecting statements containing specific keywords and generating a warning;

[2072] ...

[2073] A system including:

[2074] (Claim 2)

[2075] 10. The system of claim 1, wherein the system generates at least one of closed feedback and open feedback.

[2076] (Claim 3)

[2077] 10. The system of claim 1, further comprising: means for presenting the generated feedback in real time.

[2078] "Example 2: Combining Emotion Engines"

[2079] (Claim 1)

[2080] A means of collecting participants' comments as audio data;

[2081] means for converting voice data into text data;

[2082] a means for analyzing the text data to generate real-time feedback;

[2083] a means of providing feedback to participants;

[2084] means for recognizing user emotions through sentiment analysis;

[2085] A system including:

[2086] (Claim 2)

[2087] 10. The system of claim 1, wherein the system generates at least one of closed feedback and open feedback.

[2088] (Claim 3)

[2089] 10. The system of claim 1, further comprising means for detecting and immediately issuing a warning against intrusive remarks or harassment.

[2090] "Application example 2 when combining emotion engines"

[2091] (Claim 1)

[2092] A means of collecting participants' comments as audio data;

[2093] means for converting voice data into text data;

[2094] a means for analyzing the text data to generate real-time feedback;

[2095] a means of providing feedback to participants;

[2096] a means for recognizing sentiment in text data;

[2097] a means for generating feedback based on the recognized emotion;

[2098] A means to detect and immediately warn against coercive remarks and harassment,

[2099] A means to automatically generate meeting minutes and provide them to participants;

[2100] A system including:

[2101] (Claim 2)

[2102] 10. The system of claim 1, wherein the system generates at least one of closed feedback and open feedback.

[2103] (Claim 3)

[2104] The system of claim 1, wherein participants' comments are collected using smart devices. [Explanation of symbols]

[2105] 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 of collecting participants' comments as audio data; means for converting voice data into text data; a means for analyzing the text data to generate real-time feedback; a means of providing feedback to participants; A system including:

2. The system of claim 1 , wherein the system generates at least one of closed feedback and open feedback.

3. 10. The system of claim 1, further comprising means for detecting and immediately issuing a warning against intrusive remarks or harassment.

4. The system of claim 1 , further comprising means for recording the generated feedback and speech data and automatically generating minutes after the meeting is over.

5. The system of claim 1 further comprising means for managing the initiation of meetings and the invitation of participants.

Images