System

Abstract

A system is provided.SOLUTION: A system comprising: means for recording conference voices in real time; means for dividing the recorded conference voices into data packets and transmitting the data packets to a server; means for converting voice data received on the server side into text data; means for analyzing the text data and scoring statement contents of conference participants; means for notifying a user terminal of a scoring result; and means for storing the analysis result for a long period of time and accumulating the analysis result as evaluation data.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] Create a "problem the invention aims to solve" and a "means for solving the problem."

[0005] Traditional meetings face problems such as poor evaluation of small meetings or closed environments such as one-on-one meetings, which can lead to a decline in meeting quality, overlooking harassment, and difficulty in properly evaluating participants' contributions and leadership. Harassment issues, which can be particularly damaging to companies, require effective monitoring measures to detect and address them early. Furthermore, there is a lack of fair methods for evaluating each individual's facilitation and problem-solving skills, and there is a lack of trend analysis based on long-term data accumulation. [Means for solving the problem]

[0006] The purpose of this invention is to provide a system that records conference audio in real time, divides the audio data into data packets, and sends them to a server. The server converts the received audio data into text data and analyzes it using generative AI. Based on the analysis, the content of the participants' comments is scored and evaluated for leadership, cooperation, and the presence or absence of harassment. The results are then sent from the server to the user's device, providing detailed feedback to improve the quality of the meeting. In addition, the analysis results can be stored over the long term and accumulated as evaluation data, enabling regular backups and analysis of long-term evaluation trends. This enables fair evaluations and appropriate responses, reducing damage to the company.

[0007] "Conference audio" refers to the content of remarks and conversations made during a conference recorded as audio signals.

[0008] "Real-time" means processing and analyzing events at the same time they occur, and responding immediately without delay.

[0009] "Recording means" includes any device or method for storing audio data or other information, such as a microphone and recording device.

[0010] A "data packet" is a small unit of data that is divided into large amounts for transmission, and is used for efficient transfer over a network.

[0011] A "server" is a computer system or component thereof that stores, processes, transmits, and receives data on a network.

[0012] "Text data" is written information generated through voice recognition or manual input, and can be stored and processed in digital form.

[0013] "Analysis" is the process of examining data in detail, extracting meaning, and providing relevant information and evaluations.

[0014] "Scoring" refers to quantitatively evaluating data or statements by assigning a score based on specific criteria or indicators.

[0015] A "user terminal" is a device that is directly operated and used by a user, and examples include PCs, smartphones, tablets, etc.

[0016] "Feedback" refers to information such as advice, evaluation comments, and improvement measures provided based on the results of evaluation and analysis.

[0017] A "database" is a system or software for structuring and storing data and efficiently searching, updating, and managing it.

[0018] "Backup" is the process of making a copy of data and storing it in preparation for recovery in the event of loss or corruption.

[0019] "Evaluation trends" refer to fluctuations and patterns in evaluation results based on data collected over a long period of time.

[0020] "Trend analysis" is the use of past data to predict and analyze future trends and tendencies.

[0021] "Harassment" refers to inappropriate conduct toward another person, including threatening, abusive, discriminatory, or offensive conduct.

[0022] "Leadership" refers to the ability and behavior to effectively lead a team or group and guide them to achieve their goals.

[0023] "Cooperativeness" refers to the ability and attitude to cooperate smoothly with others and work together.

[0024] "Generative AI" refers to systems or algorithms that use artificial intelligence techniques to generate, analyze, and provide feedback on data. [Brief explanation of the drawings]

[0025] [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

[0026] 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.

[0027] First, the terms used in the following description will be explained.

[0028] 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).

[0029] 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.

[0030] 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.

[0031] 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.

[0032] 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."

[0033] [First embodiment]

[0034] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0035] 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.

[0036] 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).

[0037] 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.

[0038] 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.

[0039] 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.

[0040] 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.

[0041] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0042] 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.

[0043] 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.

[0044] 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.

[0045] 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."

[0046] This invention relates to a meeting evaluation system for improving the quality of meetings. This system records and analyzes meeting audio in real time and scores the remarks of meeting participants to monitor harassment and evaluate facilitation skills.

[0047] System configuration

[0048] The system consists of the following main components:

[0049] 1. User device (PC, smartphone, tablet, etc.)

[0050] 2. Server (a computer system that receives and analyzes voice data)

[0051] 3. Network (communications infrastructure connecting user devices and servers)

[0052] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided.

[0053] Specific operation example

[0054] A specific example of the operation of this system is shown below.

[0055] Meeting Scenarios

[0056] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[0057] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[0058] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[0059] 3. The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine (e.g., voice recognition API).

[0060] 4. The converted text data is input into a generative AI (e.g., a generative AI model) to perform a detailed analysis of the meeting content.

[0061] 5. Based on the analysis results, the server scores participants on their leadership, cooperativeness, and whether or not they engage in harassment.

[0062] For example, user A's leadership score is 85, user B's contribution score is 70, cooperativeness score is 90, and no harassment is detected.

[0063] 6. The server sends the scoring results to the user devices of User A and User B and displays the results visually through the user interface.

[0064] 7. The device will provide detailed score information and feedback comments from the generated AI, allowing users to review their results.

[0065] 8. The server records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[0066] Explanation of program processing

[0067] 1. A user starts a meeting and launches the recording app.

[0068] 2. The terminal captures the conference audio and divides it into data packets in real time.

[0069] 3. The terminal sequentially transmits voice data packets to the server.

[0070] 4. The server receives the audio data packets and concatenates them to form an audio stream.

[0071] 5. The server uses a speech recognition engine to convert the speech to text.

[0072] 6. The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[0073] 7. The server sends the analysis results to the terminal and displays them on the user interface.

[0074] 8. The server records the results in a database and stores them as long-term data.

[0075] Through these processes, the present invention can improve the quality of meetings and achieve fair evaluations. In addition, the harassment monitoring function can effectively manage risk within a company. In this way, the embodiments for carrying out the invention have been specifically described.

[0076] The processing flow will be explained below.

[0077] Step 1:

[0078] A user starts a conference and launches the recording application on their device. When the user presses the "Start Recording" button, the device captures the conference audio through the audio input device (microphone).

[0079] Step 2:

[0080] The terminal divides the recorded conference audio into data packets in real time. The terminal divides the captured audio data into small packets with a certain byte size and transmits them sequentially to the server via the network.

[0081] Step 3:

[0082] The server rearranges the received audio data packets in order and concatenates them in a stream format. The server checks the data integrity and requests packet retransmission if necessary.

[0083] Step 4:

[0084] The server inputs the formatted audio stream into a speech recognition engine (e.g., speech recognition API) and converts the audio data into text data. The server receives the converted text data from the speech recognition engine and temporarily stores it in a database.

[0085] Step 5:

[0086] The server inputs the text data into a generative AI (e.g., a generative AI model) and instructs the generative AI to "analyze the content of the meeting based on this text data and score it on indicators such as leadership, cooperation, and the presence or absence of harassment."

[0087] Step 6:

[0088] The server receives the analysis results output by the generation AI and extracts scores and detailed feedback for each indicator. For example, User A's leadership score might be 85 points, User B's contribution score 70 points, cooperation score 90 points, and no harassment detected.

[0089] Step 7:

[0090] The server sends the scoring results and detailed feedback comments in JSON format to the user's device. The server records the result transmission log and checks the success or failure status.

[0091] Step 8:

[0092] The device receives the JSON data, parses it, and displays it in the user interface, where the user can visually check the score details and feedback from the generating AI.

[0093] Step 9:

[0094] The server structures the scoring results and analysis and records them in a database. The recorded data is backed up regularly and stored as long-term evaluation data.

[0095] Step 10:

[0096] The server generates reports based on the accumulated evaluation data, analyzing long-term evaluation trends and other trends, and provides these reports to administrators to help improve the quality of corporate meetings.

[0097] This enables the meeting evaluation system to improve the quality of meetings and provide fair evaluations and appropriate feedback quickly.

[0098] Example 1

[0099] 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."

[0100] Conventional meeting evaluation systems often manually record and analyze meeting audio, making it difficult to effectively evaluate meeting quality. They also lacked a means to objectively evaluate participants' leadership and cooperation, or to automatically detect harassment. As a result, it was difficult to provide feedback to improve meeting quality, and risk management across the company was inadequate.

[0101] 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.

[0102] In this invention, the server includes a means for dividing the recorded conference audio into data packets and transmitting them to the server, a means for converting the received audio data into text data, a means for inputting the converted text data into a generative AI model for analysis and scoring the speech content of the conference participants, a means for notifying the user terminal of the scoring results, and a means for long-term storage of the analysis results and accumulating them as evaluation data. This enables effective evaluation of the quality of the conference and provides detailed feedback to participants. It also automates the evaluation of leadership and cooperation and the monitoring of harassment, strengthening risk management in companies.

[0103] "Conference audio" refers to all audio data spoken during a conference.

[0104] "Recording means" refers to the equipment or method for recording conference audio in real time and converting it into digital data.

[0105] A "data packet" refers to a data unit that divides a series of audio data into a certain size and transmits it over a network.

[0106] "Server" refers to a computing system that processes data received from a terminal, generates analysis results, and manages them.

[0107] "Audio data" refers to data in which conference audio has been converted into digital format.

[0108] "Text data" refers to character string data converted from voice data using voice recognition technology.

[0109] A "generative AI model" refers to an artificial intelligence model that performs natural language processing and analysis based on large amounts of data.

[0110] "Analyzing" refers to the process of using certain algorithms or models to make sense of input data and extract relevant information.

[0111] "Content of remarks" refers to the general content of remarks made by conference participants during the conference.

[0112] "Scoring" refers to assigning points based on analysis results according to certain evaluation criteria.

[0113] "User terminal" refers to a device used by a conference participant to record voice and display the analysis results.

[0114] "Means for notifying" refers to a method or device for transmitting information from a server to a user terminal and informing the user.

[0115] "Storage means" refers to methods and devices for retaining data for long periods of time.

[0116] "Evaluation data" refers to data including analysis results and scoring results of meetings.

[0117] "Detailed feedback" refers to specific and analytical evaluations and information on areas for improvement provided to meeting participants.

[0118] "Leadership" refers to the criteria used to evaluate meeting participants' ability to lead and provide direction.

[0119] "Collaborative ability" refers to a criterion that evaluates meeting participants' ability to cooperate and collaborate as team members.

[0120] "Harassment" refers to inappropriate comments or behavior that may occur during a meeting.

[0121] This invention relates to a meeting evaluation system for improving the quality of meetings. This system records and analyzes meeting audio in real time and scores the remarks made by meeting participants to monitor harassment and evaluate facilitation skills. The detailed configuration and operation of the system are shown below.

[0122] System configuration

[0123] The system consists of the following main components:

[0124] 1. User device (e.g. PC, smartphone, tablet, etc.)

[0125] 2. Server (a computer system that receives and analyzes voice data)

[0126] 3. Network (communications infrastructure connecting user devices and servers)

[0127] Operational Overview

[0128] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted to text data on the server side and further analyzed using a generative AI model. The analysis results are scored based on the content of the speech, and various indicators (such as leadership, cooperation, and the presence or absence of harassment) are evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided.

[0129] Specific operation example

[0130] A specific example of the operation of this system is shown below.

[0131] Meeting Scenarios

[0132] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[0133] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[0134] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[0135] 3. The server arranges and concatenates the received voice data in order and converts it into text data using a voice recognition engine (e.g., Google Cloud Speech-to-Text).

[0136] 4. The converted text data is input into a generative AI (e.g., GPT-4 model) to perform a detailed analysis of the meeting content.

[0137] 5. Based on the analysis results, the server scores participants on their leadership, cooperativeness, and whether or not they engage in harassment.

[0138] For example, user A's leadership score is 85, user B's contribution score is 70, cooperativeness score is 90, and no harassment is detected.

[0139] 6. The server sends the scoring results to the user's device and displays the results visually through the user interface.

[0140] 7. The device will provide detailed score information and feedback comments from the generated AI, allowing users to review their results.

[0141] 8. The server records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[0142] Example of input prompt for generative AI model

[0143] Example prompt:

[0144] Analyze the audio recording of the meeting below and rate each participant's leadership, cooperation, and whether or not they engaged in harassment.

[0145] text:

[0146] User A: Today's meeting is about the progress of the new project. Everyone, please give us your opinions.

[0147] User B: I'll be in charge of the progress report. Currently, Task A is 80% complete.

[0148] ...

[0149] In this way, by showing detailed embodiments of the invention, it is possible to effectively evaluate the quality of meetings and provide clear feedback to participants through scoring of leadership and cooperation, which is expected to improve meeting management and strengthen corporate risk management.

[0150] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0151] Step 1:

[0152] A user starts a meeting and launches the recording app.

[0153] Specifically, when a meeting starts, the user launches a dedicated recording application on their device.

[0154] Specific operation: For example, User A double-clicks the meeting recording app on his / her PC and clicks the "Start Recording" button.

[0155] Input: Conference audio (analog audio)

[0156] Output: Launched recording application

[0157] Step 2:

[0158] The terminal captures the conference audio and divides it into data packets in real time.

[0159] Specific description: The terminal records the conference audio through a microphone and performs the process of dividing it into small data packets.

[0160] What it does: The PC's microphone captures audio data every second, and then divides each into data packets every 0.5 seconds.

[0161] Input: Conference audio (analog audio)

[0162] Output: Data packet (digital data)

[0163] Step 3:

[0164] The terminal sequentially transmits voice data packets to the server.

[0165] Specific explanation: The terminal sequentially transmits the divided voice data packets to the server.

[0166] What it does: Your PC sends packetized voice data to a server every second via your internet connection.

[0167] Input: Data packets (digital voice data)

[0168] Output: Data packet sent to the server

[0169] Step 4:

[0170] The server receives the audio data packets and concatenates them to form an audio stream.

[0171] Specifically, the server concatenates the received audio data packets in order to create a continuous audio stream.

[0172] Specific operation: The server combines the audio data packets received every 0.5 seconds to restore the original continuous audio data.

[0173] Input: Data packets (digital voice data)

[0174] Output: Concatenated audio stream (continuous audio data)

[0175] Step 5:

[0176] The server uses a speech recognition engine to convert the speech into text.

[0177] Specifically, the server uses a speech recognition engine (e.g., Google Cloud Speech-to-Text) to convert the voice data into text data.

[0178] What happens: The server sends the audio file to a speech recognition API and retrieves the conversation in text format.

[0179] Input: Concatenated audio stream (continuous audio data)

[0180] Output: Text data (string format)

[0181] Step 6:

[0182] The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[0183] Specific explanation: The server inputs text data into a generative AI (e.g., GPT-4 model) to analyze the meeting content and score what is said.

[0184] Specific behavior: The server passes the following prompt to the generation AI:

[0185] Example prompt:

[0186] Analyze the audio recording of the meeting below and rate each participant's leadership, cooperation, and whether or not they engaged in harassment.

[0187] text:

[0188] User A: Today's meeting is about the progress of the new project. Everyone, please give us your opinions.

[0189] User B: I'll be in charge of the progress report. Currently, Task A is 80% complete.

[0190] ...

[0191] Input: Text data (string format)

[0192] Output: Analysis results and scoring data (evaluation index)

[0193] Step 7:

[0194] The server sends the analysis results to the terminal and displays them on the user interface.

[0195] Specific explanation: The server sends the analysis results returned by the generation AI to the user terminal and displays them on the user interface.

[0196] Specific operation: The server sends the evaluation results (e.g., User A's leadership score is 85 points, User B's contribution score is 70 points) to the devices of User A and User B, and visualizes them on a dedicated dashboard.

[0197] Input: Analysis results and scoring data (evaluation index)

[0198] Output: Evaluation results displayed on the user's terminal

[0199] Step 8:

[0200] The server records the results in a database and accumulates them as long-term evaluation data.

[0201] Specific explanation: The server stores the analysis results and each meeting data in a database and manages them as long-term evaluation data.

[0202] Specific operation: The server records the evaluation results and speech analysis data in a MySQL database for future comparative analysis.

[0203] Input: Analysis results and scoring data (evaluation index)

[0204] Output: Evaluation data stored in a database

[0205] This system improves the quality of meetings and enables fair evaluations. It supports productive meeting management while also providing a harassment monitoring function, strengthening risk management across the company.

[0206] (Application example 1)

[0207] 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."

[0208] Conventional meeting evaluation systems can evaluate the quality of meetings and the behavior of participants in real time, but they are unable to evaluate in-car communication or safety awareness. Furthermore, there is a lack of systems that can analyze in-car conversations in detail to evaluate the driver's attention and whether or not harassment has occurred. This has resulted in insufficient prevention of harassment and improvement of safety awareness in the in-car environment.

[0209] 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.

[0210] In this invention, the server includes means for recording conference audio in real time, means for dividing the recorded conference audio into data packets and transmitting them to the server, means for converting the received audio data into text data, means for analyzing the text data and scoring the speech content of the conference participants, means for recording in-car conversations in real time, means for scoring the driver's attention, safety awareness, and presence or absence of harassment through analysis, means for notifying the scoring results to a user terminal, and means for long-term storage of the analysis results and accumulating them as evaluation data. This makes it possible to evaluate in-car communication quality, safety awareness, and presence or absence of harassment in detail. This also makes it possible to appropriately monitor the behavior of drivers and passengers and improve the safety and comfort of the in-car environment.

[0211] "Conference audio" refers to the audio data of the participants' comments and interactions that occur during a conference.

[0212] A "data packet" is a small unit of digital data divided into pieces and formatted for efficient transmission over a network.

[0213] A "server" is a central computer system for collecting, analyzing, and storing data.

[0214] "Text data" refers to text information converted from speech using speech recognition technology.

[0215] "Scoring" is an evaluation method in which points are assigned according to specific evaluation indicators based on analyzed data.

[0216] "User terminal" refers to a device that is directly operated by a user, such as a PC, smartphone, or tablet.

[0217] "Real-time" refers to data being processed and transferred immediately with little delay.

[0218] "Attention" is an indicator of how much a driver is concentrating on traffic conditions and driving operations.

[0219] "Harassment" refers to inappropriate words, actions or behaviour towards other crew members.

[0220] "Analysis" is the process of analyzing collected data in detail to clarify its meaning and trends.

[0221] "Notification" refers to the act of informing the user of the analysis results or scoring results.

[0222] "Evaluation data" refers to data that includes analysis results and scoring results, and is used as the basis for long-term evaluation and improvement.

[0223] This invention relates to an in-car communication evaluation system. The purpose is to improve the safety and comfort of the in-car environment by recording and analyzing in-car conversations in real time and scoring passenger attention, safety awareness, and the presence or absence of harassment. This system consists of the following main components:

[0224] System configuration

[0225] 1. User Device

[0226] This refers to in-vehicle systems, smartphones, tablets, etc. that have recording functions and are responsible for recording conversations in real time.

[0227] 2. Server

[0228] It is the central computer system that collects and analyzes data and stores the results.

[0229] 3. Network

[0230] It is a communications infrastructure that connects user terminals and servers, and uses wireless communication technology.

[0231] Specific operation example

[0232] A specific example of the operation of this system is shown below.

[0233] In-car communication scenario

[0234] Consider a situation in which user A (driver) and user B (passenger) are having a conversation in a car.

[0235] 1. User terminal: The vehicle's in-vehicle system activates the recording function and presses the "Start Recording" button.

[0236] 2. Terminal: Records conversations inside the car in real time, divides the voice data into data packets, and sends them to the server sequentially.

[0237] 3. Server: The received voice data is arranged in order and concatenated, and the voice data is converted into text data using a voice recognition engine (e.g., speech_recognition library).

[0238] 4. Server: The converted text data is input into a generative AI model (e.g., GPT-3) for detailed analysis.

[0239] 5. Server: Based on the analysis results, the server scores participants on their attention, safety awareness, and whether or not they have engaged in harassment.

[0240] For example, user A's attention score is 85, user B's safety awareness score is 90, and no harassment is detected.

[0241] 6. Server: Sends the scoring results to the user's terminal and displays the results visually through the user interface.

[0242] 7. Device: Provides detailed score information and feedback comments from the generated AI, allowing users to check their results.

[0243] 8. Server: Records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[0244] Hardware and software used

[0245] Speech recognition engine: speech_recognition library

[0246] Generative AI model: GPT-3

[0247] User devices: in-vehicle systems, smartphones, tablets, etc.

[0248] Server: a computer system for data analysis and storage

[0249] Network: Wireless communication technology (e.g., Wi-Fi, 4G, 5G)

[0250] Adding specific examples

[0251] Driving conversation scenario:

[0252] Expect conversations such as, "Are you concentrating on driving?", "Drive carefully," and "We're not in a hurry, so drive safely."

[0253] Example prompts to input to a generative AI model:

[0254] Based on the conversation below, please rate the level of attention, safety awareness, and whether or not there was any harassment.

[0255] example:

[0256] How was your drive today?

[0257] Drive carefully. We're not in a hurry, so safety comes first.

[0258] I think you are being harsh towards me.

[0259] The present invention thereby provides a means for evaluating in-vehicle communications in real time and for appropriately monitoring occupant behavior.

[0260] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0261] Step 1:

[0262] The user terminal activates the recording function of the vehicle's in-vehicle system and operates the "start recording" button.

[0263] Input: Start recording

[0264] Output: Real-time recording begins

[0265] Specific operation: The user starts recording on the in-car system or smartphone. The device prepares to record the conversation in real time.

[0266] Step 2:

[0267] The terminal records conversations in the car in real time, divides the voice data into data packets, and transmits them to the server one by one.

[0268] Input: Voice data of conversations in a car

[0269] Output: sent as data packets to the server

[0270] Specific operation: The device divides the recorded audio into small data packets and sends them to the server over the network.

[0271] Step 3:

[0272] The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine.

[0273] Input: Packetized voice data

[0274] Output: Converted text data

[0275] Specific operation: The server uses a speech recognition library (e.g., speech_recognition) to analyze the voice data and convert it into text data.

[0276] Step 4:

[0277] The server inputs the converted text data into a generative AI model (e.g., GPT-3) for detailed analysis.

[0278] Input: Text data

[0279] Output: Analysis results (e.g., scores for attention, safety awareness, and whether or not there was harassment)

[0280] Specific operation: The server inputs prompts into the generative AI model, which then performs detailed analysis and scoring based on the conversation content.

[0281] Step 5:

[0282] Based on the analysis results, the server scores the person's attention, safety awareness, and whether or not they have engaged in harassment.

[0283] Input: Analysis results from a generative AI model

[0284] Output: Each score (e.g., attention 85 points, safety awareness 90 points, no harassment detected)

[0285] Specific operation: The server quantifies specific indicators as scores based on the output of the generation AI.

[0286] Step 6:

[0287] The server transmits the scoring results to the user terminal and visually displays the results through a user interface.

[0288] Input: Scoring results

[0289] Output: Visual feedback displayed in the user interface

[0290] Specific operation: The server sends the score and analysis results to the user's device, which then displays the results on the screen.

[0291] Step 7:

[0292] The device provides the user with detailed score information and feedback comments from the generated AI for review.

[0293] Input: Scoring results received from the server

[0294] Output: Feedback comments that the user sees

[0295] Specific operation: The terminal displays the scoring results and feedback comments for the user to review.

[0296] Step 8:

[0297] The server records the scoring results and analysis details in a database and accumulates them as long-term evaluation data.

[0298] Input: Scoring results and analysis details

[0299] Output: Evaluation data stored in a database

[0300] Specific operation: The server records all analysis results in a database and stores the data for future evaluation and improvement.

[0301] 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.

[0302] The present invention relates to a meeting evaluation system for improving the quality of meetings. This system not only records and analyzes meeting audio in real time, scoring the content of participants' remarks, but also monitors harassment and evaluates facilitation skills by recognizing participants' emotions.

[0303] System configuration

[0304] The system consists of the following main components:

[0305] 1. User device (PC, smartphone, tablet, etc.)

[0306] 2. Server (a computer system that receives and analyzes voice data)

[0307] 3. Network (communications infrastructure connecting user devices and servers)

[0308] 4. Emotion engine (a module that recognizes the emotions of meeting participants)

[0309] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The emotion engine also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored over the long term and accumulated as evaluation data.

[0310] Specific operation example

[0311] A specific example of the operation of this system is shown below.

[0312] Meeting Scenarios

[0313] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[0314] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[0315] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[0316] 3. The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine (e.g., voice recognition API).

[0317] 4. The converted text data is input into a generative AI (e.g., a generative AI model) to perform a detailed analysis of the meeting content.

[0318] 5. Based on the analysis results, the server scores participants on their leadership, cooperation, and harassment.

[0319] For example, user A's leadership score is 85 points, user B's contribution score is 70 points, cooperativeness score is 90 points, and no harassment is detected.

[0320] 6. The emotion engine recognizes the participant's emotions from the voice data and sends the data to the server.

[0321] 7. The server adjusts the scoring results based on the recognized emotion data and generates detailed feedback.

[0322] For example, if user A feels stressed during a meeting, feedback is provided based on that emotion.

[0323] 8. The server sends the scoring results and emotion data to the user devices of User A and User B, and visually displays the results through the user interface.

[0324] 9. The device will provide detailed scores, feedback comments from the generative AI, and an evaluation from the emotion engine, allowing users to review their results.

[0325] 10. The server records the scoring results, analysis results, and emotional data in a database and accumulates them as long-term evaluation data.

[0326] Explanation of program processing

[0327] 1. A user starts a meeting and launches the recording app.

[0328] 2. The terminal captures the conference audio and divides it into data packets in real time.

[0329] 3. The terminal sequentially transmits voice data packets to the server.

[0330] 4. The server receives the audio data packets and concatenates them to form an audio stream.

[0331] 5. The server uses a speech recognition engine to convert the speech to text.

[0332] 6. The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[0333] 7. The server sends the analysis results to the terminal and displays them on the user interface.

[0334] 8. The server uses the emotion engine to recognize the emotion data and adjust the scoring results.

[0335] 9. The server sends the adjusted results to the terminal and notifies the user.

[0336] 10. The server records the results in a database and stores them as long-term data.

[0337] Through these processes, the system improves the quality of meetings and provides fair evaluations and appropriate feedback quickly.In addition, the introduction of an emotion engine provides detailed feedback based on the user's emotional state, resulting in a better meeting experience.

[0338] The processing flow will be explained below.

[0339] Step 1:

[0340] A user starts a conference and launches the recording application on their device. When the user presses the "Start Recording" button, the device captures the conference audio through the audio input device (microphone).

[0341] Step 2:

[0342] The terminal divides the recorded conference audio into data packets in real time. The terminal divides the captured audio data into small packets with a certain byte size and transmits them sequentially to the server via the network.

[0343] Step 3:

[0344] The server reorders the received audio data packets and concatenates them into a stream. The server checks the data integrity and requests packet retransmission if necessary.

[0345] Step 4:

[0346] The server inputs the formatted audio stream into a speech recognition engine (e.g., speech recognition API) and converts the audio data into text data. The server receives the converted text data from the speech recognition engine and temporarily stores it in a database.

[0347] Step 5:

[0348] The server inputs the text data into a generative AI (e.g., a generative AI model) and instructs the generative AI to "analyze the content of the meeting based on this text data and score it on indicators such as leadership, cooperation, and the presence or absence of harassment."

[0349] Step 6:

[0350] The server receives the analysis results output by the generation AI and extracts scores and detailed feedback for each indicator. For example, User A's leadership score might be 85 points, User B's contribution score 70 points, cooperation score 90 points, and no harassment detected.

[0351] Step 7:

[0352] The emotion engine recognizes the emotions of conference participants from the voice data. The server provides the voice data to the emotion engine, which analyzes the emotional state. For example, emotions such as anger, anxiety, and joy are recognized.

[0353] Step 8:

[0354] The server receives the recognized emotion data from the emotion engine and adjusts the scoring results, such as affecting leadership and cooperation scores based on the emotion data, or adding warnings if there is a high likelihood of harassment.

[0355] Step 9:

[0356] The server sends the adjusted scoring results and detailed emotional feedback in JSON format to the user's device. The server records the result transmission log and checks the success or failure status.

[0357] Step 10:

[0358] The device receives the JSON data, parses it, and displays it on the user interface, where the user can visually check the score details, feedback from the AI, and the analysis of the user's emotional state.

[0359] Step 11:

[0360] The server structures the scoring results, analysis details, and emotion data and records them in a database. The recorded data is backed up regularly and stored as long-term evaluation data.

[0361] Step 12:

[0362] The server generates reports that analyze long-term evaluation trends and trends based on the accumulated evaluation data and emotion data. The server provides these reports to administrators, helping to improve the quality of corporate meetings.

[0363] This system improves the quality of meetings, providing fair evaluations and prompt, appropriate feedback. Furthermore, the introduction of an emotion engine provides detailed feedback based on the user's emotional state, enabling a better meeting experience.

[0364] Example 2

[0365] 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."

[0366] Accurately evaluating the content and quality of comments made in meetings and providing feedback to meeting participants is important for improving the performance of an entire organization. However, conventional meeting evaluation systems have difficulty evaluating not only the content of participants' comments but also their emotions, and it is not possible to accumulate this data over the long term and perform trend analysis. Therefore, a system that comprehensively evaluates the quality of meetings and provides detailed feedback is needed.

[0367] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for recording the conference audio in real time, a means for dividing the recorded conference audio into data packets and transmitting them to the server, a means for converting the audio data received on the server side into text data, a means for analyzing the text data and scoring the remarks of the conference participants, a means for notifying the scoring results to the user terminal, a means for long-term saving the analysis results and accumulating them as evaluation data, and a means for recognizing the emotions of the conference participants and adjusting the scoring results. This makes it possible to comprehensively evaluate the quality of the conference in real time and provide fair feedback.

[0368] "Conference audio" refers to audio data of conversations and presentations that occur during a conference.

[0369] "Real-time recording means" refers to a device or method that instantly records conference audio and stores it as digital data.

[0370] "Data packet" refers to the unit in which digital data is divided into small pieces for efficient transmission over a network.

[0371] "Means for transmitting to the server" refers to the technology or method for transmitting data packets from the user terminal to the server.

[0372] "Means for converting into text data" refers to speech recognition technology or algorithms that convert voice data into text information.

[0373] "Means for analyzing text data" refers to programs or processing methods for analyzing data converted into text and understanding its content.

[0374] "Means for scoring the content of statements made by meeting participants" refers to algorithms or methods for evaluating and quantifying the content of statements.

[0375] "Means for notifying the user terminal of the scoring results" refers to a method or system for notifying the user of the analysis and evaluation results.

[0376] "Means for storing analysis results for the long term and accumulating them as evaluation data" refers to technology that stores analysis results in a database or the like for a long period of time and uses them for future analysis and evaluation.

[0377] "Means for recognizing the emotions of meeting participants" refers to technologies and methods for identifying the emotional state of meeting participants from voice or text data.

[0378] "Means for adjusting scoring results" refers to algorithms or methods for correcting or amending initial scoring results based on recognized emotion data.

[0379] The present invention relates to a meeting evaluation system for improving the quality of meetings. This system not only records and analyzes meeting audio in real time and scores the content of participants' comments, but also monitors harassment and evaluates facilitation ability by recognizing participants' emotions.

[0380] System configuration

[0381] The system consists of the following main components:

[0382] 1. User device (PC, smartphone, tablet, etc.)

[0383] 2. Server (a computer system that receives and analyzes voice data)

[0384] 3. Network (communications infrastructure connecting user devices and servers)

[0385] 4. Emotion engine (a module that recognizes the emotions of meeting participants)

[0386] 5. Database (a storage device for long-term storage of analysis results)

[0387] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The emotion engine also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored over the long term and accumulated as evaluation data.

[0388] Specific operation example

[0389] A specific example of the operation of this system is shown below.

[0390] Meeting Scenarios

[0391] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[0392] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[0393] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[0394] 3. The server arranges and concatenates the received voice data in order and converts it into text data using a speech recognition engine (e.g., Google Cloud Speech-to-Text API).

[0395] 4. The converted text data is input into a generative AI (e.g., OpenAI GPT-4) to perform a detailed analysis of the meeting content.

[0396] 5. Based on the analysis results, the server scores participants on their leadership, cooperation, and harassment.

[0397] For example, user A's leadership score is 85 points, user B's contribution score is 70 points, cooperativeness score is 90 points, and no harassment is detected.

[0398] 6. The emotion engine recognizes the participant's emotions from the voice data and sends the data to the server.

[0399] 7. The server adjusts the scoring results based on the recognized emotion data and generates detailed feedback.

[0400] For example, if user A feels stressed during a meeting, feedback is provided based on that emotion.

[0401] 8. The server sends the scoring results and emotion data to the user devices of User A and User B, and visually displays the results through the user interface.

[0402] 9. The device will provide detailed scores, feedback comments from the generative AI, and an evaluation from the emotion engine, allowing users to review their results.

[0403] 10. The server records the scoring results, analysis results, and emotional data in a database and accumulates them as long-term evaluation data.

[0404] Prompt Sentence Examples

[0405] Here are some example prompts to input to a generative AI model (e.g., GPT-4):

[0406] The audio from a meeting was transcribed as follows:

[0407] 1. User A: "How is the project going?"

[0408] 2. User B: "It's going well, but there are some challenges."

[0409] 3. User A: "Do you need any help?"

[0410] Please rate User A's leadership, cooperation, and overall performance on a numerical scale. Also, please rate User B's contribution.

[0411] This allows the system to evaluate the quality of the meeting in real time and provide fair feedback. Furthermore, by incorporating an emotion engine, the system can provide detailed feedback based on the user's emotional state, resulting in a better meeting experience.

[0412] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0413] Step 1:

[0414] A user starts a meeting and launches the recording app.

[0415] Specific actions: User A opens the recording app on their smartphone and presses the "Start Recording" button.

[0416] Input: User action at the start of a meeting.

[0417] Output: Start recording.

[0418] Step 2:

[0419] The terminal captures the conference audio and divides it into data packets in real time.

[0420] How it works: The smartphone picks up the audio from the meeting and packets the audio data every second.

[0421] Input: Conference audio.

[0422] Output: Audio data packets.

[0423] Step 3:

[0424] The terminal sequentially transmits voice data packets to the server.

[0425] How it works: A smartphone sends voice data packets over the internet to a cloud server, minimizing latency through parallel processing.

[0426] Input: Audio data packets.

[0427] Output: Sending data packets to the server.

[0428] Step 4:

[0429] The server receives the audio data packets and concatenates them to form an audio stream.

[0430] What it does: The server reorders the audio data packets it receives and reconstructs them into a continuous audio stream.

[0431] Input: Audio data packets.

[0432] Output: Audio stream.

[0433] Step 5:

[0434] The server converts the speech to text using a speech recognition engine (e.g., a speech recognition API).

[0435] What happens: The server inputs the audio stream into the Google Cloud Speech-to-Text API and generates text data.

[0436] Input: Audio stream.

[0437] Output: Text data.

[0438] Step 6:

[0439] The server inputs the converted text data into a generative AI (e.g., a generative AI model) for detailed analysis and scoring.

[0440] Specific operation: The server sends the text data to a generation AI (e.g., GPT-4), which analyzes the content of the meeting and quantifies leadership, cooperation, and the presence or absence of harassment.

[0441] Input: Text data.

[0442] Output: Scoring results.

[0443] Step 7:

[0444] The server sends the analysis results to the terminal and displays them on the user interface.

[0445] Specific operation: The server sends the scoring results generated to the user's smartphone, and the score is displayed on the app.

[0446] Input: Scoring results.

[0447] Output: Notification and display on smartphone.

[0448] Step 8:

[0449] The server uses an emotion engine to recognize the emotion data and adjust the scoring results.

[0450] What it does: The server inputs the voice data into an emotion engine (e.g., Emotion API) to recognize the participant's emotional state, and fine-tunes the scoring results based on that.

[0451] Input: Audio data.

[0452] Output: Recognized emotions, adjusted scoring results.

[0453] Step 9:

[0454] The server sends the adjusted results to the terminal and notifies the user.

[0455] Specific operation: The server sends the adjusted scoring results and emotional data to the smartphone and displays detailed feedback.

[0456] Input: Calibrated scoring results and sentiment data.

[0457] Output: Notifications and detailed feedback displayed on your smartphone.

[0458] Step 10:

[0459] The server records the results in a database for long-term storage.

[0460] Specific operation: The server stores the scoring results and emotion data in a database and accumulates them as future evaluation data.

[0461] Input: Scoring results and sentiment data.

[0462] Output: Recorded and accumulated evaluation data in a database.

[0463] (Application example 2)

[0464] 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."

[0465] Conventional meeting evaluation systems can score the content of participants' comments, but they do not perform more detailed analysis, such as detecting the emotions and harassment contained in the comments, or evaluating cooperation. Furthermore, since there is no feedback that takes into account the emotional elements that arise during meetings, there is a problem in that measures to improve the quality of meetings cannot be taken sufficiently. Furthermore, there is a lack of analysis of evaluation trends through long-term data analysis, so there is a lack of indicators for continuously improving meeting outcomes.

[0466] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for recording conference audio in real time, means for dividing the recorded conference audio into data packets and transmitting them to a computer system, means for converting the audio data received by the computer system into text data, means for analyzing the text data and scoring the speech content of the conference participants, means for notifying the scoring results to an information display terminal, means for long-term saving of the analysis results and accumulating them as evaluation data, and means for recognizing the emotions of the conference participants using an emotion recognition module and adjusting the scoring results based on the emotion data. This enables detailed analysis and feedback to improve the quality of the conference, and enables sustainable measures to improve the outcome of the conference through adjustment of scoring results taking into account the emotional states of the conference participants and accumulation and analysis of long-term evaluation data.

[0467] A "means for recording conference audio in real time" is a device that has the function of instantly recording audio spoken during a conference.

[0468] "Means for dividing the recorded conference audio into data packets and transmitting them to a computer system" refers to a device or software that has the function of dividing the recorded audio data into packets of a certain size and transmitting them to a server via a communication network.

[0469] "Means for converting voice data received by a computer system into text data" refers to the process of converting voice data received by a server into text information using voice recognition technology.

[0470] The "means for analyzing text data and scoring the content of statements made by meeting participants" is a system that has the function of analyzing converted text data and calculating various evaluation indicators based on the content of statements made by meeting participants.

[0471] "Means for notifying the information display terminal of the scoring results" refers to a mechanism for transmitting the scoring information obtained as an analysis result to the device used by the user and displaying or notifying the user.

[0472] "Means for storing analysis results over the long term and accumulating them as evaluation data" refers to a system that stores the generated analysis results and scores in a database and maintains that data for a long period of time.

[0473] "Means for recognizing the emotions of meeting participants using an emotion recognition module and adjusting the scoring results based on that emotion data" is a component that has the function of recognizing emotions from the voices and facial expressions of meeting participants, and appropriately correcting the scoring results to reflect their emotional state.

[0474] The "means for scoring leadership, cooperation, and the presence or absence of harassment" is a system that has the function of evaluating the strength of leadership, the degree of cooperation, and whether harassment is present based on the comments and actions of meeting participants.

[0475] The "means for providing detailed feedback" is a mechanism for providing specific and useful feedback to the user based on the scoring results and emotional data.

[0476] "Means of analyzing long-term evaluation tendencies and trends based on accumulated data" refers to the process of analyzing stored historical data to identify evaluation tendencies and trends that emerge over time.

[0477] "Means that take into account the correlation with emotional data" refers to a method of analyzing emotional data in combination with other evaluation indicators to verify the causal relationship and influence between emotions and behavior.

[0478] The present invention relates to a conference evaluation system for improving the quality of conferences, and includes means for recording conference audio in real time, means for dividing the recorded audio into data packets and transmitting them to a computer system, means for converting the audio data received by the computer system into text data, means for analyzing the text data and scoring the content of comments made by conference participants, means for notifying an information display terminal of the scoring results, means for long-term storage of the analysis results and accumulation as evaluation data, and means for recognizing the emotions of conference participants using an emotion recognition module and adjusting the scoring results based on that emotion data. This system makes it possible to improve the quality of conferences.

[0479] System configuration

[0480] The system consists of the following main components:

[0481] 1. User device (PC, smartphone, tablet, etc.)

[0482] 2. Server (computer system that receives and analyzes voice data)

[0483] 3. Network (communications infrastructure connecting user devices and servers)

[0484] 4. Emotion Recognition Module (Module that recognizes the emotions of meeting participants)

[0485] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and analyzed using generative AI. An emotion recognition module also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and various indicators (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) are evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored long-term and accumulated as evaluation data.

[0486] System Operation

[0487] A specific example of the operation of this system is shown below.

[0488] Meeting Scenarios

[0489] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[0490] 1. User A starts the recording application on their device and records the conference audio in real time.

[0491] 2. The user terminal divides the recorded voice into data packets and sends them to the server over the network.

[0492] 3. The server uses a speech recognition engine (e.g., Google Speech-to-Text API) to convert the received voice data into text data.

[0493] 4. The converted text data is analyzed by a generative AI (e.g., a generative AI model such as GPT-4).

[0494] 5. Based on the analysis results, the server scores the comments made by the conference participants and notifies the user terminal of the scoring results.

[0495] 6. The emotion recognition module recognizes the emotions of meeting participants from the voice data and adjusts the scoring results based on this emotion data.

[0496] 7. The server notifies the user device of the adjusted scoring results and detailed feedback.

[0497] 8. Analysis results and emotional data will be stored for a long period of time and accumulated on the server as evaluation data.

[0498] 9. The accumulated data will be analyzed to help improve the quality of future meetings, and long-term evaluation and trend analysis will be conducted.

[0499] Specific examples of hardware and software used

[0500] User devices: PC, smartphone, tablet

[0501] Server: High-performance cloud server

[0502] Network: Internet, LAN

[0503] Speech recognition engine: Google Speech-to-Text API

[0504] Generative AI models: GPT-4 and similar models

[0505] Emotion Recognition Module: Dedicated emotion analysis software

[0506] Prompt Sentence Examples

[0507] "Please analyze this text and rate it on points like leadership, collaboration, and customer service. I'd also like you to do a sentiment analysis."

[0508] This will improve the quality of meetings and enable fair evaluations and prompt provision of appropriate feedback. Furthermore, the introduction of an emotion recognition module will provide detailed feedback based on the user's emotional state, enabling a better meeting experience.

[0509] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0510] Step 1:

[0511] A user starts a meeting and launches a recording application on their device. Recording of the meeting audio begins. The input is the user's voice, and the output is the recorded audio data. Specifically, the microphone on the user's device picks up the audio, and the recording software records it.

[0512] Step 2:

[0513] The device divides the recorded voice into data packets in real time and sends them to the server via the network. The input is the recorded voice data, and the output is the voice data divided into data packets. Specifically, the recording application on the user device divides the recorded data into small data packets and sends them to the server via the Internet.

[0514] Step 3:

[0515] The server concatenates the received audio data to form a single audio data stream. The input is audio data divided into data packets, and the output is the concatenated audio data. Specifically, the server program reassembles the arriving data packets into the correct order and saves them as a stream of audio data.

[0516] Step 4:

[0517] The server converts the concatenated voice data into text data using a speech recognition engine (e.g., Google Speech-to-Text API). The input is the concatenated voice data, and the output is text data. Specifically, the speech recognition engine analyzes the voice samples and generates corresponding text.

[0518] Step 5:

[0519] The server inputs the text data into a generative AI model (e.g., GPT-4) for detailed analysis. The input is text data, and the output is analysis results and scoring data. An example prompt is, "Analyze this text and evaluate points such as leadership, cooperation, and customer responsiveness. Please also perform sentiment analysis." Specifically, the generative AI model analyzes the text data and generates a score based on each indicator.

[0520] Step 6:

[0521] The emotion recognition module recognizes the emotions of conference participants from the voice data and generates emotion data. The input is voice data and the output is emotion data. Specifically, the emotion recognition software analyzes the tone and rate of the voice to determine emotional states such as stress, joy, or anger.

[0522] Step 7:

[0523] The server adjusts the initial scoring result based on the emotion data. The input is the initial scoring result and emotion data, and the output is the adjusted scoring result. Specifically, the scoring algorithm incorporates the emotion data and recalculates the scoring result.

[0524] Step 8:

[0525] The server sends the adjusted scoring results and detailed feedback to the user terminal. The input is the adjusted scoring results and feedback data, and the output is a notification to the user terminal. Specifically, the server sends the generated detailed feedback comments to the user terminal and displays them through the application interface.

[0526] Step 9:

[0527] The analysis results and emotion data are stored over the long term and accumulated on the server as evaluation data. The input is the analysis results and emotion data, and the output is data stored in the database. Specifically, the server's database management system efficiently records this data and makes it available for future trend analysis and evaluation.

[0528] Step 10:

[0529] The server analyzes long-term evaluation trends and trends based on the accumulated data and generates reports. The input is the long-term stored evaluation data, and the output is the evaluation report. Specifically, data analysis software analyzes past data, extracts notable trends and patterns, and generates a report to be provided to the user.

[0530] 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.

[0531] 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.

[0532] 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.

[0533] [Second embodiment]

[0534] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0535] 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.

[0536] 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).

[0537] 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.

[0538] 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.

[0539] 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).

[0540] 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.

[0541] 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.

[0542] 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.

[0543] 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.

[0544] 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.

[0545] 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."

[0546] This invention relates to a meeting evaluation system for improving the quality of meetings. This system records and analyzes meeting audio in real time and scores the remarks of meeting participants to monitor harassment and evaluate facilitation skills.

[0547] System configuration

[0548] The system consists of the following main components:

[0549] 1. User device (PC, smartphone, tablet, etc.)

[0550] 2. Server (a computer system that receives and analyzes voice data)

[0551] 3. Network (communications infrastructure connecting user devices and servers)

[0552] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided.

[0553] Specific operation example

[0554] A specific example of the operation of this system is shown below.

[0555] Meeting Scenarios

[0556] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[0557] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[0558] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[0559] 3. The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine (e.g., voice recognition API).

[0560] 4. The converted text data is input into a generative AI (e.g., a generative AI model) to perform a detailed analysis of the meeting content.

[0561] 5. Based on the analysis results, the server scores participants on their leadership, cooperativeness, and whether or not they engage in harassment.

[0562] For example, user A's leadership score is 85, user B's contribution score is 70, cooperativeness score is 90, and no harassment is detected.

[0563] 6. The server sends the scoring results to the user devices of User A and User B and displays the results visually through the user interface.

[0564] 7. The device will provide detailed score information and feedback comments from the generated AI, allowing users to review their results.

[0565] 8. The server records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[0566] Explanation of program processing

[0567] 1. A user starts a meeting and launches the recording app.

[0568] 2. The terminal captures the conference audio and divides it into data packets in real time.

[0569] 3. The terminal sequentially transmits voice data packets to the server.

[0570] 4. The server receives the audio data packets and concatenates them to form an audio stream.

[0571] 5. The server uses a speech recognition engine to convert the speech to text.

[0572] 6. The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[0573] 7. The server sends the analysis results to the terminal and displays them on the user interface.

[0574] 8. The server records the results in a database and stores them as long-term data.

[0575] Through these processes, the present invention can improve the quality of meetings and achieve fair evaluations. In addition, the harassment monitoring function can effectively manage risk within a company. In this way, the embodiments for carrying out the invention have been specifically described.

[0576] The processing flow will be explained below.

[0577] Step 1:

[0578] A user starts a conference and launches the recording application on their device. When the user presses the "Start Recording" button, the device captures the conference audio through the audio input device (microphone).

[0579] Step 2:

[0580] The terminal divides the recorded conference audio into data packets in real time. The terminal divides the captured audio data into small packets with a certain byte size and transmits them sequentially to the server via the network.

[0581] Step 3:

[0582] The server rearranges the received audio data packets in order and concatenates them in a stream format. The server checks the data integrity and requests packet retransmission if necessary.

[0583] Step 4:

[0584] The server inputs the formatted audio stream into a speech recognition engine (e.g., speech recognition API) and converts the audio data into text data. The server receives the converted text data from the speech recognition engine and temporarily stores it in a database.

[0585] Step 5:

[0586] The server inputs the text data into a generative AI (e.g., a generative AI model) and instructs the generative AI to "analyze the content of the meeting based on this text data and score it on indicators such as leadership, cooperation, and the presence or absence of harassment."

[0587] Step 6:

[0588] The server receives the analysis results output by the generation AI and extracts scores and detailed feedback for each indicator. For example, User A's leadership score might be 85 points, User B's contribution score 70 points, cooperation score 90 points, and no harassment detected.

[0589] Step 7:

[0590] The server sends the scoring results and detailed feedback comments in JSON format to the user's device. The server records the result transmission log and checks the success or failure status.

[0591] Step 8:

[0592] The device receives the JSON data, parses it, and displays it in the user interface, where the user can visually check the score details and feedback from the generating AI.

[0593] Step 9:

[0594] The server structures the scoring results and analysis and records them in a database. The recorded data is backed up regularly and stored as long-term evaluation data.

[0595] Step 10:

[0596] The server generates reports based on the accumulated evaluation data, analyzing long-term evaluation trends and other trends, and provides these reports to administrators to help improve the quality of corporate meetings.

[0597] This enables the meeting evaluation system to improve the quality of meetings and provide fair evaluations and appropriate feedback quickly.

[0598] Example 1

[0599] 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."

[0600] Conventional meeting evaluation systems often manually record and analyze meeting audio, making it difficult to effectively evaluate meeting quality. They also lacked a means to objectively evaluate participants' leadership and cooperation, or to automatically detect harassment. As a result, it was difficult to provide feedback to improve meeting quality, and risk management across the company was inadequate.

[0601] 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.

[0602] In this invention, the server includes a means for dividing the recorded conference audio into data packets and transmitting them to the server, a means for converting the received audio data into text data, a means for inputting the converted text data into a generative AI model for analysis and scoring the speech content of the conference participants, a means for notifying the user terminal of the scoring results, and a means for long-term storage of the analysis results and accumulating them as evaluation data. This enables effective evaluation of the quality of the conference and provides detailed feedback to participants. It also automates the evaluation of leadership and cooperation and the monitoring of harassment, strengthening risk management in companies.

[0603] "Conference audio" refers to all audio data spoken during a conference.

[0604] "Recording means" refers to the equipment or method for recording conference audio in real time and converting it into digital data.

[0605] A "data packet" refers to a data unit that divides a series of audio data into a certain size and transmits it over a network.

[0606] "Server" refers to a computing system that processes data received from a terminal, generates analysis results, and manages them.

[0607] "Audio data" refers to data in which conference audio has been converted into digital format.

[0608] "Text data" refers to character string data converted from voice data using voice recognition technology.

[0609] A "generative AI model" refers to an artificial intelligence model that performs natural language processing and analysis based on large amounts of data.

[0610] "Analyzing" refers to the process of using certain algorithms or models to make sense of input data and extract relevant information.

[0611] "Content of remarks" refers to the general content of remarks made by conference participants during the conference.

[0612] "Scoring" refers to assigning points based on analysis results according to certain evaluation criteria.

[0613] "User terminal" refers to a device used by a conference participant to record voice and display the analysis results.

[0614] "Means for notifying" refers to a method or device for transmitting information from a server to a user terminal and informing the user.

[0615] "Storage means" refers to methods and devices for retaining data for long periods of time.

[0616] "Evaluation data" refers to data including analysis results and scoring results of meetings.

[0617] "Detailed feedback" refers to specific and analytical evaluations and information on areas for improvement provided to meeting participants.

[0618] "Leadership" refers to the criteria used to evaluate meeting participants' ability to lead and provide direction.

[0619] "Collaborative ability" refers to a criterion that evaluates meeting participants' ability to cooperate and collaborate as team members.

[0620] "Harassment" refers to inappropriate comments or behavior that may occur during a meeting.

[0621] This invention relates to a meeting evaluation system for improving the quality of meetings. This system records and analyzes meeting audio in real time and scores the remarks made by meeting participants to monitor harassment and evaluate facilitation skills. The detailed configuration and operation of the system are shown below.

[0622] System configuration

[0623] The system consists of the following main components:

[0624] 1. User device (e.g. PC, smartphone, tablet, etc.)

[0625] 2. Server (a computer system that receives and analyzes voice data)

[0626] 3. Network (communications infrastructure connecting user devices and servers)

[0627] Operational Overview

[0628] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted to text data on the server side and further analyzed using a generative AI model. The analysis results are scored based on the content of the speech, and various indicators (such as leadership, cooperation, and the presence or absence of harassment) are evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided.

[0629] Specific operation example

[0630] A specific example of the operation of this system is shown below.

[0631] Meeting Scenarios

[0632] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[0633] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[0634] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[0635] 3. The server arranges and concatenates the received voice data in order and converts it into text data using a voice recognition engine (e.g., Google Cloud Speech-to-Text).

[0636] 4. The converted text data is input into a generative AI (e.g., GPT-4 model) to perform a detailed analysis of the meeting content.

[0637] 5. Based on the analysis results, the server scores participants on their leadership, cooperativeness, and whether or not they engage in harassment.

[0638] For example, user A's leadership score is 85, user B's contribution score is 70, cooperativeness score is 90, and no harassment is detected.

[0639] 6. The server sends the scoring results to the user's device and displays the results visually through the user interface.

[0640] 7. The device will provide detailed score information and feedback comments from the generated AI, allowing users to review their results.

[0641] 8. The server records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[0642] Example of input prompt for generative AI model

[0643] Example prompt:

[0644] Analyze the audio recording of the meeting below and rate each participant's leadership, cooperation, and whether or not they engaged in harassment.

[0645] text:

[0646] User A: Today's meeting is about the progress of the new project. Everyone, please give us your opinions.

[0647] User B: I'll be in charge of the progress report. Currently, Task A is 80% complete.

[0648] ...

[0649] In this way, by showing detailed embodiments of the invention, it is possible to effectively evaluate the quality of meetings and provide clear feedback to participants through scoring of leadership and cooperation, which is expected to improve meeting management and strengthen corporate risk management.

[0650] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0651] Step 1:

[0652] A user starts a meeting and launches the recording app.

[0653] Specifically, when a meeting starts, the user launches a dedicated recording application on their device.

[0654] Specific operation: For example, User A double-clicks the meeting recording app on his / her PC and clicks the "Start Recording" button.

[0655] Input: Conference audio (analog audio)

[0656] Output: Launched recording application

[0657] Step 2:

[0658] The terminal captures the conference audio and divides it into data packets in real time.

[0659] Specific description: The terminal records the conference audio through a microphone and performs the process of dividing it into small data packets.

[0660] What it does: The PC's microphone captures audio data every second, and then divides each into data packets every 0.5 seconds.

[0661] Input: Conference audio (analog audio)

[0662] Output: Data packet (digital data)

[0663] Step 3:

[0664] The terminal sequentially transmits voice data packets to the server.

[0665] Specific explanation: The terminal sequentially transmits the divided voice data packets to the server.

[0666] What it does: Your PC sends packetized voice data to a server every second via your internet connection.

[0667] Input: Data packets (digital voice data)

[0668] Output: Data packet sent to the server

[0669] Step 4:

[0670] The server receives the audio data packets and concatenates them to form an audio stream.

[0671] Specifically, the server concatenates the received audio data packets in order to create a continuous audio stream.

[0672] Specific operation: The server combines the audio data packets received every 0.5 seconds to restore the original continuous audio data.

[0673] Input: Data packets (digital voice data)

[0674] Output: Concatenated audio stream (continuous audio data)

[0675] Step 5:

[0676] The server uses a speech recognition engine to convert the speech into text.

[0677] Specifically, the server uses a speech recognition engine (e.g., Google Cloud Speech-to-Text) to convert the voice data into text data.

[0678] What happens: The server sends the audio file to a speech recognition API and retrieves the conversation in text format.

[0679] Input: Concatenated audio stream (continuous audio data)

[0680] Output: Text data (string format)

[0681] Step 6:

[0682] The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[0683] Specific explanation: The server inputs text data into a generative AI (e.g., GPT-4 model) to analyze the meeting content and score what is said.

[0684] Specific behavior: The server passes the following prompt to the generation AI:

[0685] Example prompt:

[0686] Analyze the audio recording of the meeting below and rate each participant's leadership, cooperation, and whether or not they engaged in harassment.

[0687] text:

[0688] User A: Today's meeting is about the progress of the new project. Everyone, please give us your opinions.

[0689] User B: I'll be in charge of the progress report. Currently, Task A is 80% complete.

[0690] ...

[0691] Input: Text data (string format)

[0692] Output: Analysis results and scoring data (evaluation index)

[0693] Step 7:

[0694] The server sends the analysis results to the terminal and displays them on the user interface.

[0695] Specific explanation: The server sends the analysis results returned by the generation AI to the user terminal and displays them on the user interface.

[0696] Specific operation: The server sends the evaluation results (e.g., User A's leadership score is 85 points, User B's contribution score is 70 points) to the devices of User A and User B, and visualizes them on a dedicated dashboard.

[0697] Input: Analysis results and scoring data (evaluation index)

[0698] Output: Evaluation results displayed on the user's terminal

[0699] Step 8:

[0700] The server records the results in a database and accumulates them as long-term evaluation data.

[0701] Specific explanation: The server stores the analysis results and each meeting data in a database and manages them as long-term evaluation data.

[0702] Specific operation: The server records the evaluation results and speech analysis data in a MySQL database for future comparative analysis.

[0703] Input: Analysis results and scoring data (evaluation index)

[0704] Output: Evaluation data stored in a database

[0705] This system improves the quality of meetings and enables fair evaluations. It supports productive meeting management while also providing a harassment monitoring function, strengthening risk management across the company.

[0706] (Application example 1)

[0707] 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."

[0708] Conventional meeting evaluation systems can evaluate the quality of meetings and the behavior of participants in real time, but they are unable to evaluate in-car communication or safety awareness. Furthermore, there is a lack of systems that can analyze in-car conversations in detail to evaluate the driver's attention and whether or not harassment has occurred. This has resulted in insufficient prevention of harassment and improvement of safety awareness in the in-car environment.

[0709] 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.

[0710] In this invention, the server includes means for recording conference audio in real time, means for dividing the recorded conference audio into data packets and transmitting them to the server, means for converting the received audio data into text data, means for analyzing the text data and scoring the speech content of the conference participants, means for recording in-car conversations in real time, means for scoring the driver's attention, safety awareness, and presence or absence of harassment through analysis, means for notifying the scoring results to a user terminal, and means for long-term storage of the analysis results and accumulating them as evaluation data. This makes it possible to evaluate in-car communication quality, safety awareness, and presence or absence of harassment in detail. This also makes it possible to appropriately monitor the behavior of drivers and passengers and improve the safety and comfort of the in-car environment.

[0711] "Conference audio" refers to the audio data of the participants' comments and interactions that occur during a conference.

[0712] A "data packet" is a small unit of digital data divided into pieces and formatted for efficient transmission over a network.

[0713] A "server" is a central computer system for collecting, analyzing, and storing data.

[0714] "Text data" refers to text information converted from speech using speech recognition technology.

[0715] "Scoring" is an evaluation method in which points are assigned according to specific evaluation indicators based on analyzed data.

[0716] "User terminal" refers to a device that is directly operated by a user, such as a PC, smartphone, or tablet.

[0717] "Real-time" refers to data being processed and transferred immediately with little delay.

[0718] "Attention" is an indicator of how much a driver is concentrating on traffic conditions and driving operations.

[0719] "Harassment" refers to inappropriate words, actions or behaviour towards other crew members.

[0720] "Analysis" is the process of analyzing collected data in detail to clarify its meaning and trends.

[0721] "Notification" refers to the act of informing the user of the analysis results or scoring results.

[0722] "Evaluation data" refers to data that includes analysis results and scoring results, and is used as the basis for long-term evaluation and improvement.

[0723] This invention relates to an in-car communication evaluation system. The purpose is to improve the safety and comfort of the in-car environment by recording and analyzing in-car conversations in real time and scoring passenger attention, safety awareness, and the presence or absence of harassment. This system consists of the following main components:

[0724] System configuration

[0725] 1. User Device

[0726] This refers to in-vehicle systems, smartphones, tablets, etc. that have recording functions and are responsible for recording conversations in real time.

[0727] 2. Server

[0728] It is the central computer system that collects and analyzes data and stores the results.

[0729] 3. Network

[0730] It is a communications infrastructure that connects user terminals and servers, and uses wireless communication technology.

[0731] Specific operation example

[0732] A specific example of the operation of this system is shown below.

[0733] In-car communication scenario

[0734] Consider a situation in which user A (driver) and user B (passenger) are having a conversation in a car.

[0735] 1. User terminal: The vehicle's in-vehicle system activates the recording function and presses the "Start Recording" button.

[0736] 2. Terminal: Records conversations inside the car in real time, divides the voice data into data packets, and sends them to the server sequentially.

[0737] 3. Server: The received voice data is arranged in order and concatenated, and the voice data is converted into text data using a voice recognition engine (e.g., speech_recognition library).

[0738] 4. Server: The converted text data is input into a generative AI model (e.g., GPT-3) for detailed analysis.

[0739] 5. Server: Based on the analysis results, the server scores participants on their attention, safety awareness, and whether or not they have engaged in harassment.

[0740] For example, user A's attention score is 85, user B's safety awareness score is 90, and no harassment is detected.

[0741] 6. Server: Sends the scoring results to the user's terminal and displays the results visually through the user interface.

[0742] 7. Device: Provides detailed score information and feedback comments from the generated AI, allowing users to check their results.

[0743] 8. Server: Records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[0744] Hardware and software used

[0745] Speech recognition engine: speech_recognition library

[0746] Generative AI model: GPT-3

[0747] User devices: in-vehicle systems, smartphones, tablets, etc.

[0748] Server: a computer system for data analysis and storage

[0749] Network: Wireless communication technology (e.g., Wi-Fi, 4G, 5G)

[0750] Adding specific examples

[0751] Driving conversation scenario:

[0752] Expect conversations such as, "Are you concentrating on driving?", "Drive carefully," and "We're not in a hurry, so drive safely."

[0753] Example prompts to input to a generative AI model:

[0754] Based on the conversation below, please rate the level of attention, safety awareness, and whether or not there was any harassment.

[0755] example:

[0756] How was your drive today?

[0757] Drive carefully. We're not in a hurry, so safety comes first.

[0758] I think you are being harsh towards me.

[0759] The present invention thereby provides a means for evaluating in-vehicle communications in real time and for appropriately monitoring occupant behavior.

[0760] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0761] Step 1:

[0762] The user terminal activates the recording function of the vehicle's in-vehicle system and operates the "start recording" button.

[0763] Input: Start recording

[0764] Output: Real-time recording begins

[0765] Specific operation: The user starts recording on the in-car system or smartphone. The device prepares to record the conversation in real time.

[0766] Step 2:

[0767] The terminal records conversations in the car in real time, divides the voice data into data packets, and transmits them to the server one by one.

[0768] Input: Voice data of conversations in a car

[0769] Output: sent as data packets to the server

[0770] Specific operation: The device divides the recorded audio into small data packets and sends them to the server over the network.

[0771] Step 3:

[0772] The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine.

[0773] Input: Packetized voice data

[0774] Output: Converted text data

[0775] Specific operation: The server uses a speech recognition library (e.g., speech_recognition) to analyze the voice data and convert it into text data.

[0776] Step 4:

[0777] The server inputs the converted text data into a generative AI model (e.g., GPT-3) for detailed analysis.

[0778] Input: Text data

[0779] Output: Analysis results (e.g., scores for attention, safety awareness, and whether or not there was harassment)

[0780] Specific operation: The server inputs prompts into the generative AI model, which then performs detailed analysis and scoring based on the conversation content.

[0781] Step 5:

[0782] Based on the analysis results, the server scores the person's attention, safety awareness, and whether or not they have engaged in harassment.

[0783] Input: Analysis results from a generative AI model

[0784] Output: Each score (e.g., attention 85 points, safety awareness 90 points, no harassment detected)

[0785] Specific operation: The server quantifies specific indicators as scores based on the output of the generation AI.

[0786] Step 6:

[0787] The server transmits the scoring results to the user terminal and visually displays the results through a user interface.

[0788] Input: Scoring results

[0789] Output: Visual feedback displayed in the user interface

[0790] Specific operation: The server sends the score and analysis results to the user's device, which then displays the results on the screen.

[0791] Step 7:

[0792] The device provides the user with detailed score information and feedback comments from the generated AI for review.

[0793] Input: Scoring results received from the server

[0794] Output: Feedback comments that the user sees

[0795] Specific operation: The terminal displays the scoring results and feedback comments for the user to review.

[0796] Step 8:

[0797] The server records the scoring results and analysis details in a database and accumulates them as long-term evaluation data.

[0798] Input: Scoring results and analysis details

[0799] Output: Evaluation data stored in a database

[0800] Specific operation: The server records all analysis results in a database and stores the data for future evaluation and improvement.

[0801] 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.

[0802] The present invention relates to a meeting evaluation system for improving the quality of meetings. This system not only records and analyzes meeting audio in real time, scoring the content of participants' remarks, but also monitors harassment and evaluates facilitation skills by recognizing participants' emotions.

[0803] System configuration

[0804] The system consists of the following main components:

[0805] 1. User device (PC, smartphone, tablet, etc.)

[0806] 2. Server (a computer system that receives and analyzes voice data)

[0807] 3. Network (communications infrastructure connecting user devices and servers)

[0808] 4. Emotion engine (a module that recognizes the emotions of meeting participants)

[0809] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The emotion engine also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored over the long term and accumulated as evaluation data.

[0810] Specific operation example

[0811] A specific example of the operation of this system is shown below.

[0812] Meeting Scenarios

[0813] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[0814] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[0815] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[0816] 3. The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine (e.g., voice recognition API).

[0817] 4. The converted text data is input into a generative AI (e.g., a generative AI model) to perform a detailed analysis of the meeting content.

[0818] 5. Based on the analysis results, the server scores participants on their leadership, cooperation, and harassment.

[0819] For example, user A's leadership score is 85 points, user B's contribution score is 70 points, cooperativeness score is 90 points, and no harassment is detected.

[0820] 6. The emotion engine recognizes the participant's emotions from the voice data and sends the data to the server.

[0821] 7. The server adjusts the scoring results based on the recognized emotion data and generates detailed feedback.

[0822] For example, if user A feels stressed during a meeting, feedback is provided based on that emotion.

[0823] 8. The server sends the scoring results and emotion data to the user devices of User A and User B, and visually displays the results through the user interface.

[0824] 9. The device will provide detailed scores, feedback comments from the generative AI, and an evaluation from the emotion engine, allowing users to review their results.

[0825] 10. The server records the scoring results, analysis results, and emotional data in a database and accumulates them as long-term evaluation data.

[0826] Explanation of program processing

[0827] 1. A user starts a meeting and launches the recording app.

[0828] 2. The terminal captures the conference audio and divides it into data packets in real time.

[0829] 3. The terminal sequentially transmits voice data packets to the server.

[0830] 4. The server receives the audio data packets and concatenates them to form an audio stream.

[0831] 5. The server uses a speech recognition engine to convert the speech to text.

[0832] 6. The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[0833] 7. The server sends the analysis results to the terminal and displays them on the user interface.

[0834] 8. The server uses the emotion engine to recognize the emotion data and adjust the scoring results.

[0835] 9. The server sends the adjusted results to the terminal and notifies the user.

[0836] 10. The server records the results in a database and stores them as long-term data.

[0837] Through these processes, the system improves the quality of meetings and provides fair evaluations and appropriate feedback quickly.In addition, the introduction of an emotion engine provides detailed feedback based on the user's emotional state, resulting in a better meeting experience.

[0838] The processing flow will be explained below.

[0839] Step 1:

[0840] A user starts a conference and launches the recording application on their device. When the user presses the "Start Recording" button, the device captures the conference audio through the audio input device (microphone).

[0841] Step 2:

[0842] The terminal divides the recorded conference audio into data packets in real time. The terminal divides the captured audio data into small packets with a certain byte size and transmits them sequentially to the server via the network.

[0843] Step 3:

[0844] The server reorders the received audio data packets and concatenates them into a stream. The server checks the data integrity and requests packet retransmission if necessary.

[0845] Step 4:

[0846] The server inputs the formatted audio stream into a speech recognition engine (e.g., speech recognition API) and converts the audio data into text data. The server receives the converted text data from the speech recognition engine and temporarily stores it in a database.

[0847] Step 5:

[0848] The server inputs the text data into a generative AI (e.g., a generative AI model) and instructs the generative AI to "analyze the content of the meeting based on this text data and score it on indicators such as leadership, cooperation, and the presence or absence of harassment."

[0849] Step 6:

[0850] The server receives the analysis results output by the generation AI and extracts scores and detailed feedback for each indicator. For example, User A's leadership score might be 85 points, User B's contribution score 70 points, cooperation score 90 points, and no harassment detected.

[0851] Step 7:

[0852] The emotion engine recognizes the emotions of conference participants from the voice data. The server provides the voice data to the emotion engine, which analyzes the emotional state. For example, emotions such as anger, anxiety, and joy are recognized.

[0853] Step 8:

[0854] The server receives the recognized emotion data from the emotion engine and adjusts the scoring results, such as affecting leadership and cooperation scores based on the emotion data, or adding warnings if there is a high likelihood of harassment.

[0855] Step 9:

[0856] The server sends the adjusted scoring results and detailed emotional feedback in JSON format to the user's device. The server records the result transmission log and checks the success or failure status.

[0857] Step 10:

[0858] The device receives the JSON data, parses it, and displays it on the user interface, where the user can visually check the score details, feedback from the AI, and the analysis of the user's emotional state.

[0859] Step 11:

[0860] The server structures the scoring results, analysis details, and emotion data and records them in a database. The recorded data is backed up regularly and stored as long-term evaluation data.

[0861] Step 12:

[0862] The server generates reports that analyze long-term evaluation trends and trends based on the accumulated evaluation data and emotion data. The server provides these reports to administrators, helping to improve the quality of corporate meetings.

[0863] This system improves the quality of meetings, providing fair evaluations and prompt, appropriate feedback. Furthermore, the introduction of an emotion engine provides detailed feedback based on the user's emotional state, enabling a better meeting experience.

[0864] Example 2

[0865] 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."

[0866] Accurately evaluating the content and quality of comments made in meetings and providing feedback to meeting participants is important for improving the performance of an entire organization. However, conventional meeting evaluation systems have difficulty evaluating not only the content of participants' comments but also their emotions, and it is not possible to accumulate this data over the long term and perform trend analysis. Therefore, a system that comprehensively evaluates the quality of meetings and provides detailed feedback is needed.

[0867] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for recording the conference audio in real time, a means for dividing the recorded conference audio into data packets and transmitting them to the server, a means for converting the audio data received on the server side into text data, a means for analyzing the text data and scoring the remarks of the conference participants, a means for notifying the scoring results to the user terminal, a means for long-term saving the analysis results and accumulating them as evaluation data, and a means for recognizing the emotions of the conference participants and adjusting the scoring results. This makes it possible to comprehensively evaluate the quality of the conference in real time and provide fair feedback.

[0868] "Conference audio" refers to audio data of conversations and presentations that occur during a conference.

[0869] "Real-time recording means" refers to a device or method that instantly records conference audio and stores it as digital data.

[0870] "Data packet" refers to the unit in which digital data is divided into small pieces for efficient transmission over a network.

[0871] "Means for transmitting to the server" refers to the technology or method for transmitting data packets from the user terminal to the server.

[0872] "Means for converting into text data" refers to speech recognition technology or algorithms that convert voice data into text information.

[0873] "Means for analyzing text data" refers to programs or processing methods for analyzing data converted into text and understanding its content.

[0874] "Means for scoring the content of statements made by meeting participants" refers to algorithms or methods for evaluating and quantifying the content of statements.

[0875] "Means for notifying the user terminal of the scoring results" refers to a method or system for notifying the user of the analysis and evaluation results.

[0876] "Means for storing analysis results for the long term and accumulating them as evaluation data" refers to technology that stores analysis results in a database or the like for a long period of time and uses them for future analysis and evaluation.

[0877] "Means for recognizing the emotions of meeting participants" refers to technologies and methods for identifying the emotional state of meeting participants from voice or text data.

[0878] "Means for adjusting scoring results" refers to algorithms or methods for correcting or amending initial scoring results based on recognized emotion data.

[0879] The present invention relates to a meeting evaluation system for improving the quality of meetings. This system not only records and analyzes meeting audio in real time and scores the content of participants' comments, but also monitors harassment and evaluates facilitation ability by recognizing participants' emotions.

[0880] System configuration

[0881] The system consists of the following main components:

[0882] 1. User device (PC, smartphone, tablet, etc.)

[0883] 2. Server (a computer system that receives and analyzes voice data)

[0884] 3. Network (communications infrastructure connecting user devices and servers)

[0885] 4. Emotion engine (a module that recognizes the emotions of meeting participants)

[0886] 5. Database (a storage device for long-term storage of analysis results)

[0887] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The emotion engine also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored over the long term and accumulated as evaluation data.

[0888] Specific operation example

[0889] A specific example of the operation of this system is shown below.

[0890] Meeting Scenarios

[0891] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[0892] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[0893] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[0894] 3. The server arranges and concatenates the received voice data in order and converts it into text data using a speech recognition engine (e.g., Google Cloud Speech-to-Text API).

[0895] 4. The converted text data is input into a generative AI (e.g., OpenAI GPT-4) to perform a detailed analysis of the meeting content.

[0896] 5. Based on the analysis results, the server scores participants on their leadership, cooperation, and harassment.

[0897] For example, user A's leadership score is 85 points, user B's contribution score is 70 points, cooperativeness score is 90 points, and no harassment is detected.

[0898] 6. The emotion engine recognizes the participant's emotions from the voice data and sends the data to the server.

[0899] 7. The server adjusts the scoring results based on the recognized emotion data and generates detailed feedback.

[0900] For example, if user A feels stressed during a meeting, feedback is provided based on that emotion.

[0901] 8. The server sends the scoring results and emotion data to the user devices of User A and User B, and visually displays the results through the user interface.

[0902] 9. The device will provide detailed scores, feedback comments from the generative AI, and an evaluation from the emotion engine, allowing users to review their results.

[0903] 10. The server records the scoring results, analysis results, and emotional data in a database and accumulates them as long-term evaluation data.

[0904] Prompt Sentence Examples

[0905] Here are some example prompts to input to a generative AI model (e.g., GPT-4):

[0906] The audio from a meeting was transcribed as follows:

[0907] 1. User A: "How is the project going?"

[0908] 2. User B: "It's going well, but there are some challenges."

[0909] 3. User A: "Do you need any help?"

[0910] Please rate User A's leadership, cooperation, and overall performance on a numerical scale. Also, please rate User B's contribution.

[0911] This allows the system to evaluate the quality of the meeting in real time and provide fair feedback. Furthermore, by incorporating an emotion engine, the system can provide detailed feedback based on the user's emotional state, resulting in a better meeting experience.

[0912] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0913] Step 1:

[0914] A user starts a meeting and launches the recording app.

[0915] Specific actions: User A opens the recording app on their smartphone and presses the "Start Recording" button.

[0916] Input: User action at the start of a meeting.

[0917] Output: Start recording.

[0918] Step 2:

[0919] The terminal captures the conference audio and divides it into data packets in real time.

[0920] How it works: The smartphone picks up the audio from the meeting and packets the audio data every second.

[0921] Input: Conference audio.

[0922] Output: Audio data packets.

[0923] Step 3:

[0924] The terminal sequentially transmits voice data packets to the server.

[0925] How it works: A smartphone sends voice data packets over the internet to a cloud server, minimizing latency through parallel processing.

[0926] Input: Audio data packets.

[0927] Output: Sending data packets to the server.

[0928] Step 4:

[0929] The server receives the audio data packets and concatenates them to form an audio stream.

[0930] What it does: The server reorders the audio data packets it receives and reconstructs them into a continuous audio stream.

[0931] Input: Audio data packets.

[0932] Output: Audio stream.

[0933] Step 5:

[0934] The server converts the speech to text using a speech recognition engine (e.g., a speech recognition API).

[0935] What happens: The server inputs the audio stream into the Google Cloud Speech-to-Text API and generates text data.

[0936] Input: Audio stream.

[0937] Output: Text data.

[0938] Step 6:

[0939] The server inputs the converted text data into a generative AI (e.g., a generative AI model) for detailed analysis and scoring.

[0940] Specific operation: The server sends the text data to a generation AI (e.g., GPT-4), which analyzes the content of the meeting and quantifies leadership, cooperation, and the presence or absence of harassment.

[0941] Input: Text data.

[0942] Output: Scoring results.

[0943] Step 7:

[0944] The server sends the analysis results to the terminal and displays them on the user interface.

[0945] Specific operation: The server sends the scoring results generated to the user's smartphone, and the score is displayed on the app.

[0946] Input: Scoring results.

[0947] Output: Notification and display on smartphone.

[0948] Step 8:

[0949] The server uses an emotion engine to recognize the emotion data and adjust the scoring results.

[0950] What it does: The server inputs the voice data into an emotion engine (e.g., Emotion API) to recognize the participant's emotional state, and fine-tunes the scoring results based on that.

[0951] Input: Audio data.

[0952] Output: Recognized emotions, adjusted scoring results.

[0953] Step 9:

[0954] The server sends the adjusted results to the terminal and notifies the user.

[0955] Specific operation: The server sends the adjusted scoring results and emotional data to the smartphone and displays detailed feedback.

[0956] Input: Calibrated scoring results and sentiment data.

[0957] Output: Notifications and detailed feedback displayed on your smartphone.

[0958] Step 10:

[0959] The server records the results in a database for long-term storage.

[0960] Specific operation: The server stores the scoring results and emotion data in a database and accumulates them as future evaluation data.

[0961] Input: Scoring results and sentiment data.

[0962] Output: Recorded and accumulated evaluation data in a database.

[0963] (Application example 2)

[0964] 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."

[0965] Conventional meeting evaluation systems can score the content of participants' comments, but they do not perform more detailed analysis, such as detecting the emotions and harassment contained in the comments, or evaluating cooperation. Furthermore, since there is no feedback that takes into account the emotional elements that arise during meetings, there is a problem in that measures to improve the quality of meetings cannot be taken sufficiently. Furthermore, there is a lack of analysis of evaluation trends through long-term data analysis, so there is a lack of indicators for continuously improving meeting outcomes.

[0966] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for recording conference audio in real time, means for dividing the recorded conference audio into data packets and transmitting them to a computer system, means for converting the audio data received by the computer system into text data, means for analyzing the text data and scoring the speech content of the conference participants, means for notifying the scoring results to an information display terminal, means for long-term saving of the analysis results and accumulating them as evaluation data, and means for recognizing the emotions of the conference participants using an emotion recognition module and adjusting the scoring results based on the emotion data. This enables detailed analysis and feedback to improve the quality of the conference, and enables sustainable measures to improve the outcome of the conference through adjustment of scoring results taking into account the emotional states of the conference participants and accumulation and analysis of long-term evaluation data.

[0967] A "means for recording conference audio in real time" is a device that has the function of instantly recording audio spoken during a conference.

[0968] "Means for dividing the recorded conference audio into data packets and transmitting them to a computer system" refers to a device or software that has the function of dividing the recorded audio data into packets of a certain size and transmitting them to a server via a communication network.

[0969] "Means for converting voice data received by a computer system into text data" refers to the process of converting voice data received by a server into text information using voice recognition technology.

[0970] The "means for analyzing text data and scoring the content of statements made by meeting participants" is a system that has the function of analyzing converted text data and calculating various evaluation indicators based on the content of statements made by meeting participants.

[0971] "Means for notifying the information display terminal of the scoring results" refers to a mechanism for transmitting the scoring information obtained as an analysis result to the device used by the user and displaying or notifying the user.

[0972] "Means for storing analysis results over the long term and accumulating them as evaluation data" refers to a system that stores the generated analysis results and scores in a database and maintains that data for a long period of time.

[0973] "Means for recognizing the emotions of meeting participants using an emotion recognition module and adjusting the scoring results based on that emotion data" is a component that has the function of recognizing emotions from the voices and facial expressions of meeting participants, and appropriately correcting the scoring results to reflect their emotional state.

[0974] The "means for scoring leadership, cooperation, and the presence or absence of harassment" is a system that has the function of evaluating the strength of leadership, the degree of cooperation, and whether harassment is present based on the comments and actions of meeting participants.

[0975] The "means for providing detailed feedback" is a mechanism for providing specific and useful feedback to the user based on the scoring results and emotional data.

[0976] "Means of analyzing long-term evaluation tendencies and trends based on accumulated data" refers to the process of analyzing stored historical data to identify evaluation tendencies and trends that emerge over time.

[0977] "Means that take into account the correlation with emotional data" refers to a method of analyzing emotional data in combination with other evaluation indicators to verify the causal relationship and influence between emotions and behavior.

[0978] The present invention relates to a conference evaluation system for improving the quality of conferences, and includes means for recording conference audio in real time, means for dividing the recorded audio into data packets and transmitting them to a computer system, means for converting the audio data received by the computer system into text data, means for analyzing the text data and scoring the content of comments made by conference participants, means for notifying an information display terminal of the scoring results, means for long-term storage of the analysis results and accumulation as evaluation data, and means for recognizing the emotions of conference participants using an emotion recognition module and adjusting the scoring results based on that emotion data. This system makes it possible to improve the quality of conferences.

[0979] System configuration

[0980] The system consists of the following main components:

[0981] 1. User device (PC, smartphone, tablet, etc.)

[0982] 2. Server (computer system that receives and analyzes voice data)

[0983] 3. Network (communications infrastructure connecting user devices and servers)

[0984] 4. Emotion Recognition Module (Module that recognizes the emotions of meeting participants)

[0985] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and analyzed using generative AI. An emotion recognition module also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and various indicators (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) are evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored long-term and accumulated as evaluation data.

[0986] System Operation

[0987] A specific example of the operation of this system is shown below.

[0988] Meeting Scenarios

[0989] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[0990] 1. User A starts the recording application on their device and records the conference audio in real time.

[0991] 2. The user terminal divides the recorded voice into data packets and sends them to the server over the network.

[0992] 3. The server uses a speech recognition engine (e.g., Google Speech-to-Text API) to convert the received voice data into text data.

[0993] 4. The converted text data is analyzed by a generative AI (e.g., a generative AI model such as GPT-4).

[0994] 5. Based on the analysis results, the server scores the comments made by the conference participants and notifies the user terminal of the scoring results.

[0995] 6. The emotion recognition module recognizes the emotions of meeting participants from the voice data and adjusts the scoring results based on this emotion data.

[0996] 7. The server notifies the user device of the adjusted scoring results and detailed feedback.

[0997] 8. Analysis results and emotional data will be stored for a long period of time and accumulated on the server as evaluation data.

[0998] 9. The accumulated data will be analyzed to help improve the quality of future meetings, and long-term evaluation and trend analysis will be conducted.

[0999] Specific examples of hardware and software used

[1000] User devices: PC, smartphone, tablet

[1001] Server: High-performance cloud server

[1002] Network: Internet, LAN

[1003] Speech recognition engine: Google Speech-to-Text API

[1004] Generative AI models: GPT-4 and similar models

[1005] Emotion Recognition Module: Dedicated emotion analysis software

[1006] Prompt Sentence Examples

[1007] "Please analyze this text and rate it on points like leadership, collaboration, and customer service. I'd also like you to do a sentiment analysis."

[1008] This will improve the quality of meetings and enable fair evaluations and prompt provision of appropriate feedback. Furthermore, the introduction of an emotion recognition module will provide detailed feedback based on the user's emotional state, enabling a better meeting experience.

[1009] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1010] Step 1:

[1011] A user starts a meeting and launches a recording application on their device. Recording of the meeting audio begins. The input is the user's voice, and the output is the recorded audio data. Specifically, the microphone on the user's device picks up the audio, and the recording software records it.

[1012] Step 2:

[1013] The device divides the recorded voice into data packets in real time and sends them to the server via the network. The input is the recorded voice data, and the output is the voice data divided into data packets. Specifically, the recording application on the user device divides the recorded data into small data packets and sends them to the server via the Internet.

[1014] Step 3:

[1015] The server concatenates the received audio data to form a single audio data stream. The input is audio data divided into data packets, and the output is the concatenated audio data. Specifically, the server program reassembles the arriving data packets into the correct order and saves them as a stream of audio data.

[1016] Step 4:

[1017] The server converts the concatenated voice data into text data using a speech recognition engine (e.g., Google Speech-to-Text API). The input is the concatenated voice data, and the output is text data. Specifically, the speech recognition engine analyzes the voice samples and generates corresponding text.

[1018] Step 5:

[1019] The server inputs the text data into a generative AI model (e.g., GPT-4) for detailed analysis. The input is text data, and the output is analysis results and scoring data. An example prompt is, "Analyze this text and evaluate points such as leadership, cooperation, and customer responsiveness. Please also perform sentiment analysis." Specifically, the generative AI model analyzes the text data and generates a score based on each indicator.

[1020] Step 6:

[1021] The emotion recognition module recognizes the emotions of conference participants from the voice data and generates emotion data. The input is voice data and the output is emotion data. Specifically, the emotion recognition software analyzes the tone and rate of the voice to determine emotional states such as stress, joy, or anger.

[1022] Step 7:

[1023] The server adjusts the initial scoring result based on the emotion data. The input is the initial scoring result and emotion data, and the output is the adjusted scoring result. Specifically, the scoring algorithm incorporates the emotion data and recalculates the scoring result.

[1024] Step 8:

[1025] The server sends the adjusted scoring results and detailed feedback to the user terminal. The input is the adjusted scoring results and feedback data, and the output is a notification to the user terminal. Specifically, the server sends the generated detailed feedback comments to the user terminal and displays them through the application interface.

[1026] Step 9:

[1027] The analysis results and emotion data are stored over the long term and accumulated on the server as evaluation data. The input is the analysis results and emotion data, and the output is data stored in the database. Specifically, the server's database management system efficiently records this data and makes it available for future trend analysis and evaluation.

[1028] Step 10:

[1029] The server analyzes long-term evaluation trends and trends based on the accumulated data and generates reports. The input is the long-term stored evaluation data, and the output is the evaluation report. Specifically, data analysis software analyzes past data, extracts notable trends and patterns, and generates a report to be provided to the user.

[1030] 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.

[1031] 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.

[1032] 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.

[1033] [Third embodiment]

[1034] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1035] 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.

[1036] 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).

[1037] 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.

[1038] 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.

[1039] 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).

[1040] 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.

[1041] 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.

[1042] 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.

[1043] 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.

[1044] 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.

[1045] 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."

[1046] This invention relates to a meeting evaluation system for improving the quality of meetings. This system records and analyzes meeting audio in real time and scores the remarks of meeting participants to monitor harassment and evaluate facilitation skills.

[1047] System configuration

[1048] The system consists of the following main components:

[1049] 1. User device (PC, smartphone, tablet, etc.)

[1050] 2. Server (a computer system that receives and analyzes voice data)

[1051] 3. Network (communications infrastructure connecting user devices and servers)

[1052] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided.

[1053] Specific operation example

[1054] A specific example of the operation of this system is shown below.

[1055] Meeting Scenarios

[1056] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[1057] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[1058] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[1059] 3. The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine (e.g., voice recognition API).

[1060] 4. The converted text data is input into a generative AI (e.g., a generative AI model) to perform a detailed analysis of the meeting content.

[1061] 5. Based on the analysis results, the server scores participants on their leadership, cooperativeness, and whether or not they engage in harassment.

[1062] For example, user A's leadership score is 85, user B's contribution score is 70, cooperativeness score is 90, and no harassment is detected.

[1063] 6. The server sends the scoring results to the user devices of User A and User B and displays the results visually through the user interface.

[1064] 7. The device will provide detailed score information and feedback comments from the generated AI, allowing users to review their results.

[1065] 8. The server records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[1066] Explanation of program processing

[1067] 1. A user starts a meeting and launches the recording app.

[1068] 2. The terminal captures the conference audio and divides it into data packets in real time.

[1069] 3. The terminal sequentially transmits voice data packets to the server.

[1070] 4. The server receives the audio data packets and concatenates them to form an audio stream.

[1071] 5. The server uses a speech recognition engine to convert the speech to text.

[1072] 6. The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[1073] 7. The server sends the analysis results to the terminal and displays them on the user interface.

[1074] 8. The server records the results in a database and stores them as long-term data.

[1075] Through these processes, the present invention can improve the quality of meetings and achieve fair evaluations. In addition, the harassment monitoring function can effectively manage risk within a company. In this way, the embodiments for carrying out the invention have been specifically described.

[1076] The processing flow will be explained below.

[1077] Step 1:

[1078] A user starts a conference and launches the recording application on their device. When the user presses the "Start Recording" button, the device captures the conference audio through the audio input device (microphone).

[1079] Step 2:

[1080] The terminal divides the recorded conference audio into data packets in real time. The terminal divides the captured audio data into small packets with a certain byte size and transmits them sequentially to the server via the network.

[1081] Step 3:

[1082] The server rearranges the received audio data packets in order and concatenates them in a stream format. The server checks the data integrity and requests packet retransmission if necessary.

[1083] Step 4:

[1084] The server inputs the formatted audio stream into a speech recognition engine (e.g., speech recognition API) and converts the audio data into text data. The server receives the converted text data from the speech recognition engine and temporarily stores it in a database.

[1085] Step 5:

[1086] The server inputs the text data into a generative AI (e.g., a generative AI model) and instructs the generative AI to "analyze the content of the meeting based on this text data and score it on indicators such as leadership, cooperation, and the presence or absence of harassment."

[1087] Step 6:

[1088] The server receives the analysis results output by the generation AI and extracts scores and detailed feedback for each indicator. For example, User A's leadership score might be 85 points, User B's contribution score 70 points, cooperation score 90 points, and no harassment detected.

[1089] Step 7:

[1090] The server sends the scoring results and detailed feedback comments in JSON format to the user's device. The server records the result transmission log and checks the success or failure status.

[1091] Step 8:

[1092] The device receives the JSON data, parses it, and displays it in the user interface, where the user can visually check the score details and feedback from the generating AI.

[1093] Step 9:

[1094] The server structures the scoring results and analysis and records them in a database. The recorded data is backed up regularly and stored as long-term evaluation data.

[1095] Step 10:

[1096] The server generates reports based on the accumulated evaluation data, analyzing long-term evaluation trends and other trends, and provides these reports to administrators to help improve the quality of corporate meetings.

[1097] This enables the meeting evaluation system to improve the quality of meetings and provide fair evaluations and appropriate feedback quickly.

[1098] Example 1

[1099] 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."

[1100] Conventional meeting evaluation systems often manually record and analyze meeting audio, making it difficult to effectively evaluate meeting quality. They also lacked a means to objectively evaluate participants' leadership and cooperation, or to automatically detect harassment. As a result, it was difficult to provide feedback to improve meeting quality, and risk management across the company was inadequate.

[1101] 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.

[1102] In this invention, the server includes a means for dividing the recorded conference audio into data packets and transmitting them to the server, a means for converting the received audio data into text data, a means for inputting the converted text data into a generative AI model for analysis and scoring the speech content of the conference participants, a means for notifying the user terminal of the scoring results, and a means for long-term storage of the analysis results and accumulating them as evaluation data. This enables effective evaluation of the quality of the conference and provides detailed feedback to participants. It also automates the evaluation of leadership and cooperation and the monitoring of harassment, strengthening risk management in companies.

[1103] "Conference audio" refers to all audio data spoken during a conference.

[1104] "Recording means" refers to the equipment or method for recording conference audio in real time and converting it into digital data.

[1105] A "data packet" refers to a data unit that divides a series of audio data into a certain size and transmits it over a network.

[1106] "Server" refers to a computing system that processes data received from a terminal, generates analysis results, and manages them.

[1107] "Audio data" refers to data in which conference audio has been converted into digital format.

[1108] "Text data" refers to character string data converted from voice data using voice recognition technology.

[1109] A "generative AI model" refers to an artificial intelligence model that performs natural language processing and analysis based on large amounts of data.

[1110] "Analyzing" refers to the process of using certain algorithms or models to make sense of input data and extract relevant information.

[1111] "Content of remarks" refers to the general content of remarks made by conference participants during the conference.

[1112] "Scoring" refers to assigning points based on analysis results according to certain evaluation criteria.

[1113] "User terminal" refers to a device used by a conference participant to record voice and display the analysis results.

[1114] "Means for notifying" refers to a method or device for transmitting information from a server to a user terminal and informing the user.

[1115] "Storage means" refers to methods and devices for retaining data for long periods of time.

[1116] "Evaluation data" refers to data including analysis results and scoring results of meetings.

[1117] "Detailed feedback" refers to specific and analytical evaluations and information on areas for improvement provided to meeting participants.

[1118] "Leadership" refers to the criteria used to evaluate meeting participants' ability to lead and provide direction.

[1119] "Collaborative ability" refers to a criterion that evaluates meeting participants' ability to cooperate and collaborate as team members.

[1120] "Harassment" refers to inappropriate comments or behavior that may occur during a meeting.

[1121] This invention relates to a meeting evaluation system for improving the quality of meetings. This system records and analyzes meeting audio in real time and scores the remarks made by meeting participants to monitor harassment and evaluate facilitation skills. The detailed configuration and operation of the system are shown below.

[1122] System configuration

[1123] The system consists of the following main components:

[1124] 1. User device (e.g. PC, smartphone, tablet, etc.)

[1125] 2. Server (a computer system that receives and analyzes voice data)

[1126] 3. Network (communications infrastructure connecting user devices and servers)

[1127] Operational Overview

[1128] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted to text data on the server side and further analyzed using a generative AI model. The analysis results are scored based on the content of the speech, and various indicators (such as leadership, cooperation, and the presence or absence of harassment) are evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided.

[1129] Specific operation example

[1130] A specific example of the operation of this system is shown below.

[1131] Meeting Scenarios

[1132] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[1133] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[1134] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[1135] 3. The server arranges and concatenates the received voice data in order and converts it into text data using a voice recognition engine (e.g., Google Cloud Speech-to-Text).

[1136] 4. The converted text data is input into a generative AI (e.g., GPT-4 model) to perform a detailed analysis of the meeting content.

[1137] 5. Based on the analysis results, the server scores participants on their leadership, cooperativeness, and whether or not they engage in harassment.

[1138] For example, user A's leadership score is 85, user B's contribution score is 70, cooperativeness score is 90, and no harassment is detected.

[1139] 6. The server sends the scoring results to the user's device and displays the results visually through the user interface.

[1140] 7. The device will provide detailed score information and feedback comments from the generated AI, allowing users to review their results.

[1141] 8. The server records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[1142] Example of input prompt for generative AI model

[1143] Example prompt:

[1144] Analyze the audio recording of the meeting below and rate each participant's leadership, cooperation, and whether or not they engaged in harassment.

[1145] text:

[1146] User A: Today's meeting is about the progress of the new project. Everyone, please give us your opinions.

[1147] User B: I'll be in charge of the progress report. Currently, Task A is 80% complete.

[1148] ...

[1149] In this way, by showing detailed embodiments of the invention, it is possible to effectively evaluate the quality of meetings and provide clear feedback to participants through scoring of leadership and cooperation, which is expected to improve meeting management and strengthen corporate risk management.

[1150] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1151] Step 1:

[1152] A user starts a meeting and launches the recording app.

[1153] Specifically, when a meeting starts, the user launches a dedicated recording application on their device.

[1154] Specific operation: For example, User A double-clicks the meeting recording app on his / her PC and clicks the "Start Recording" button.

[1155] Input: Conference audio (analog audio)

[1156] Output: Launched recording application

[1157] Step 2:

[1158] The terminal captures the conference audio and divides it into data packets in real time.

[1159] Specific description: The terminal records the conference audio through a microphone and performs the process of dividing it into small data packets.

[1160] What it does: The PC's microphone captures audio data every second, and then divides each into data packets every 0.5 seconds.

[1161] Input: Conference audio (analog audio)

[1162] Output: Data packet (digital data)

[1163] Step 3:

[1164] The terminal sequentially transmits voice data packets to the server.

[1165] Specific explanation: The terminal sequentially transmits the divided voice data packets to the server.

[1166] What it does: Your PC sends packetized voice data to a server every second via your internet connection.

[1167] Input: Data packets (digital voice data)

[1168] Output: Data packet sent to the server

[1169] Step 4:

[1170] The server receives the audio data packets and concatenates them to form an audio stream.

[1171] Specifically, the server concatenates the received audio data packets in order to create a continuous audio stream.

[1172] Specific operation: The server combines the audio data packets received every 0.5 seconds to restore the original continuous audio data.

[1173] Input: Data packets (digital voice data)

[1174] Output: Concatenated audio stream (continuous audio data)

[1175] Step 5:

[1176] The server uses a speech recognition engine to convert the speech into text.

[1177] Specifically, the server uses a speech recognition engine (e.g., Google Cloud Speech-to-Text) to convert the voice data into text data.

[1178] What happens: The server sends the audio file to a speech recognition API and retrieves the conversation in text format.

[1179] Input: Concatenated audio stream (continuous audio data)

[1180] Output: Text data (string format)

[1181] Step 6:

[1182] The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[1183] Specific explanation: The server inputs text data into a generative AI (e.g., GPT-4 model) to analyze the meeting content and score what is said.

[1184] Specific behavior: The server passes the following prompt to the generation AI:

[1185] Example prompt:

[1186] Analyze the audio recording of the meeting below and rate each participant's leadership, cooperation, and whether or not they engaged in harassment.

[1187] text:

[1188] User A: Today's meeting is about the progress of the new project. Everyone, please give us your opinions.

[1189] User B: I'll be in charge of the progress report. Currently, Task A is 80% complete.

[1190] ...

[1191] Input: Text data (string format)

[1192] Output: Analysis results and scoring data (evaluation index)

[1193] Step 7:

[1194] The server sends the analysis results to the terminal and displays them on the user interface.

[1195] Specific explanation: The server sends the analysis results returned by the generation AI to the user terminal and displays them on the user interface.

[1196] Specific operation: The server sends the evaluation results (e.g., User A's leadership score is 85 points, User B's contribution score is 70 points) to the devices of User A and User B, and visualizes them on a dedicated dashboard.

[1197] Input: Analysis results and scoring data (evaluation index)

[1198] Output: Evaluation results displayed on the user's terminal

[1199] Step 8:

[1200] The server records the results in a database and accumulates them as long-term evaluation data.

[1201] Specific explanation: The server stores the analysis results and each meeting data in a database and manages them as long-term evaluation data.

[1202] Specific operation: The server records the evaluation results and speech analysis data in a MySQL database for future comparative analysis.

[1203] Input: Analysis results and scoring data (evaluation index)

[1204] Output: Evaluation data stored in a database

[1205] This system improves the quality of meetings and enables fair evaluations. It supports productive meeting management while also providing a harassment monitoring function, strengthening risk management across the company.

[1206] (Application example 1)

[1207] 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."

[1208] Conventional meeting evaluation systems can evaluate the quality of meetings and the behavior of participants in real time, but they are unable to evaluate in-car communication or safety awareness. Furthermore, there is a lack of systems that can analyze in-car conversations in detail to evaluate the driver's attention and whether or not harassment has occurred. This has resulted in insufficient prevention of harassment and improvement of safety awareness in the in-car environment.

[1209] 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.

[1210] In this invention, the server includes means for recording conference audio in real time, means for dividing the recorded conference audio into data packets and transmitting them to the server, means for converting the received audio data into text data, means for analyzing the text data and scoring the speech content of the conference participants, means for recording in-car conversations in real time, means for scoring the driver's attention, safety awareness, and presence or absence of harassment through analysis, means for notifying the scoring results to a user terminal, and means for long-term storage of the analysis results and accumulating them as evaluation data. This makes it possible to evaluate in-car communication quality, safety awareness, and presence or absence of harassment in detail. This also makes it possible to appropriately monitor the behavior of drivers and passengers and improve the safety and comfort of the in-car environment.

[1211] "Conference audio" refers to the audio data of the participants' comments and interactions that occur during a conference.

[1212] A "data packet" is a small unit of digital data divided into pieces and formatted for efficient transmission over a network.

[1213] A "server" is a central computer system for collecting, analyzing, and storing data.

[1214] "Text data" refers to text information converted from speech using speech recognition technology.

[1215] "Scoring" is an evaluation method in which points are assigned according to specific evaluation indicators based on analyzed data.

[1216] "User terminal" refers to a device that is directly operated by a user, such as a PC, smartphone, or tablet.

[1217] "Real-time" refers to data being processed and transferred immediately with little delay.

[1218] "Attention" is an indicator of how much a driver is concentrating on traffic conditions and driving operations.

[1219] "Harassment" refers to inappropriate words, actions or behaviour towards other crew members.

[1220] "Analysis" is the process of analyzing collected data in detail to clarify its meaning and trends.

[1221] "Notification" refers to the act of informing the user of the analysis results or scoring results.

[1222] "Evaluation data" refers to data that includes analysis results and scoring results, and is used as the basis for long-term evaluation and improvement.

[1223] This invention relates to an in-car communication evaluation system. The purpose is to improve the safety and comfort of the in-car environment by recording and analyzing in-car conversations in real time and scoring passenger attention, safety awareness, and the presence or absence of harassment. This system consists of the following main components:

[1224] System configuration

[1225] 1. User Device

[1226] This refers to in-vehicle systems, smartphones, tablets, etc. that have recording functions and are responsible for recording conversations in real time.

[1227] 2. Server

[1228] It is the central computer system that collects and analyzes data and stores the results.

[1229] 3. Network

[1230] It is a communications infrastructure that connects user terminals and servers, and uses wireless communication technology.

[1231] Specific operation example

[1232] A specific example of the operation of this system is shown below.

[1233] In-car communication scenario

[1234] Consider a situation in which user A (driver) and user B (passenger) are having a conversation in a car.

[1235] 1. User terminal: The vehicle's in-vehicle system activates the recording function and presses the "Start Recording" button.

[1236] 2. Terminal: Records conversations inside the car in real time, divides the voice data into data packets, and sends them to the server sequentially.

[1237] 3. Server: The received voice data is arranged in order and concatenated, and the voice data is converted into text data using a voice recognition engine (e.g., speech_recognition library).

[1238] 4. Server: The converted text data is input into a generative AI model (e.g., GPT-3) for detailed analysis.

[1239] 5. Server: Based on the analysis results, the server scores participants on their attention, safety awareness, and whether or not they have engaged in harassment.

[1240] For example, user A's attention score is 85, user B's safety awareness score is 90, and no harassment is detected.

[1241] 6. Server: Sends the scoring results to the user's terminal and displays the results visually through the user interface.

[1242] 7. Device: Provides detailed score information and feedback comments from the generated AI, allowing users to check their results.

[1243] 8. Server: Records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[1244] Hardware and software used

[1245] Speech recognition engine: speech_recognition library

[1246] Generative AI model: GPT-3

[1247] User devices: in-vehicle systems, smartphones, tablets, etc.

[1248] Server: a computer system for data analysis and storage

[1249] Network: Wireless communication technology (e.g., Wi-Fi, 4G, 5G)

[1250] Adding specific examples

[1251] Driving conversation scenario:

[1252] Expect conversations such as, "Are you concentrating on driving?", "Drive carefully," and "We're not in a hurry, so drive safely."

[1253] Example prompts to input to a generative AI model:

[1254] Based on the conversation below, please rate the level of attention, safety awareness, and whether or not there was any harassment.

[1255] example:

[1256] How was your drive today?

[1257] Drive carefully. We're not in a hurry, so safety comes first.

[1258] I think you are being harsh towards me.

[1259] The present invention thereby provides a means for evaluating in-vehicle communications in real time and for appropriately monitoring occupant behavior.

[1260] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1261] Step 1:

[1262] The user terminal activates the recording function of the vehicle's in-vehicle system and operates the "start recording" button.

[1263] Input: Start recording

[1264] Output: Real-time recording begins

[1265] Specific operation: The user starts recording on the in-car system or smartphone. The device prepares to record the conversation in real time.

[1266] Step 2:

[1267] The terminal records conversations in the car in real time, divides the voice data into data packets, and transmits them to the server one by one.

[1268] Input: Voice data of conversations in a car

[1269] Output: sent as data packets to the server

[1270] Specific operation: The device divides the recorded audio into small data packets and sends them to the server over the network.

[1271] Step 3:

[1272] The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine.

[1273] Input: Packetized voice data

[1274] Output: Converted text data

[1275] Specific operation: The server uses a speech recognition library (e.g., speech_recognition) to analyze the voice data and convert it into text data.

[1276] Step 4:

[1277] The server inputs the converted text data into a generative AI model (e.g., GPT-3) for detailed analysis.

[1278] Input: Text data

[1279] Output: Analysis results (e.g., scores for attention, safety awareness, and whether or not there was harassment)

[1280] Specific operation: The server inputs prompts into the generative AI model, which then performs detailed analysis and scoring based on the conversation content.

[1281] Step 5:

[1282] Based on the analysis results, the server scores the person's attention, safety awareness, and whether or not they have engaged in harassment.

[1283] Input: Analysis results from a generative AI model

[1284] Output: Each score (e.g., attention 85 points, safety awareness 90 points, no harassment detected)

[1285] Specific operation: The server quantifies specific indicators as scores based on the output of the generation AI.

[1286] Step 6:

[1287] The server transmits the scoring results to the user terminal and visually displays the results through a user interface.

[1288] Input: Scoring results

[1289] Output: Visual feedback displayed in the user interface

[1290] Specific operation: The server sends the score and analysis results to the user's device, which then displays the results on the screen.

[1291] Step 7:

[1292] The device provides the user with detailed score information and feedback comments from the generated AI for review.

[1293] Input: Scoring results received from the server

[1294] Output: Feedback comments that the user sees

[1295] Specific operation: The terminal displays the scoring results and feedback comments for the user to review.

[1296] Step 8:

[1297] The server records the scoring results and analysis details in a database and accumulates them as long-term evaluation data.

[1298] Input: Scoring results and analysis details

[1299] Output: Evaluation data stored in a database

[1300] Specific operation: The server records all analysis results in a database and stores the data for future evaluation and improvement.

[1301] 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.

[1302] The present invention relates to a meeting evaluation system for improving the quality of meetings. This system not only records and analyzes meeting audio in real time, scoring the content of participants' remarks, but also monitors harassment and evaluates facilitation skills by recognizing participants' emotions.

[1303] System configuration

[1304] The system consists of the following main components:

[1305] 1. User device (PC, smartphone, tablet, etc.)

[1306] 2. Server (a computer system that receives and analyzes voice data)

[1307] 3. Network (communications infrastructure connecting user devices and servers)

[1308] 4. Emotion engine (a module that recognizes the emotions of meeting participants)

[1309] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The emotion engine also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored over the long term and accumulated as evaluation data.

[1310] Specific operation example

[1311] A specific example of the operation of this system is shown below.

[1312] Meeting Scenarios

[1313] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[1314] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[1315] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[1316] 3. The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine (e.g., voice recognition API).

[1317] 4. The converted text data is input into a generative AI (e.g., a generative AI model) to perform a detailed analysis of the meeting content.

[1318] 5. Based on the analysis results, the server scores participants on their leadership, cooperation, and harassment.

[1319] For example, user A's leadership score is 85 points, user B's contribution score is 70 points, cooperativeness score is 90 points, and no harassment is detected.

[1320] 6. The emotion engine recognizes the participant's emotions from the voice data and sends the data to the server.

[1321] 7. The server adjusts the scoring results based on the recognized emotion data and generates detailed feedback.

[1322] For example, if user A feels stressed during a meeting, feedback is provided based on that emotion.

[1323] 8. The server sends the scoring results and emotion data to the user devices of User A and User B, and visually displays the results through the user interface.

[1324] 9. The device will provide detailed scores, feedback comments from the generative AI, and an evaluation from the emotion engine, allowing users to review their results.

[1325] 10. The server records the scoring results, analysis results, and emotional data in a database and accumulates them as long-term evaluation data.

[1326] Explanation of program processing

[1327] 1. A user starts a meeting and launches the recording app.

[1328] 2. The terminal captures the conference audio and divides it into data packets in real time.

[1329] 3. The terminal sequentially transmits voice data packets to the server.

[1330] 4. The server receives the audio data packets and concatenates them to form an audio stream.

[1331] 5. The server uses a speech recognition engine to convert the speech to text.

[1332] 6. The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[1333] 7. The server sends the analysis results to the terminal and displays them on the user interface.

[1334] 8. The server uses the emotion engine to recognize the emotion data and adjust the scoring results.

[1335] 9. The server sends the adjusted results to the terminal and notifies the user.

[1336] 10. The server records the results in a database and stores them as long-term data.

[1337] Through these processes, the system improves the quality of meetings and provides fair evaluations and appropriate feedback quickly.In addition, the introduction of an emotion engine provides detailed feedback based on the user's emotional state, resulting in a better meeting experience.

[1338] The processing flow will be explained below.

[1339] Step 1:

[1340] A user starts a conference and launches the recording application on their device. When the user presses the "Start Recording" button, the device captures the conference audio through the audio input device (microphone).

[1341] Step 2:

[1342] The terminal divides the recorded conference audio into data packets in real time. The terminal divides the captured audio data into small packets with a certain byte size and transmits them sequentially to the server via the network.

[1343] Step 3:

[1344] The server reorders the received audio data packets and concatenates them into a stream. The server checks the data integrity and requests packet retransmission if necessary.

[1345] Step 4:

[1346] The server inputs the formatted audio stream into a speech recognition engine (e.g., speech recognition API) and converts the audio data into text data. The server receives the converted text data from the speech recognition engine and temporarily stores it in a database.

[1347] Step 5:

[1348] The server inputs the text data into a generative AI (e.g., a generative AI model) and instructs the generative AI to "analyze the content of the meeting based on this text data and score it on indicators such as leadership, cooperation, and the presence or absence of harassment."

[1349] Step 6:

[1350] The server receives the analysis results output by the generation AI and extracts scores and detailed feedback for each indicator. For example, User A's leadership score might be 85 points, User B's contribution score 70 points, cooperation score 90 points, and no harassment detected.

[1351] Step 7:

[1352] The emotion engine recognizes the emotions of conference participants from the voice data. The server provides the voice data to the emotion engine, which analyzes the emotional state. For example, emotions such as anger, anxiety, and joy are recognized.

[1353] Step 8:

[1354] The server receives the recognized emotion data from the emotion engine and adjusts the scoring results, such as affecting leadership and cooperation scores based on the emotion data, or adding warnings if there is a high likelihood of harassment.

[1355] Step 9:

[1356] The server sends the adjusted scoring results and detailed emotional feedback in JSON format to the user's device. The server records the result transmission log and checks the success or failure status.

[1357] Step 10:

[1358] The device receives the JSON data, parses it, and displays it on the user interface, where the user can visually check the score details, feedback from the AI, and the analysis of the user's emotional state.

[1359] Step 11:

[1360] The server structures the scoring results, analysis details, and emotion data and records them in a database. The recorded data is backed up regularly and stored as long-term evaluation data.

[1361] Step 12:

[1362] The server generates reports that analyze long-term evaluation trends and trends based on the accumulated evaluation data and emotion data. The server provides these reports to administrators, helping to improve the quality of corporate meetings.

[1363] This system improves the quality of meetings, providing fair evaluations and prompt, appropriate feedback. Furthermore, the introduction of an emotion engine provides detailed feedback based on the user's emotional state, enabling a better meeting experience.

[1364] Example 2

[1365] 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."

[1366] Accurately evaluating the content and quality of comments made in meetings and providing feedback to meeting participants is important for improving the performance of an entire organization. However, conventional meeting evaluation systems have difficulty evaluating not only the content of participants' comments but also their emotions, and it is not possible to accumulate this data over the long term and perform trend analysis. Therefore, a system that comprehensively evaluates the quality of meetings and provides detailed feedback is needed.

[1367] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for recording the conference audio in real time, a means for dividing the recorded conference audio into data packets and transmitting them to the server, a means for converting the audio data received on the server side into text data, a means for analyzing the text data and scoring the remarks of the conference participants, a means for notifying the scoring results to the user terminal, a means for long-term saving the analysis results and accumulating them as evaluation data, and a means for recognizing the emotions of the conference participants and adjusting the scoring results. This makes it possible to comprehensively evaluate the quality of the conference in real time and provide fair feedback.

[1368] "Conference audio" refers to audio data of conversations and presentations that occur during a conference.

[1369] "Real-time recording means" refers to a device or method that instantly records conference audio and stores it as digital data.

[1370] "Data packet" refers to the unit in which digital data is divided into small pieces for efficient transmission over a network.

[1371] "Means for transmitting to the server" refers to the technology or method for transmitting data packets from the user terminal to the server.

[1372] "Means for converting into text data" refers to speech recognition technology or algorithms that convert voice data into text information.

[1373] "Means for analyzing text data" refers to programs or processing methods for analyzing data converted into text and understanding its content.

[1374] "Means for scoring the content of statements made by meeting participants" refers to algorithms or methods for evaluating and quantifying the content of statements.

[1375] "Means for notifying the user terminal of the scoring results" refers to a method or system for notifying the user of the analysis and evaluation results.

[1376] "Means for storing analysis results for the long term and accumulating them as evaluation data" refers to technology that stores analysis results in a database or the like for a long period of time and uses them for future analysis and evaluation.

[1377] "Means for recognizing the emotions of meeting participants" refers to technologies and methods for identifying the emotional state of meeting participants from voice or text data.

[1378] "Means for adjusting scoring results" refers to algorithms or methods for correcting or amending initial scoring results based on recognized emotion data.

[1379] The present invention relates to a meeting evaluation system for improving the quality of meetings. This system not only records and analyzes meeting audio in real time and scores the content of participants' comments, but also monitors harassment and evaluates facilitation ability by recognizing participants' emotions.

[1380] System configuration

[1381] The system consists of the following main components:

[1382] 1. User device (PC, smartphone, tablet, etc.)

[1383] 2. Server (a computer system that receives and analyzes voice data)

[1384] 3. Network (communications infrastructure connecting user devices and servers)

[1385] 4. Emotion engine (a module that recognizes the emotions of meeting participants)

[1386] 5. Database (a storage device for long-term storage of analysis results)

[1387] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The emotion engine also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored over the long term and accumulated as evaluation data.

[1388] Specific operation example

[1389] A specific example of the operation of this system is shown below.

[1390] Meeting Scenarios

[1391] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[1392] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[1393] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[1394] 3. The server arranges and concatenates the received voice data in order and converts it into text data using a speech recognition engine (e.g., Google Cloud Speech-to-Text API).

[1395] 4. The converted text data is input into a generative AI (e.g., OpenAI GPT-4) to perform a detailed analysis of the meeting content.

[1396] 5. Based on the analysis results, the server scores participants on their leadership, cooperation, and harassment.

[1397] For example, user A's leadership score is 85 points, user B's contribution score is 70 points, cooperativeness score is 90 points, and no harassment is detected.

[1398] 6. The emotion engine recognizes the participant's emotions from the voice data and sends the data to the server.

[1399] 7. The server adjusts the scoring results based on the recognized emotion data and generates detailed feedback.

[1400] For example, if user A feels stressed during a meeting, feedback is provided based on that emotion.

[1401] 8. The server sends the scoring results and emotion data to the user devices of User A and User B, and visually displays the results through the user interface.

[1402] 9. The device will provide detailed scores, feedback comments from the generative AI, and an evaluation from the emotion engine, allowing users to review their results.

[1403] 10. The server records the scoring results, analysis results, and emotional data in a database and accumulates them as long-term evaluation data.

[1404] Prompt Sentence Examples

[1405] Here are some example prompts to input to a generative AI model (e.g., GPT-4):

[1406] The audio from a meeting was transcribed as follows:

[1407] 1. User A: "How is the project going?"

[1408] 2. User B: "It's going well, but there are some challenges."

[1409] 3. User A: "Do you need any help?"

[1410] Please rate User A's leadership, cooperation, and overall performance on a numerical scale. Also, please rate User B's contribution.

[1411] This allows the system to evaluate the quality of the meeting in real time and provide fair feedback. Furthermore, by incorporating an emotion engine, the system can provide detailed feedback based on the user's emotional state, resulting in a better meeting experience.

[1412] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1413] Step 1:

[1414] A user starts a meeting and launches the recording app.

[1415] Specific actions: User A opens the recording app on their smartphone and presses the "Start Recording" button.

[1416] Input: User action at the start of a meeting.

[1417] Output: Start recording.

[1418] Step 2:

[1419] The terminal captures the conference audio and divides it into data packets in real time.

[1420] How it works: The smartphone picks up the audio from the meeting and packets the audio data every second.

[1421] Input: Conference audio.

[1422] Output: Audio data packets.

[1423] Step 3:

[1424] The terminal sequentially transmits voice data packets to the server.

[1425] How it works: A smartphone sends voice data packets over the internet to a cloud server, minimizing latency through parallel processing.

[1426] Input: Audio data packets.

[1427] Output: Sending data packets to the server.

[1428] Step 4:

[1429] The server receives the audio data packets and concatenates them to form an audio stream.

[1430] What it does: The server reorders the audio data packets it receives and reconstructs them into a continuous audio stream.

[1431] Input: Audio data packets.

[1432] Output: Audio stream.

[1433] Step 5:

[1434] The server converts the speech to text using a speech recognition engine (e.g., a speech recognition API).

[1435] What happens: The server inputs the audio stream into the Google Cloud Speech-to-Text API and generates text data.

[1436] Input: Audio stream.

[1437] Output: Text data.

[1438] Step 6:

[1439] The server inputs the converted text data into a generative AI (e.g., a generative AI model) for detailed analysis and scoring.

[1440] Specific operation: The server sends the text data to a generation AI (e.g., GPT-4), which analyzes the content of the meeting and quantifies leadership, cooperation, and the presence or absence of harassment.

[1441] Input: Text data.

[1442] Output: Scoring results.

[1443] Step 7:

[1444] The server sends the analysis results to the terminal and displays them on the user interface.

[1445] Specific operation: The server sends the scoring results generated to the user's smartphone, and the score is displayed on the app.

[1446] Input: Scoring results.

[1447] Output: Notification and display on smartphone.

[1448] Step 8:

[1449] The server uses an emotion engine to recognize the emotion data and adjust the scoring results.

[1450] What it does: The server inputs the voice data into an emotion engine (e.g., Emotion API) to recognize the participant's emotional state, and fine-tunes the scoring results based on that.

[1451] Input: Audio data.

[1452] Output: Recognized emotions, adjusted scoring results.

[1453] Step 9:

[1454] The server sends the adjusted results to the terminal and notifies the user.

[1455] Specific operation: The server sends the adjusted scoring results and emotional data to the smartphone and displays detailed feedback.

[1456] Input: Calibrated scoring results and sentiment data.

[1457] Output: Notifications and detailed feedback displayed on your smartphone.

[1458] Step 10:

[1459] The server records the results in a database for long-term storage.

[1460] Specific operation: The server stores the scoring results and emotion data in a database and accumulates them as future evaluation data.

[1461] Input: Scoring results and sentiment data.

[1462] Output: Recorded and accumulated evaluation data in a database.

[1463] (Application example 2)

[1464] 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."

[1465] Conventional meeting evaluation systems can score the content of participants' comments, but they do not perform more detailed analysis, such as detecting the emotions and harassment contained in the comments, or evaluating cooperation. Furthermore, since there is no feedback that takes into account the emotional elements that arise during meetings, there is a problem in that measures to improve the quality of meetings cannot be taken sufficiently. Furthermore, there is a lack of analysis of evaluation trends through long-term data analysis, so there is a lack of indicators for continuously improving meeting outcomes.

[1466] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for recording conference audio in real time, means for dividing the recorded conference audio into data packets and transmitting them to a computer system, means for converting the audio data received by the computer system into text data, means for analyzing the text data and scoring the speech content of the conference participants, means for notifying the scoring results to an information display terminal, means for long-term saving of the analysis results and accumulating them as evaluation data, and means for recognizing the emotions of the conference participants using an emotion recognition module and adjusting the scoring results based on the emotion data. This enables detailed analysis and feedback to improve the quality of the conference, and enables sustainable measures to improve the outcome of the conference through adjustment of scoring results taking into account the emotional states of the conference participants and accumulation and analysis of long-term evaluation data.

[1467] A "means for recording conference audio in real time" is a device that has the function of instantly recording audio spoken during a conference.

[1468] "Means for dividing the recorded conference audio into data packets and transmitting them to a computer system" refers to a device or software that has the function of dividing the recorded audio data into packets of a certain size and transmitting them to a server via a communication network.

[1469] "Means for converting voice data received by a computer system into text data" refers to the process of converting voice data received by a server into text information using voice recognition technology.

[1470] The "means for analyzing text data and scoring the content of statements made by meeting participants" is a system that has the function of analyzing converted text data and calculating various evaluation indicators based on the content of statements made by meeting participants.

[1471] "Means for notifying the information display terminal of the scoring results" refers to a mechanism for transmitting the scoring information obtained as an analysis result to the device used by the user and displaying or notifying the user.

[1472] "Means for storing analysis results over the long term and accumulating them as evaluation data" refers to a system that stores the generated analysis results and scores in a database and maintains that data for a long period of time.

[1473] "Means for recognizing the emotions of meeting participants using an emotion recognition module and adjusting the scoring results based on that emotion data" is a component that has the function of recognizing emotions from the voices and facial expressions of meeting participants, and appropriately correcting the scoring results to reflect their emotional state.

[1474] The "means for scoring leadership, cooperation, and the presence or absence of harassment" is a system that has the function of evaluating the strength of leadership, the degree of cooperation, and whether harassment is present based on the comments and actions of meeting participants.

[1475] The "means for providing detailed feedback" is a mechanism for providing specific and useful feedback to the user based on the scoring results and emotional data.

[1476] "Means of analyzing long-term evaluation tendencies and trends based on accumulated data" refers to the process of analyzing stored historical data to identify evaluation tendencies and trends that emerge over time.

[1477] "Means that take into account the correlation with emotional data" refers to a method of analyzing emotional data in combination with other evaluation indicators to verify the causal relationship and influence between emotions and behavior.

[1478] The present invention relates to a conference evaluation system for improving the quality of conferences, and includes means for recording conference audio in real time, means for dividing the recorded audio into data packets and transmitting them to a computer system, means for converting the audio data received by the computer system into text data, means for analyzing the text data and scoring the content of comments made by conference participants, means for notifying an information display terminal of the scoring results, means for long-term storage of the analysis results and accumulation as evaluation data, and means for recognizing the emotions of conference participants using an emotion recognition module and adjusting the scoring results based on that emotion data. This system makes it possible to improve the quality of conferences.

[1479] System configuration

[1480] The system consists of the following main components:

[1481] 1. User device (PC, smartphone, tablet, etc.)

[1482] 2. Server (computer system that receives and analyzes voice data)

[1483] 3. Network (communications infrastructure connecting user devices and servers)

[1484] 4. Emotion Recognition Module (Module that recognizes the emotions of meeting participants)

[1485] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and analyzed using generative AI. An emotion recognition module also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and various indicators (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) are evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored long-term and accumulated as evaluation data.

[1486] System Operation

[1487] A specific example of the operation of this system is shown below.

[1488] Meeting Scenarios

[1489] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[1490] 1. User A starts the recording application on their device and records the conference audio in real time.

[1491] 2. The user terminal divides the recorded voice into data packets and sends them to the server over the network.

[1492] 3. The server uses a speech recognition engine (e.g., Google Speech-to-Text API) to convert the received voice data into text data.

[1493] 4. The converted text data is analyzed by a generative AI (e.g., a generative AI model such as GPT-4).

[1494] 5. Based on the analysis results, the server scores the comments made by the conference participants and notifies the user terminal of the scoring results.

[1495] 6. The emotion recognition module recognizes the emotions of meeting participants from the voice data and adjusts the scoring results based on this emotion data.

[1496] 7. The server notifies the user device of the adjusted scoring results and detailed feedback.

[1497] 8. Analysis results and emotional data will be stored for a long period of time and accumulated on the server as evaluation data.

[1498] 9. The accumulated data will be analyzed to help improve the quality of future meetings, and long-term evaluation and trend analysis will be conducted.

[1499] Specific examples of hardware and software used

[1500] User devices: PC, smartphone, tablet

[1501] Server: High-performance cloud server

[1502] Network: Internet, LAN

[1503] Speech recognition engine: Google Speech-to-Text API

[1504] Generative AI models: GPT-4 and similar models

[1505] Emotion Recognition Module: Dedicated emotion analysis software

[1506] Prompt Sentence Examples

[1507] "Please analyze this text and rate it on points like leadership, collaboration, and customer service. I'd also like you to do a sentiment analysis."

[1508] This will improve the quality of meetings and enable fair evaluations and prompt provision of appropriate feedback. Furthermore, the introduction of an emotion recognition module will provide detailed feedback based on the user's emotional state, enabling a better meeting experience.

[1509] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1510] Step 1:

[1511] A user starts a meeting and launches a recording application on their device. Recording of the meeting audio begins. The input is the user's voice, and the output is the recorded audio data. Specifically, the microphone on the user's device picks up the audio, and the recording software records it.

[1512] Step 2:

[1513] The device divides the recorded voice into data packets in real time and sends them to the server via the network. The input is the recorded voice data, and the output is the voice data divided into data packets. Specifically, the recording application on the user device divides the recorded data into small data packets and sends them to the server via the Internet.

[1514] Step 3:

[1515] The server concatenates the received audio data to form a single audio data stream. The input is audio data divided into data packets, and the output is the concatenated audio data. Specifically, the server program reassembles the arriving data packets into the correct order and saves them as a stream of audio data.

[1516] Step 4:

[1517] The server converts the concatenated voice data into text data using a speech recognition engine (e.g., Google Speech-to-Text API). The input is the concatenated voice data, and the output is text data. Specifically, the speech recognition engine analyzes the voice samples and generates corresponding text.

[1518] Step 5:

[1519] The server inputs the text data into a generative AI model (e.g., GPT-4) for detailed analysis. The input is text data, and the output is analysis results and scoring data. An example prompt is, "Analyze this text and evaluate points such as leadership, cooperation, and customer responsiveness. Please also perform sentiment analysis." Specifically, the generative AI model analyzes the text data and generates a score based on each indicator.

[1520] Step 6:

[1521] The emotion recognition module recognizes the emotions of conference participants from the voice data and generates emotion data. The input is voice data and the output is emotion data. Specifically, the emotion recognition software analyzes the tone and rate of the voice to determine emotional states such as stress, joy, or anger.

[1522] Step 7:

[1523] The server adjusts the initial scoring result based on the emotion data. The input is the initial scoring result and emotion data, and the output is the adjusted scoring result. Specifically, the scoring algorithm incorporates the emotion data and recalculates the scoring result.

[1524] Step 8:

[1525] The server sends the adjusted scoring results and detailed feedback to the user terminal. The input is the adjusted scoring results and feedback data, and the output is a notification to the user terminal. Specifically, the server sends the generated detailed feedback comments to the user terminal and displays them through the application interface.

[1526] Step 9:

[1527] The analysis results and emotion data are stored over the long term and accumulated on the server as evaluation data. The input is the analysis results and emotion data, and the output is data stored in the database. Specifically, the server's database management system efficiently records this data and makes it available for future trend analysis and evaluation.

[1528] Step 10:

[1529] The server analyzes long-term evaluation trends and trends based on the accumulated data and generates reports. The input is the long-term stored evaluation data, and the output is the evaluation report. Specifically, data analysis software analyzes past data, extracts notable trends and patterns, and generates a report to be provided to the user.

[1530] 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.

[1531] 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.

[1532] 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.

[1533] [Fourth embodiment]

[1534] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1535] 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.

[1536] 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).

[1537] 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.

[1538] 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.

[1539] 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).

[1540] 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.

[1541] 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.

[1542] 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.

[1543] 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.

[1544] 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.

[1545] 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.

[1546] 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."

[1547] This invention relates to a meeting evaluation system for improving the quality of meetings. This system records and analyzes meeting audio in real time and scores the remarks of meeting participants to monitor harassment and evaluate facilitation skills.

[1548] System configuration

[1549] The system consists of the following main components:

[1550] 1. User device (PC, smartphone, tablet, etc.)

[1551] 2. Server (a computer system that receives and analyzes voice data)

[1552] 3. Network (communications infrastructure connecting user devices and servers)

[1553] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided.

[1554] Specific operation example

[1555] A specific example of the operation of this system is shown below.

[1556] Meeting Scenarios

[1557] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[1558] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[1559] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[1560] 3. The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine (e.g., voice recognition API).

[1561] 4. The converted text data is input into a generative AI (e.g., a generative AI model) to perform a detailed analysis of the meeting content.

[1562] 5. Based on the analysis results, the server scores participants on their leadership, cooperativeness, and whether or not they engage in harassment.

[1563] For example, user A's leadership score is 85, user B's contribution score is 70, cooperativeness score is 90, and no harassment is detected.

[1564] 6. The server sends the scoring results to the user devices of User A and User B and displays the results visually through the user interface.

[1565] 7. The device will provide detailed score information and feedback comments from the generated AI, allowing users to review their results.

[1566] 8. The server records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[1567] Explanation of program processing

[1568] 1. A user starts a meeting and launches the recording app.

[1569] 2. The terminal captures the conference audio and divides it into data packets in real time.

[1570] 3. The terminal sequentially transmits voice data packets to the server.

[1571] 4. The server receives the audio data packets and concatenates them to form an audio stream.

[1572] 5. The server uses a speech recognition engine to convert the speech to text.

[1573] 6. The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[1574] 7. The server sends the analysis results to the terminal and displays them on the user interface.

[1575] 8. The server records the results in a database and stores them as long-term data.

[1576] Through these processes, the present invention can improve the quality of meetings and achieve fair evaluations. In addition, the harassment monitoring function can effectively manage risk within a company. In this way, the embodiments for carrying out the invention have been specifically described.

[1577] The processing flow will be explained below.

[1578] Step 1:

[1579] A user starts a conference and launches the recording application on their device. When the user presses the "Start Recording" button, the device captures the conference audio through the audio input device (microphone).

[1580] Step 2:

[1581] The terminal divides the recorded conference audio into data packets in real time. The terminal divides the captured audio data into small packets with a certain byte size and transmits them sequentially to the server via the network.

[1582] Step 3:

[1583] The server rearranges the received audio data packets in order and concatenates them in a stream format. The server checks the data integrity and requests packet retransmission if necessary.

[1584] Step 4:

[1585] The server inputs the formatted audio stream into a speech recognition engine (e.g., speech recognition API) and converts the audio data into text data. The server receives the converted text data from the speech recognition engine and temporarily stores it in a database.

[1586] Step 5:

[1587] The server inputs the text data into a generative AI (e.g., a generative AI model) and instructs the generative AI to "analyze the content of the meeting based on this text data and score it on indicators such as leadership, cooperation, and the presence or absence of harassment."

[1588] Step 6:

[1589] The server receives the analysis results output by the generation AI and extracts scores and detailed feedback for each indicator. For example, User A's leadership score might be 85 points, User B's contribution score 70 points, cooperation score 90 points, and no harassment detected.

[1590] Step 7:

[1591] The server sends the scoring results and detailed feedback comments in JSON format to the user's device. The server records the result transmission log and checks the success or failure status.

[1592] Step 8:

[1593] The device receives the JSON data, parses it, and displays it in the user interface, where the user can visually check the score details and feedback from the generating AI.

[1594] Step 9:

[1595] The server structures the scoring results and analysis and records them in a database. The recorded data is backed up regularly and stored as long-term evaluation data.

[1596] Step 10:

[1597] The server generates reports based on the accumulated evaluation data, analyzing long-term evaluation trends and other trends, and provides these reports to administrators to help improve the quality of corporate meetings.

[1598] This enables the meeting evaluation system to improve the quality of meetings and provide fair evaluations and appropriate feedback quickly.

[1599] Example 1

[1600] 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."

[1601] Conventional meeting evaluation systems often manually record and analyze meeting audio, making it difficult to effectively evaluate meeting quality. They also lacked a means to objectively evaluate participants' leadership and cooperation, or to automatically detect harassment. As a result, it was difficult to provide feedback to improve meeting quality, and risk management across the company was inadequate.

[1602] 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.

[1603] In this invention, the server includes a means for dividing the recorded conference audio into data packets and transmitting them to the server, a means for converting the received audio data into text data, a means for inputting the converted text data into a generative AI model for analysis and scoring the speech content of the conference participants, a means for notifying the user terminal of the scoring results, and a means for long-term storage of the analysis results and accumulating them as evaluation data. This enables effective evaluation of the quality of the conference and provides detailed feedback to participants. It also automates the evaluation of leadership and cooperation and the monitoring of harassment, strengthening risk management in companies.

[1604] "Conference audio" refers to all audio data spoken during a conference.

[1605] "Recording means" refers to the equipment or method for recording conference audio in real time and converting it into digital data.

[1606] A "data packet" refers to a data unit that divides a series of audio data into a certain size and transmits it over a network.

[1607] "Server" refers to a computing system that processes data received from a terminal, generates analysis results, and manages them.

[1608] "Audio data" refers to data in which conference audio has been converted into digital format.

[1609] "Text data" refers to character string data converted from voice data using voice recognition technology.

[1610] A "generative AI model" refers to an artificial intelligence model that performs natural language processing and analysis based on large amounts of data.

[1611] "Analyzing" refers to the process of using certain algorithms or models to make sense of input data and extract relevant information.

[1612] "Content of remarks" refers to the general content of remarks made by conference participants during the conference.

[1613] "Scoring" refers to assigning points based on analysis results according to certain evaluation criteria.

[1614] "User terminal" refers to a device used by a conference participant to record voice and display the analysis results.

[1615] "Means for notifying" refers to a method or device for transmitting information from a server to a user terminal and informing the user.

[1616] "Storage means" refers to methods and devices for retaining data for long periods of time.

[1617] "Evaluation data" refers to data including analysis results and scoring results of meetings.

[1618] "Detailed feedback" refers to specific and analytical evaluations and information on areas for improvement provided to meeting participants.

[1619] "Leadership" refers to the criteria used to evaluate meeting participants' ability to lead and provide direction.

[1620] "Collaborative ability" refers to a criterion that evaluates meeting participants' ability to cooperate and collaborate as team members.

[1621] "Harassment" refers to inappropriate comments or behavior that may occur during a meeting.

[1622] This invention relates to a meeting evaluation system for improving the quality of meetings. This system records and analyzes meeting audio in real time and scores the remarks made by meeting participants to monitor harassment and evaluate facilitation skills. The detailed configuration and operation of the system are shown below.

[1623] System configuration

[1624] The system consists of the following main components:

[1625] 1. User device (e.g. PC, smartphone, tablet, etc.)

[1626] 2. Server (a computer system that receives and analyzes voice data)

[1627] 3. Network (communications infrastructure connecting user devices and servers)

[1628] Operational Overview

[1629] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted to text data on the server side and further analyzed using a generative AI model. The analysis results are scored based on the content of the speech, and various indicators (such as leadership, cooperation, and the presence or absence of harassment) are evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided.

[1630] Specific operation example

[1631] A specific example of the operation of this system is shown below.

[1632] Meeting Scenarios

[1633] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[1634] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[1635] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[1636] 3. The server arranges and concatenates the received voice data in order and converts it into text data using a voice recognition engine (e.g., Google Cloud Speech-to-Text).

[1637] 4. The converted text data is input into a generative AI (e.g., GPT-4 model) to perform a detailed analysis of the meeting content.

[1638] 5. Based on the analysis results, the server scores participants on their leadership, cooperativeness, and whether or not they engage in harassment.

[1639] For example, user A's leadership score is 85, user B's contribution score is 70, cooperativeness score is 90, and no harassment is detected.

[1640] 6. The server sends the scoring results to the user's device and displays the results visually through the user interface.

[1641] 7. The device will provide detailed score information and feedback comments from the generated AI, allowing users to review their results.

[1642] 8. The server records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[1643] Example of input prompt for generative AI model

[1644] Example prompt:

[1645] Analyze the audio recording of the meeting below and rate each participant's leadership, cooperation, and whether or not they engaged in harassment.

[1646] text:

[1647] User A: Today's meeting is about the progress of the new project. Everyone, please give us your opinions.

[1648] User B: I'll be in charge of the progress report. Currently, Task A is 80% complete.

[1649] ...

[1650] In this way, by showing detailed embodiments of the invention, it is possible to effectively evaluate the quality of meetings and provide clear feedback to participants through scoring of leadership and cooperation, which is expected to improve meeting management and strengthen corporate risk management.

[1651] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1652] Step 1:

[1653] A user starts a meeting and launches the recording app.

[1654] Specifically, when a meeting starts, the user launches a dedicated recording application on their device.

[1655] Specific operation: For example, User A double-clicks the meeting recording app on his / her PC and clicks the "Start Recording" button.

[1656] Input: Conference audio (analog audio)

[1657] Output: Launched recording application

[1658] Step 2:

[1659] The terminal captures the conference audio and divides it into data packets in real time.

[1660] Specific description: The terminal records the conference audio through a microphone and performs the process of dividing it into small data packets.

[1661] What it does: The PC's microphone captures audio data every second, and then divides each into data packets every 0.5 seconds.

[1662] Input: Conference audio (analog audio)

[1663] Output: Data packet (digital data)

[1664] Step 3:

[1665] The terminal sequentially transmits voice data packets to the server.

[1666] Specific explanation: The terminal sequentially transmits the divided voice data packets to the server.

[1667] What it does: Your PC sends packetized voice data to a server every second via your internet connection.

[1668] Input: Data packets (digital voice data)

[1669] Output: Data packet sent to the server

[1670] Step 4:

[1671] The server receives the audio data packets and concatenates them to form an audio stream.

[1672] Specifically, the server concatenates the received audio data packets in order to create a continuous audio stream.

[1673] Specific operation: The server combines the audio data packets received every 0.5 seconds to restore the original continuous audio data.

[1674] Input: Data packets (digital voice data)

[1675] Output: Concatenated audio stream (continuous audio data)

[1676] Step 5:

[1677] The server uses a speech recognition engine to convert the speech into text.

[1678] Specifically, the server uses a speech recognition engine (e.g., Google Cloud Speech-to-Text) to convert the voice data into text data.

[1679] What happens: The server sends the audio file to a speech recognition API and retrieves the conversation in text format.

[1680] Input: Concatenated audio stream (continuous audio data)

[1681] Output: Text data (string format)

[1682] Step 6:

[1683] The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[1684] Specific explanation: The server inputs text data into a generative AI (e.g., GPT-4 model) to analyze the meeting content and score what is said.

[1685] Specific behavior: The server passes the following prompt to the generation AI:

[1686] Example prompt:

[1687] Analyze the audio recording of the meeting below and rate each participant's leadership, cooperation, and whether or not they engaged in harassment.

[1688] text:

[1689] User A: Today's meeting is about the progress of the new project. Everyone, please give us your opinions.

[1690] User B: I'll be in charge of the progress report. Currently, Task A is 80% complete.

[1691] ...

[1692] Input: Text data (string format)

[1693] Output: Analysis results and scoring data (evaluation index)

[1694] Step 7:

[1695] The server sends the analysis results to the terminal and displays them on the user interface.

[1696] Specific explanation: The server sends the analysis results returned by the generation AI to the user terminal and displays them on the user interface.

[1697] Specific operation: The server sends the evaluation results (e.g., User A's leadership score is 85 points, User B's contribution score is 70 points) to the devices of User A and User B, and visualizes them on a dedicated dashboard.

[1698] Input: Analysis results and scoring data (evaluation index)

[1699] Output: Evaluation results displayed on the user's terminal

[1700] Step 8:

[1701] The server records the results in a database and accumulates them as long-term evaluation data.

[1702] Specific explanation: The server stores the analysis results and each meeting data in a database and manages them as long-term evaluation data.

[1703] Specific operation: The server records the evaluation results and speech analysis data in a MySQL database for future comparative analysis.

[1704] Input: Analysis results and scoring data (evaluation index)

[1705] Output: Evaluation data stored in a database

[1706] This system improves the quality of meetings and enables fair evaluations. It supports productive meeting management while also providing a harassment monitoring function, strengthening risk management across the company.

[1707] (Application example 1)

[1708] 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."

[1709] Conventional meeting evaluation systems can evaluate the quality of meetings and the behavior of participants in real time, but they are unable to evaluate in-car communication or safety awareness. Furthermore, there is a lack of systems that can analyze in-car conversations in detail to evaluate the driver's attention and whether or not harassment has occurred. This has resulted in insufficient prevention of harassment and improvement of safety awareness in the in-car environment.

[1710] 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.

[1711] In this invention, the server includes means for recording conference audio in real time, means for dividing the recorded conference audio into data packets and transmitting them to the server, means for converting the received audio data into text data, means for analyzing the text data and scoring the speech content of the conference participants, means for recording in-car conversations in real time, means for scoring the driver's attention, safety awareness, and presence or absence of harassment through analysis, means for notifying the scoring results to a user terminal, and means for long-term storage of the analysis results and accumulating them as evaluation data. This makes it possible to evaluate in-car communication quality, safety awareness, and presence or absence of harassment in detail. This also makes it possible to appropriately monitor the behavior of drivers and passengers and improve the safety and comfort of the in-car environment.

[1712] "Conference audio" refers to the audio data of the participants' comments and interactions that occur during a conference.

[1713] A "data packet" is a small unit of digital data divided into pieces and formatted for efficient transmission over a network.

[1714] A "server" is a central computer system for collecting, analyzing, and storing data.

[1715] "Text data" refers to text information converted from speech using speech recognition technology.

[1716] "Scoring" is an evaluation method in which points are assigned according to specific evaluation indicators based on analyzed data.

[1717] "User terminal" refers to a device that is directly operated by a user, such as a PC, smartphone, or tablet.

[1718] "Real-time" refers to data being processed and transferred immediately with little delay.

[1719] "Attention" is an indicator of how much a driver is concentrating on traffic conditions and driving operations.

[1720] "Harassment" refers to inappropriate words, actions or behaviour towards other crew members.

[1721] "Analysis" is the process of analyzing collected data in detail to clarify its meaning and trends.

[1722] "Notification" refers to the act of informing the user of the analysis results or scoring results.

[1723] "Evaluation data" refers to data that includes analysis results and scoring results, and is used as the basis for long-term evaluation and improvement.

[1724] This invention relates to an in-car communication evaluation system. The purpose is to improve the safety and comfort of the in-car environment by recording and analyzing in-car conversations in real time and scoring passenger attention, safety awareness, and the presence or absence of harassment. This system consists of the following main components:

[1725] System configuration

[1726] 1. User Device

[1727] This refers to in-vehicle systems, smartphones, tablets, etc. that have recording functions and are responsible for recording conversations in real time.

[1728] 2. Server

[1729] It is the central computer system that collects and analyzes data and stores the results.

[1730] 3. Network

[1731] It is a communications infrastructure that connects user terminals and servers, and uses wireless communication technology.

[1732] Specific operation example

[1733] A specific example of the operation of this system is shown below.

[1734] In-car communication scenario

[1735] Consider a situation in which user A (driver) and user B (passenger) are having a conversation in a car.

[1736] 1. User terminal: The vehicle's in-vehicle system activates the recording function and presses the "Start Recording" button.

[1737] 2. Terminal: Records conversations inside the car in real time, divides the voice data into data packets, and sends them to the server sequentially.

[1738] 3. Server: The received voice data is arranged in order and concatenated, and the voice data is converted into text data using a voice recognition engine (e.g., speech_recognition library).

[1739] 4. Server: The converted text data is input into a generative AI model (e.g., GPT-3) for detailed analysis.

[1740] 5. Server: Based on the analysis results, the server scores participants on their attention, safety awareness, and whether or not they have engaged in harassment.

[1741] For example, user A's attention score is 85, user B's safety awareness score is 90, and no harassment is detected.

[1742] 6. Server: Sends the scoring results to the user's terminal and displays the results visually through the user interface.

[1743] 7. Device: Provides detailed score information and feedback comments from the generated AI, allowing users to check their results.

[1744] 8. Server: Records the scoring results and analysis results in a database and accumulates them as long-term evaluation data.

[1745] Hardware and software used

[1746] Speech recognition engine: speech_recognition library

[1747] Generative AI model: GPT-3

[1748] User devices: in-vehicle systems, smartphones, tablets, etc.

[1749] Server: a computer system for data analysis and storage

[1750] Network: Wireless communication technology (e.g., Wi-Fi, 4G, 5G)

[1751] Adding specific examples

[1752] Driving conversation scenario:

[1753] Expect conversations such as, "Are you concentrating on driving?", "Drive carefully," and "We're not in a hurry, so drive safely."

[1754] Example prompts to input to a generative AI model:

[1755] Based on the conversation below, please rate the level of attention, safety awareness, and whether or not there was any harassment.

[1756] example:

[1757] How was your drive today?

[1758] Drive carefully. We're not in a hurry, so safety comes first.

[1759] I think you are being harsh towards me.

[1760] The present invention thereby provides a means for evaluating in-vehicle communications in real time and for appropriately monitoring occupant behavior.

[1761] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1762] Step 1:

[1763] The user terminal activates the recording function of the vehicle's in-vehicle system and operates the "start recording" button.

[1764] Input: Start recording

[1765] Output: Real-time recording begins

[1766] Specific operation: The user starts recording on the in-car system or smartphone. The device prepares to record the conversation in real time.

[1767] Step 2:

[1768] The terminal records conversations in the car in real time, divides the voice data into data packets, and transmits them to the server one by one.

[1769] Input: Voice data of conversations in a car

[1770] Output: sent as data packets to the server

[1771] Specific operation: The device divides the recorded audio into small data packets and sends them to the server over the network.

[1772] Step 3:

[1773] The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine.

[1774] Input: Packetized voice data

[1775] Output: Converted text data

[1776] Specific operation: The server uses a speech recognition library (e.g., speech_recognition) to analyze the voice data and convert it into text data.

[1777] Step 4:

[1778] The server inputs the converted text data into a generative AI model (e.g., GPT-3) for detailed analysis.

[1779] Input: Text data

[1780] Output: Analysis results (e.g., scores for attention, safety awareness, and whether or not there was harassment)

[1781] Specific operation: The server inputs prompts into the generative AI model, which then performs detailed analysis and scoring based on the conversation content.

[1782] Step 5:

[1783] Based on the analysis results, the server scores the person's attention, safety awareness, and whether or not they have engaged in harassment.

[1784] Input: Analysis results from a generative AI model

[1785] Output: Each score (e.g., attention 85 points, safety awareness 90 points, no harassment detected)

[1786] Specific operation: The server quantifies specific indicators as scores based on the output of the generation AI.

[1787] Step 6:

[1788] The server transmits the scoring results to the user terminal and visually displays the results through a user interface.

[1789] Input: Scoring results

[1790] Output: Visual feedback displayed in the user interface

[1791] Specific operation: The server sends the score and analysis results to the user's device, which then displays the results on the screen.

[1792] Step 7:

[1793] The device provides the user with detailed score information and feedback comments from the generated AI for review.

[1794] Input: Scoring results received from the server

[1795] Output: Feedback comments that the user sees

[1796] Specific operation: The terminal displays the scoring results and feedback comments for the user to review.

[1797] Step 8:

[1798] The server records the scoring results and analysis details in a database and accumulates them as long-term evaluation data.

[1799] Input: Scoring results and analysis details

[1800] Output: Evaluation data stored in a database

[1801] Specific operation: The server records all analysis results in a database and stores the data for future evaluation and improvement.

[1802] 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.

[1803] The present invention relates to a meeting evaluation system for improving the quality of meetings. This system not only records and analyzes meeting audio in real time, scoring the content of participants' remarks, but also monitors harassment and evaluates facilitation skills by recognizing participants' emotions.

[1804] System configuration

[1805] The system consists of the following main components:

[1806] 1. User device (PC, smartphone, tablet, etc.)

[1807] 2. Server (a computer system that receives and analyzes voice data)

[1808] 3. Network (communications infrastructure connecting user devices and servers)

[1809] 4. Emotion engine (a module that recognizes the emotions of meeting participants)

[1810] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The emotion engine also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored over the long term and accumulated as evaluation data.

[1811] Specific operation example

[1812] A specific example of the operation of this system is shown below.

[1813] Meeting Scenarios

[1814] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[1815] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[1816] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[1817] 3. The server arranges and concatenates the received voice data in order, and converts the voice data into text data using a voice recognition engine (e.g., voice recognition API).

[1818] 4. The converted text data is input into a generative AI (e.g., a generative AI model) to perform a detailed analysis of the meeting content.

[1819] 5. Based on the analysis results, the server scores participants on their leadership, cooperation, and harassment.

[1820] For example, user A's leadership score is 85 points, user B's contribution score is 70 points, cooperativeness score is 90 points, and no harassment is detected.

[1821] 6. The emotion engine recognizes the participant's emotions from the voice data and sends the data to the server.

[1822] 7. The server adjusts the scoring results based on the recognized emotion data and generates detailed feedback.

[1823] For example, if user A feels stressed during a meeting, feedback is provided based on that emotion.

[1824] 8. The server sends the scoring results and emotion data to the user devices of User A and User B, and visually displays the results through the user interface.

[1825] 9. The device will provide detailed scores, feedback comments from the generative AI, and an evaluation from the emotion engine, allowing users to review their results.

[1826] 10. The server records the scoring results, analysis results, and emotional data in a database and accumulates them as long-term evaluation data.

[1827] Explanation of program processing

[1828] 1. A user starts a meeting and launches the recording app.

[1829] 2. The terminal captures the conference audio and divides it into data packets in real time.

[1830] 3. The terminal sequentially transmits voice data packets to the server.

[1831] 4. The server receives the audio data packets and concatenates them to form an audio stream.

[1832] 5. The server uses a speech recognition engine to convert the speech to text.

[1833] 6. The server inputs the converted text data into the generation AI for detailed analysis and scoring.

[1834] 7. The server sends the analysis results to the terminal and displays them on the user interface.

[1835] 8. The server uses the emotion engine to recognize the emotion data and adjust the scoring results.

[1836] 9. The server sends the adjusted results to the terminal and notifies the user.

[1837] 10. The server records the results in a database and stores them as long-term data.

[1838] Through these processes, the system improves the quality of meetings and provides fair evaluations and appropriate feedback quickly.In addition, the introduction of an emotion engine provides detailed feedback based on the user's emotional state, resulting in a better meeting experience.

[1839] The processing flow will be explained below.

[1840] Step 1:

[1841] A user starts a conference and launches the recording application on their device. When the user presses the "Start Recording" button, the device captures the conference audio through the audio input device (microphone).

[1842] Step 2:

[1843] The terminal divides the recorded conference audio into data packets in real time. The terminal divides the captured audio data into small packets with a certain byte size and transmits them sequentially to the server via the network.

[1844] Step 3:

[1845] The server reorders the received audio data packets and concatenates them into a stream. The server checks the data integrity and requests packet retransmission if necessary.

[1846] Step 4:

[1847] The server inputs the formatted audio stream into a speech recognition engine (e.g., speech recognition API) and converts the audio data into text data. The server receives the converted text data from the speech recognition engine and temporarily stores it in a database.

[1848] Step 5:

[1849] The server inputs the text data into a generative AI (e.g., a generative AI model) and instructs the generative AI to "analyze the content of the meeting based on this text data and score it on indicators such as leadership, cooperation, and the presence or absence of harassment."

[1850] Step 6:

[1851] The server receives the analysis results output by the generation AI and extracts scores and detailed feedback for each indicator. For example, User A's leadership score might be 85 points, User B's contribution score 70 points, cooperation score 90 points, and no harassment detected.

[1852] Step 7:

[1853] The emotion engine recognizes the emotions of conference participants from the voice data. The server provides the voice data to the emotion engine, which analyzes the emotional state. For example, emotions such as anger, anxiety, and joy are recognized.

[1854] Step 8:

[1855] The server receives the recognized emotion data from the emotion engine and adjusts the scoring results, such as affecting leadership and cooperation scores based on the emotion data, or adding warnings if there is a high likelihood of harassment.

[1856] Step 9:

[1857] The server sends the adjusted scoring results and detailed emotional feedback in JSON format to the user's device. The server records the result transmission log and checks the success or failure status.

[1858] Step 10:

[1859] The device receives the JSON data, parses it, and displays it on the user interface, where the user can visually check the score details, feedback from the AI, and the analysis of the user's emotional state.

[1860] Step 11:

[1861] The server structures the scoring results, analysis details, and emotion data and records them in a database. The recorded data is backed up regularly and stored as long-term evaluation data.

[1862] Step 12:

[1863] The server generates reports that analyze long-term evaluation trends and trends based on the accumulated evaluation data and emotion data. The server provides these reports to administrators, helping to improve the quality of corporate meetings.

[1864] This system improves the quality of meetings, providing fair evaluations and prompt, appropriate feedback. Furthermore, the introduction of an emotion engine provides detailed feedback based on the user's emotional state, enabling a better meeting experience.

[1865] Example 2

[1866] 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."

[1867] Accurately evaluating the content and quality of comments made in meetings and providing feedback to meeting participants is important for improving the performance of an entire organization. However, conventional meeting evaluation systems have difficulty evaluating not only the content of participants' comments but also their emotions, and it is not possible to accumulate this data over the long term and perform trend analysis. Therefore, a system that comprehensively evaluates the quality of meetings and provides detailed feedback is needed.

[1868] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for recording the conference audio in real time, a means for dividing the recorded conference audio into data packets and transmitting them to the server, a means for converting the audio data received on the server side into text data, a means for analyzing the text data and scoring the remarks of the conference participants, a means for notifying the scoring results to the user terminal, a means for long-term saving the analysis results and accumulating them as evaluation data, and a means for recognizing the emotions of the conference participants and adjusting the scoring results. This makes it possible to comprehensively evaluate the quality of the conference in real time and provide fair feedback.

[1869] "Conference audio" refers to audio data of conversations and presentations that occur during a conference.

[1870] "Real-time recording means" refers to a device or method that instantly records conference audio and stores it as digital data.

[1871] "Data packet" refers to the unit in which digital data is divided into small pieces for efficient transmission over a network.

[1872] "Means for transmitting to the server" refers to the technology or method for transmitting data packets from the user terminal to the server.

[1873] "Means for converting into text data" refers to speech recognition technology or algorithms that convert voice data into text information.

[1874] "Means for analyzing text data" refers to programs or processing methods for analyzing data converted into text and understanding its content.

[1875] "Means for scoring the content of statements made by meeting participants" refers to algorithms or methods for evaluating and quantifying the content of statements.

[1876] "Means for notifying the user terminal of the scoring results" refers to a method or system for notifying the user of the analysis and evaluation results.

[1877] "Means for storing analysis results for the long term and accumulating them as evaluation data" refers to technology that stores analysis results in a database or the like for a long period of time and uses them for future analysis and evaluation.

[1878] "Means for recognizing the emotions of meeting participants" refers to technologies and methods for identifying the emotional state of meeting participants from voice or text data.

[1879] "Means for adjusting scoring results" refers to algorithms or methods for correcting or amending initial scoring results based on recognized emotion data.

[1880] The present invention relates to a meeting evaluation system for improving the quality of meetings. This system not only records and analyzes meeting audio in real time and scores the content of participants' comments, but also monitors harassment and evaluates facilitation ability by recognizing participants' emotions.

[1881] System configuration

[1882] The system consists of the following main components:

[1883] 1. User device (PC, smartphone, tablet, etc.)

[1884] 2. Server (a computer system that receives and analyzes voice data)

[1885] 3. Network (communications infrastructure connecting user devices and servers)

[1886] 4. Emotion engine (a module that recognizes the emotions of meeting participants)

[1887] 5. Database (a storage device for long-term storage of analysis results)

[1888] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and further analyzed using generative AI. The emotion engine also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and each indicator (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) is evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored over the long term and accumulated as evaluation data.

[1889] Specific operation example

[1890] A specific example of the operation of this system is shown below.

[1891] Meeting Scenarios

[1892] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[1893] 1. User A launches the recording app on their device and presses the "Start Recording" button.

[1894] 2. The terminal records the conference audio in real time, divides the audio data into packets, and sends them to the server sequentially.

[1895] 3. The server arranges and concatenates the received voice data in order and converts it into text data using a speech recognition engine (e.g., Google Cloud Speech-to-Text API).

[1896] 4. The converted text data is input into a generative AI (e.g., OpenAI GPT-4) to perform a detailed analysis of the meeting content.

[1897] 5. Based on the analysis results, the server scores participants on their leadership, cooperation, and harassment.

[1898] For example, user A's leadership score is 85 points, user B's contribution score is 70 points, cooperativeness score is 90 points, and no harassment is detected.

[1899] 6. The emotion engine recognizes the participant's emotions from the voice data and sends the data to the server.

[1900] 7. The server adjusts the scoring results based on the recognized emotion data and generates detailed feedback.

[1901] For example, if user A feels stressed during a meeting, feedback is provided based on that emotion.

[1902] 8. The server sends the scoring results and emotion data to the user devices of User A and User B, and visually displays the results through the user interface.

[1903] 9. The device will provide detailed scores, feedback comments from the generative AI, and an evaluation from the emotion engine, allowing users to review their results.

[1904] 10. The server records the scoring results, analysis results, and emotional data in a database and accumulates them as long-term evaluation data.

[1905] Prompt Sentence Examples

[1906] Here are some example prompts to input to a generative AI model (e.g., GPT-4):

[1907] The audio from a meeting was transcribed as follows:

[1908] 1. User A: "How is the project going?"

[1909] 2. User B: "It's going well, but there are some challenges."

[1910] 3. User A: "Do you need any help?"

[1911] Please rate User A's leadership, cooperation, and overall performance on a numerical scale. Also, please rate User B's contribution.

[1912] This allows the system to evaluate the quality of the meeting in real time and provide fair feedback. Furthermore, by incorporating an emotion engine, the system can provide detailed feedback based on the user's emotional state, resulting in a better meeting experience.

[1913] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1914] Step 1:

[1915] A user starts a meeting and launches the recording app.

[1916] Specific actions: User A opens the recording app on their smartphone and presses the "Start Recording" button.

[1917] Input: User action at the start of a meeting.

[1918] Output: Start recording.

[1919] Step 2:

[1920] The terminal captures the conference audio and divides it into data packets in real time.

[1921] How it works: The smartphone picks up the audio from the meeting and packets the audio data every second.

[1922] Input: Conference audio.

[1923] Output: Audio data packets.

[1924] Step 3:

[1925] The terminal sequentially transmits voice data packets to the server.

[1926] How it works: A smartphone sends voice data packets over the internet to a cloud server, minimizing latency through parallel processing.

[1927] Input: Audio data packets.

[1928] Output: Sending data packets to the server.

[1929] Step 4:

[1930] The server receives the audio data packets and concatenates them to form an audio stream.

[1931] What it does: The server reorders the audio data packets it receives and reconstructs them into a continuous audio stream.

[1932] Input: Audio data packets.

[1933] Output: Audio stream.

[1934] Step 5:

[1935] The server converts the speech to text using a speech recognition engine (e.g., a speech recognition API).

[1936] What happens: The server inputs the audio stream into the Google Cloud Speech-to-Text API and generates text data.

[1937] Input: Audio stream.

[1938] Output: Text data.

[1939] Step 6:

[1940] The server inputs the converted text data into a generative AI (e.g., a generative AI model) for detailed analysis and scoring.

[1941] Specific operation: The server sends the text data to a generation AI (e.g., GPT-4), which analyzes the content of the meeting and quantifies leadership, cooperation, and the presence or absence of harassment.

[1942] Input: Text data.

[1943] Output: Scoring results.

[1944] Step 7:

[1945] The server sends the analysis results to the terminal and displays them on the user interface.

[1946] Specific operation: The server sends the scoring results generated to the user's smartphone, and the score is displayed on the app.

[1947] Input: Scoring results.

[1948] Output: Notification and display on smartphone.

[1949] Step 8:

[1950] The server uses an emotion engine to recognize the emotion data and adjust the scoring results.

[1951] What it does: The server inputs the voice data into an emotion engine (e.g., Emotion API) to recognize the participant's emotional state, and fine-tunes the scoring results based on that.

[1952] Input: Audio data.

[1953] Output: Recognized emotions, adjusted scoring results.

[1954] Step 9:

[1955] The server sends the adjusted results to the terminal and notifies the user.

[1956] Specific operation: The server sends the adjusted scoring results and emotional data to the smartphone and displays detailed feedback.

[1957] Input: Calibrated scoring results and sentiment data.

[1958] Output: Notifications and detailed feedback displayed on your smartphone.

[1959] Step 10:

[1960] The server records the results in a database for long-term storage.

[1961] Specific operation: The server stores the scoring results and emotion data in a database and accumulates them as future evaluation data.

[1962] Input: Scoring results and sentiment data.

[1963] Output: Recorded and accumulated evaluation data in a database.

[1964] (Application example 2)

[1965] 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."

[1966] Conventional meeting evaluation systems can score the content of participants' comments, but they do not perform more detailed analysis, such as detecting the emotions and harassment contained in the comments, or evaluating cooperation. Furthermore, since there is no feedback that takes into account the emotional elements that arise during meetings, there is a problem in that measures to improve the quality of meetings cannot be taken sufficiently. Furthermore, there is a lack of analysis of evaluation trends through long-term data analysis, so there is a lack of indicators for continuously improving meeting outcomes.

[1967] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for recording conference audio in real time, means for dividing the recorded conference audio into data packets and transmitting them to a computer system, means for converting the audio data received by the computer system into text data, means for analyzing the text data and scoring the speech content of the conference participants, means for notifying the scoring results to an information display terminal, means for long-term saving of the analysis results and accumulating them as evaluation data, and means for recognizing the emotions of the conference participants using an emotion recognition module and adjusting the scoring results based on the emotion data. This enables detailed analysis and feedback to improve the quality of the conference, and enables sustainable measures to improve the outcome of the conference through adjustment of scoring results taking into account the emotional states of the conference participants and accumulation and analysis of long-term evaluation data.

[1968] A "means for recording conference audio in real time" is a device that has the function of instantly recording audio spoken during a conference.

[1969] "Means for dividing the recorded conference audio into data packets and transmitting them to a computer system" refers to a device or software that has the function of dividing the recorded audio data into packets of a certain size and transmitting them to a server via a communication network.

[1970] "Means for converting voice data received by a computer system into text data" refers to the process of converting voice data received by a server into text information using voice recognition technology.

[1971] The "means for analyzing text data and scoring the content of statements made by meeting participants" is a system that has the function of analyzing converted text data and calculating various evaluation indicators based on the content of statements made by meeting participants.

[1972] "Means for notifying the information display terminal of the scoring results" refers to a mechanism for transmitting the scoring information obtained as an analysis result to the device used by the user and displaying or notifying the user.

[1973] "Means for storing analysis results over the long term and accumulating them as evaluation data" refers to a system that stores the generated analysis results and scores in a database and maintains that data for a long period of time.

[1974] "Means for recognizing the emotions of meeting participants using an emotion recognition module and adjusting the scoring results based on that emotion data" is a component that has the function of recognizing emotions from the voices and facial expressions of meeting participants, and appropriately correcting the scoring results to reflect their emotional state.

[1975] The "means for scoring leadership, cooperation, and the presence or absence of harassment" is a system that has the function of evaluating the strength of leadership, the degree of cooperation, and whether harassment is present based on the comments and actions of meeting participants.

[1976] The "means for providing detailed feedback" is a mechanism for providing specific and useful feedback to the user based on the scoring results and emotional data.

[1977] "Means of analyzing long-term evaluation tendencies and trends based on accumulated data" refers to the process of analyzing stored historical data to identify evaluation tendencies and trends that emerge over time.

[1978] "Means that take into account the correlation with emotional data" refers to a method of analyzing emotional data in combination with other evaluation indicators to verify the causal relationship and influence between emotions and behavior.

[1979] The present invention relates to a conference evaluation system for improving the quality of conferences, and includes means for recording conference audio in real time, means for dividing the recorded audio into data packets and transmitting them to a computer system, means for converting the audio data received by the computer system into text data, means for analyzing the text data and scoring the content of comments made by conference participants, means for notifying an information display terminal of the scoring results, means for long-term storage of the analysis results and accumulation as evaluation data, and means for recognizing the emotions of conference participants using an emotion recognition module and adjusting the scoring results based on that emotion data. This system makes it possible to improve the quality of conferences.

[1980] System configuration

[1981] The system consists of the following main components:

[1982] 1. User device (PC, smartphone, tablet, etc.)

[1983] 2. Server (computer system that receives and analyzes voice data)

[1984] 3. Network (communications infrastructure connecting user devices and servers)

[1985] 4. Emotion Recognition Module (Module that recognizes the emotions of meeting participants)

[1986] The user device records the audio of the meeting in real time, divides it into data packets, and sends them to the server. This audio data is converted into text data on the server side and analyzed using generative AI. An emotion recognition module also recognizes the emotions of the participants from the audio data and adjusts the scoring results based on those emotions. The analysis results are scored based on the content of the speech of the meeting participants, and various indicators (leadership, cooperation, presence or absence of harassment, emotional evaluation, etc.) are evaluated. Finally, these scoring results are notified to the user device, and detailed feedback is provided. The analysis results and emotional data are also stored long-term and accumulated as evaluation data.

[1987] System Operation

[1988] A specific example of the operation of this system is shown below.

[1989] Meeting Scenarios

[1990] Consider a situation where User A (manager) and User B (subordinate) are having a one-on-one meeting.

[1991] 1. User A starts the recording application on their device and records the conference audio in real time.

[1992] 2. The user terminal divides the recorded voice into data packets and sends them to the server over the network.

[1993] 3. The server uses a speech recognition engine (e.g., Google Speech-to-Text API) to convert the received voice data into text data.

[1994] 4. The converted text data is analyzed by a generative AI (e.g., a generative AI model such as GPT-4).

[1995] 5. Based on the analysis results, the server scores the comments made by the conference participants and notifies the user terminal of the scoring results.

[1996] 6. The emotion recognition module recognizes the emotions of meeting participants from the voice data and adjusts the scoring results based on this emotion data.

[1997] 7. The server notifies the user device of the adjusted scoring results and detailed feedback.

[1998] 8. Analysis results and emotional data will be stored for a long period of time and accumulated on the server as evaluation data.

[1999] 9. The accumulated data will be analyzed to help improve the quality of future meetings, and long-term evaluation and trend analysis will be conducted.

[2000] Specific examples of hardware and software used

[2001] User devices: PC, smartphone, tablet

[2002] Server: High-performance cloud server

[2003] Network: Internet, LAN

[2004] Speech recognition engine: Google Speech-to-Text API

[2005] Generative AI models: GPT-4 and similar models

[2006] Emotion Recognition Module: Dedicated emotion analysis software

[2007] Prompt Sentence Examples

[2008] "Please analyze this text and rate it on points like leadership, collaboration, and customer service. I'd also like you to do a sentiment analysis."

[2009] This will improve the quality of meetings and enable fair evaluations and prompt provision of appropriate feedback. Furthermore, the introduction of an emotion recognition module will provide detailed feedback based on the user's emotional state, enabling a better meeting experience.

[2010] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[2011] Step 1:

[2012] A user starts a meeting and launches a recording application on their device. Recording of the meeting audio begins. The input is the user's voice, and the output is the recorded audio data. Specifically, the microphone on the user's device picks up the audio, and the recording software records it.

[2013] Step 2:

[2014] The device divides the recorded voice into data packets in real time and sends them to the server via the network. The input is the recorded voice data, and the output is the voice data divided into data packets. Specifically, the recording application on the user device divides the recorded data into small data packets and sends them to the server via the Internet.

[2015] Step 3:

[2016] The server concatenates the received audio data to form a single audio data stream. The input is audio data divided into data packets, and the output is the concatenated audio data. Specifically, the server program reassembles the arriving data packets into the correct order and saves them as a stream of audio data.

[2017] Step 4:

[2018] The server converts the concatenated voice data into text data using a speech recognition engine (e.g., Google Speech-to-Text API). The input is the concatenated voice data, and the output is text data. Specifically, the speech recognition engine analyzes the voice samples and generates corresponding text.

[2019] Step 5:

[2020] The server inputs the text data into a generative AI model (e.g., GPT-4) for detailed analysis. The input is text data, and the output is analysis results and scoring data. An example prompt is, "Analyze this text and evaluate points such as leadership, cooperation, and customer responsiveness. Please also perform sentiment analysis." Specifically, the generative AI model analyzes the text data and generates a score based on each indicator.

[2021] Step 6:

[2022] The emotion recognition module recognizes the emotions of conference participants from the voice data and generates emotion data. The input is voice data and the output is emotion data. Specifically, the emotion recognition software analyzes the tone and rate of the voice to determine emotional states such as stress, joy, or anger.

[2023] Step 7:

[2024] The server adjusts the initial scoring result based on the emotion data. The input is the initial scoring result and emotion data, and the output is the adjusted scoring result. Specifically, the scoring algorithm incorporates the emotion data and recalculates the scoring result.

[2025] Step 8:

[2026] The server sends the adjusted scoring results and detailed feedback to the user terminal. The input is the adjusted scoring results and feedback data, and the output is a notification to the user terminal. Specifically, the server sends the generated detailed feedback comments to the user terminal and displays them through the application interface.

[2027] Step 9:

[2028] The analysis results and emotion data are stored over the long term and accumulated on the server as evaluation data. The input is the analysis results and emotion data, and the output is data stored in the database. Specifically, the server's database management system efficiently records this data and makes it available for future trend analysis and evaluation.

[2029] Step 10:

[2030] The server analyzes long-term evaluation trends and trends based on the accumulated data and generates reports. The input is the long-term stored evaluation data, and the output is the evaluation report. Specifically, data analysis software analyzes past data, extracts notable trends and patterns, and generates a report to be provided to the user.

[2031] 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.

[2032] 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.

[2033] 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.

[2034] 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.

[2035] 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.

[2036] 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.

[2037] 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).

[2038] 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.

[2039] 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."

[2040] 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.

[2041] 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).

[2042] 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.

[2043] 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.

[2044] 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.

[2045] 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.

[2046] 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 F...

Claims

1. A means for recording conference audio in real time; means for dividing the recorded conference audio into data packets and transmitting the packets to a server; A means for converting the received voice data into text data on the server side; A means for analyzing text data and scoring the content of statements made by meeting participants; A means for notifying a user terminal of the scoring result; A means to store the analysis results over the long term and accumulate them as evaluation data, A system including:

2. 2. The system according to claim 1, further comprising means for analyzing the content of statements made by conference participants, scoring them in terms of leadership, cooperation, and the presence or absence of harassment, and providing detailed feedback.

3. A means of recording the scoring results in a database and backing them up periodically; 10. The system of claim 1, further comprising means for generating reports based on the accumulated data for long-term evaluation and trend analysis.

Images