System

Abstract

A system is provided.SOLUTION: Means for collecting voice data for summarizing a main point and an important point of a discussion in an online conference in which a participant participates late and converting the voice data into text, means for analyzing text data converted from the voice data by natural language processing and extracting the main point and the important point of the discussion, means for collecting chat data, analyzing the chat data, and extracting an important remark, A system, comprising: means for identifying a conference mood; and means for providing a summary of the summarized discussion, key points, and conference mood to participants.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] The use of online meetings has increased sharply due to the impact of COVID-19, resulting in frequent overlapping meeting times. This has led to participants often joining meetings late, which takes time to understand the content of the meeting and prevents them from smoothly participating in the discussion. This problem reduces meeting productivity and increases stress for participants, so an effective solution is needed. [Means for solving the problem]

[0005] To solve this problem, the present invention provides the following means. A system is constructed that instantly summarizes the main points and important points of discussions, even when participants join an online conference late, and organizes and provides the meeting content in real time using voice data and chat data. Specifically, the system includes a means for collecting voice data and converting it into text using voice recognition technology. It also includes a means for extracting the main points and important points of discussions from the converted text data using natural language processing technology. It also includes a means for collecting chat data and analyzing it to extract important comments. It also includes a means for analyzing the voice data and chat data using sentiment analysis to identify the atmosphere of the meeting. The system then provides a means for providing participants with the summarized main points and important points of discussions, as well as the atmosphere of the meeting. This allows late participants to quickly and efficiently return to the meeting, improving productivity and reducing stress.

[0006] "Audio data" refers to digital audio files that record what participants say in an online conference.

[0007] "Text data" refers to character information obtained by converting voice data using voice recognition technology.

[0008] "Natural language processing" is a technology that allows computers to understand, analyze, and generate human language.

[0009] "Discussion points" are the important topics and conclusions discussed at a meeting.

[0010] "Key points" are items or key information that deserve special attention in the discussion.

[0011] "Timeline messages" are real-time messages exchanged between participants via text during an online meeting.

[0012] "Sentiment analysis" is a technology that mathematically analyzes emotional nuances from text or audio data and classifies emotions as positive, negative, neutral, etc.

[0013] "Meeting atmosphere" refers to the emotional state and overall psychological tone of the meeting participants.

[0014] "Summary information" is information that summarizes the main points and important points of the discussion in a concise manner.

[0015] "Real-time" refers to data acquisition and processing occurring almost immediately.

[0016] "Providing to participants" refers to the act of immediately showing the processing results to users who join late via a user interface.

[0017] A "natural language processing model" is a mathematical model or algorithm used to implement natural language processing techniques.

[0018] An "online meeting platform" is software or a service that allows multiple people to hold meetings over the Internet, such as Zoom, Meet, or Teams. [Brief explanation of the drawings]

[0019] [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

[0020] 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.

[0021] First, the terms used in the following description will be explained.

[0022] 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).

[0023] 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.

[0024] 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.

[0025] 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.

[0026] 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."

[0027] [First embodiment]

[0028] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0029] 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.

[0030] 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).

[0031] 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.

[0032] 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.

[0033] 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.

[0034] 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.

[0035] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0036] 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.

[0037] 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.

[0038] 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.

[0039] 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."

[0040] The present invention is a system that helps users who join an online conference late quickly understand the content of the discussion and participate in the conference smoothly, and processing is mainly carried out between three parties: a server, a terminal, and the user.

[0041] Overview of program processing

[0042] Audio data processing

[0043] At the start of an online meeting, the server acquires audio data through the conference platform's API. This audio data is streamed in real time and converted into text data by a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The converted text data is sent to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion.

[0044] Processing chat data

[0045] The server collects chat messages sent during online meetings in real time. The collected chat data is analyzed to extract important comments and topics. This analysis is performed using natural language processing technology and is integrated with summary information obtained from the audio data.

[0046] Conducting sentiment analysis

[0047] The server inputs the acquired voice and chat data into an emotion analysis model (such as VADER or DeepMoji) to analyze the emotional state of the meeting participants, which then classifies the overall mood of the meeting as positive, negative, neutral, etc.

[0048] Real-time provision

[0049] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[0050] Specific examples

[0051] For example, in a meeting initiated through an online conferencing platform, the server captures the audio stream in real time and uses a speech recognition engine to convert the speech "To discuss this year's sales targets" into text. The converted text is then analyzed using a natural language processing model to extract key points such as "sales targets," "key dates," and "market strategy."

[0052] At the same time, the server also collects chat data and analyzes and extracts important statements such as "Sales target is 20% higher than last year" and "Market strategy needs to be reviewed." This allows summaries of the voice data and chat data to be integrated.

[0053] The sentiment analysis model analyzes that the atmosphere in the meeting is positive, and information such as "There are many positive opinions regarding the sales target" can be obtained.

[0054] When a request is sent from a user device joining late, the server immediately sends summary information, and the user interface displays "Agenda: This year's sales target," "Important points: Sales target is to increase by 20% from last year," and "Meeting atmosphere: Mostly positive opinions." This allows users to quickly understand the content of the discussion and smoothly participate in the meeting.

[0055] In this way, this system provides an optimal means for online meeting participants to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[0056] The processing flow will be explained below.

[0057] Step 1:

[0058] The server confirms that the meeting has started through the API of the online meeting platform and starts streaming audio data, thereby capturing all spoken audio during the meeting.

[0059] Step 2:

[0060] The server sends the acquired voice data to a speech recognition engine (e.g., Google Cloud Speech-to-Text API) in real time and converts the voice data into text data. During this process, the voice data is processed continuously without interruption.

[0061] Step 3:

[0062] The server inputs the converted text data into a natural language processing model (e.g., BERT or GPT) and performs various analyses to extract the main points and important points of the discussion.

[0063] Step 4:

[0064] The server simultaneously receives chat data from the conference platform and analyzes it in real time, extracting important comments and frequently occurring topics from the chat data and saving them as text data.

[0065] Step 5:

[0066] The server combines the summaries of the voice data and chat data, organizing them so that the flow of the discussion and the main points are clear.

[0067] Step 6:

[0068] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the mood of the meeting (positive, negative, neutral). The results of the sentiment analysis are used to determine the overall mood of the meeting.

[0069] Step 7:

[0070] When a user terminal joins a conference late, it requests the information from the server, and this request includes a timestamp of when the user accessed the conference.

[0071] Step 8:

[0072] Upon receiving a user request, the server sends a summary of the discussion, key points, and the atmosphere of the meeting to the user's terminal in real time, allowing the user to instantly grasp the status of the meeting.

[0073] Step 9:

[0074] The user device displays the received information on the user interface. Specifically, the "agenda," "important points," "meeting atmosphere," etc. are displayed on the screen. Based on this information, the user can quickly join the meeting and make comments and exchange opinions.

[0075] This series of processes provides a system that allows late-joining users to efficiently understand the content of the discussion and contribute to the smooth progress of the meeting.

[0076] Example 1

[0077] 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."

[0078] Users who join online meetings late often have difficulty quickly grasping the main points and key points of the discussion and smoothly participating in the meeting. This problem arises because it takes a lot of time and effort to understand the entire content of the meeting, and because the information they need to know is not provided appropriately. It is also important to understand the atmosphere and emotional state of the meeting, and if this is not possible, users will be unable to participate effectively in the discussion.

[0079] 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.

[0080] In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data, and extracting important comments, means for analyzing the voice data and the chat data using sentiment analysis to identify the atmosphere of the meeting, means for providing the summarized main points and important points of the discussion and the atmosphere of the meeting to participants, means for integrating the main points extracted using natural language processing technology with the summary of the chat data, means for displaying the summarized information on a user interface, means for acquiring voice data in real time from an online conference platform, and means for providing the summarized information in real time in response to a request from a user terminal joining the online conference late, thereby enabling even users who join late to quickly understand the content of the discussion and effectively participate in the conference.

[0081] "Audio data" refers to data that records participants' comments and discussions that occur during an online conference.

[0082] "Text data" is voice data converted into character information, and is data generated by a voice recognition engine.

[0083] "Natural language processing" is a technology for analyzing text data and extracting the main points and important points of an argument.

[0084] "Chat data" refers to text message data sent by participants during an online conference.

[0085] "Sentiment analysis" is a technology that analyzes the emotional state of participants based on voice and chat data to identify the atmosphere of a meeting.

[0086] "Summary information" is information that includes the main points of the discussion, important points, and the atmosphere of the meeting, extracted from the voice data and chat data.

[0087] "User interface" refers to functions such as a screen or operation panel that allow a user to interact with a system.

[0088] An "online conference platform" is a technological infrastructure that allows multiple participants to communicate in real time via the Internet.

[0089] "Real-time" refers to data being generated and processed almost simultaneously, with very little delay.

[0090] A "user terminal" is a device such as a computer or smartphone that a user uses to access the system and obtain information.

[0091] A "request" refers to a request from a user terminal to a server to obtain information.

[0092] The present invention is a system that supports users who join an online conference late by quickly understanding the content of the discussion and allowing them to smoothly participate in the conference. Processing is mainly carried out between three parties: a server, a terminal, and the user.

[0093] Acquiring and converting audio data

[0094] At the start of an online meeting, the server acquires audio data through the API of the meeting platform (e.g., commonly known as a meeting API). The acquired audio data is streamed in real time. The audio data is converted into text data using a speech recognition engine (e.g., commonly known as a speech recognition API). The converted text data is used to extract the main points of the discussion.

[0095] Text summarization and analysis

[0096] The server sends the converted text data to a natural language processing model (e.g., an NLP model), which extracts the main points and key points of the discussion, which are key to quickly understanding the content of the online meeting.

[0097] Chat data collection and analysis

[0098] The server collects chat messages sent during online meetings in real time. The collected chat data is also analyzed using a natural language processing model to extract important statements and topics. This data is then combined with summary information obtained from the audio data.

[0099] Conducting sentiment analysis

[0100] The server inputs the acquired voice and chat data into a sentiment analysis model (e.g., a sentiment analysis model) to analyze the emotional state of the meeting participants, which can then categorize the mood of the meeting into categories such as positive, negative, or neutral.

[0101] Real-time provision

[0102] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[0103] Specific examples

[0104] For example, in a meeting initiated through a conferencing platform (commonly known as a conferencing API), the server captures the audio stream in real time. A speech recognition engine is used to convert the speech "To discuss this year's sales targets" into text. The converted text is then analyzed using a natural language processing model to extract key points such as "sales targets," "key dates," and "market strategy."

[0105] At the same time, the server also collects chat data using the chat API of the conference platform, analyzing and extracting important statements such as "Sales target is 20% higher than last year" and "Market strategy needs to be reviewed." This allows summaries of the voice and chat data to be integrated.

[0106] Through the sentiment analysis model, it is analyzed that the atmosphere of the meeting is positive, and information such as "there are many positive opinions regarding sales targets" can be obtained.

[0107] When a request is sent from a user terminal joining late, the server immediately sends summary information, and the user interface displays "Agenda: This year's sales target," "Important points: Sales target is to increase by 20% from last year," and "Meeting atmosphere: Mostly positive opinions." This allows the user to quickly understand the content of the discussion and smoothly participate in the meeting.

[0108] Prompt Sentence Examples

[0109] The following prompt sentences are examples of inputs to a generative AI model to generate a summary of an online meeting.

[0110] "To discuss this year's sales targets"

[0111] "Sales target is 20% increase from last year"

[0112] "Market strategy needs to be reviewed"

[0113] Sentiment analysis results: "Mostly positive opinions"

[0114] Generated summary: "At the meeting, the sales target for this fiscal year was discussed. The sales target was set at a 20% increase from last year, and it was decided that a review of the market strategy was necessary. The opinions of the meeting participants were positive, with many constructive suggestions being offered."

[0115] As a result, the present invention can support users who join an online conference late to efficiently understand the content of the discussion and effectively participate in the conference.

[0116] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0117] Step 1:

[0118] When an online conference starts, the server acquires audio data through the conference platform's API (for example, the common name conference API). Specifically, it detects the conference's "start event" and acquires the audio stream using the API. The acquired audio data is temporarily stored in a buffer. The input is the audio stream from the conference platform's API, and the output is the audio data stored in the buffer.

[0119] Step 2:

[0120] The server sends the voice data stored in the buffer to a speech recognition engine (for example, a speech recognition API) and converts it into text data. Specifically, it divides the voice data into chunks of a certain size and sends each chunk to the speech recognition engine. The text data returned by the engine is then combined into a single document. The input is the voice data stored in the buffer, and the output is text data.

[0121] Step 3:

[0122] The server sends the converted text data to a natural language processing model (e.g., an NLP model) to extract the main points and key points of the discussion. Specifically, the text data is input into the natural language processing model, which extracts the main points and keywords of the discussion. The input is the text data, and the output is the extracted main points and keywords.

[0123] Step 4:

[0124] The server collects chat messages sent during online meetings in real time through a chat API. Specifically, it monitors the chat's "new message received event" and captures and logs each new message posted. The input is the real-time message from the chat API, and the output is the collected chat data.

[0125] Step 5:

[0126] The server sends the collected chat data to a natural language processing model (NLP model) to extract important comments and topics. The acquired chat data is also analyzed by the NLP model to extract important comments and topics. The input is chat data, and the output is important comments and topics.

[0127] Step 6:

[0128] The server integrates summary information from the voice data and summary information from the chat data. Specifically, it combines each element (main points of the discussion, important keywords, important comments) and formats them into a single summary. The input is summary information from the voice data and summary information from the chat data, and the output is the integrated summary information.

[0129] Step 7:

[0130] The server inputs the integrated summary information into a sentiment analysis model (e.g., a sentiment analysis model) to analyze the emotional states of the meeting participants. Specifically, the text data of the summary information is input into the sentiment analysis model to calculate a positive, negative, or neutral sentiment score. The input is the integrated summary information, and the output is a sentiment score and information about the atmosphere of the meeting.

[0131] Step 8:

[0132] When a user terminal joins a conference late, the user terminal sends an information request to the server. Specifically, when the "join late" button is pressed on the user interface of the user terminal, the information request is sent to the server. The input is a request from the user interface, and the output is a request to the server.

[0133] Step 9:

[0134] The server receives requests from user terminals and transmits the summarized discussion content, key points, and meeting atmosphere to the user terminals in real time. Specifically, it analyzes the requests and transmits the stored summary information to the user terminals in real time. The input is the request from the user terminal, and the output is the transmission of the summary information.

[0135] Step 10:

[0136] The user terminal displays the information received from the server on the user interface. Specifically, the received summary information is displayed at an appropriate position on the screen to allow the user to quickly understand it. The input is the summary information received from the server, and the output is the information displayed on the user interface.

[0137] Through the above processing steps, this system enables even late-joining users to quickly understand the content of the discussion and smoothly participate in the conference.

[0138] (Application example 1)

[0139] 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."

[0140] In online meetings, there is a need for a system that allows late-joining users to quickly understand the content of the discussion and participate smoothly in the meeting. However, current systems make it difficult for late-joining users to immediately understand the content of the discussion, resulting in a decrease in productivity. In addition, there is a lack of means to grasp the overall atmosphere of the meeting, including sentiment analysis. This means that there is an urgent need to develop a system that allows users to participate in the meeting without falling behind.

[0141] 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.

[0142] In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data, and extracting important comments, means for analyzing the voice data and the chat data using sentiment analysis to identify the atmosphere of the meeting, means for providing the summarized main points and important points of the discussion and the atmosphere of the meeting to participants, data acquisition means for acquiring the voice data and the chat data, means for performing on-demand processing in real time using the data acquisition means, and means for visually displaying the processed data on a user interface, thereby enabling users to quickly understand the content of the discussion and smoothly participate in the meeting even if they join late.

[0143] "Audio data" refers to a digital representation of audio signals such as speech captured during an online conference.

[0144] "Text" is information that is created by transcribing audio data and expressing it as text.

[0145] "Natural language processing" is a general term for technology that processes and analyzes human language using computers.

[0146] "Chat data" refers to data of messages sent as text chat during an online conference.

[0147] "Sentiment analysis" is a technique for analyzing and classifying emotional nuances from text and audio data.

[0148] "Data acquisition means" refers to a method or device for acquiring necessary voice data or chat data from a server or user terminal.

[0149] "On-demand processing" is processing that analyzes data in real time as needed and provides results.

[0150] "User interface" is a general term for the screens and input devices that allow a user to interact with a system.

[0151] "Real-time" refers to immediate processing or response with little or no delay.

[0152] "Discussion points" are key points such as main opinions and conclusions discussed during an online meeting.

[0153] "Important points" refer to statements or information that are particularly noteworthy during online meetings.

[0154] "Atmosphere" refers to the atmosphere or mood that reflects the emotions and reactions of all meeting participants.

[0155] This system summarizes the main points and important points of discussions in online meetings, helping participants who join late to quickly understand the situation of the meeting. The process is mainly carried out between the server, the terminal, and the user.

[0156] Audio data processing

[0157] At the start of an online meeting, the server captures real-time audio data through the conference platform's API. This audio data is converted into text using a speech recognition engine such as Google Cloud Speech-to-Text API. This text data is then analyzed using natural language processing models such as BERT and GPT to extract the main points and key points of the discussion.

[0158] Processing chat data

[0159] The server also collects chat data sent during online meetings in real time. This chat data is analyzed using natural language processing technology to extract important comments and topics. This information is then combined with summary information obtained from the audio data.

[0160] Conducting sentiment analysis

[0161] The server inputs the acquired voice and chat data into emotion analysis models such as VADER and DeepMoji to analyze the emotional state of the meeting participants, which then classifies the meeting atmosphere as positive, negative, neutral, etc.

[0162] Real-time provision

[0163] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal visually displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[0164] Specific examples

[0165] For example, in a meeting initiated through an online conference platform, the server captures voice data in real time. It uses a speech recognition engine to convert speech, such as "To discuss this year's sales target," into text. This converted text is then analyzed using a natural language processing model to extract key points. At the same time, the server also collects chat data and analyzes and extracts important statements, such as "The sales target is to be increased by 20% from last year." This process integrates summaries of the voice and chat data. Using a sentiment analysis model, it is determined that the meeting atmosphere is positive, resulting in information such as "There is a lot of positive feedback regarding the sales target." When a request is sent from a user device joining late, the server immediately sends summary information, displaying it on the user interface as "Agenda: This year's sales target," "Key points: The sales target is to be increased by 20% from last year," and "Meeting atmosphere: There is a lot of positive feedback." This allows users to quickly understand the content of the discussion and smoothly participate in the meeting.

[0166] Example prompt sentence:

[0167] Prompts for generating key points

[0168] """

[0169] Meeting audio data: {audio data}

[0170] Chat content: {Chat content}

[0171] Please prepare your abstract in the following format:

[0172] summary:

[0173] 1. Agenda: {Agenda}

[0174] 2. Key Points: {Key Points}

[0175] 3. Meeting atmosphere: {Meeting atmosphere}

[0176] """

[0177] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0178] Step 1:

[0179] Acquiring audio data

[0180] Subject: Server

[0181] The server acquires audio data in real time using the API of the conference platform at the start of the online conference, using the audio stream of the conference platform as input and obtaining audio data as output, which is stored in a digital format and used for subsequent processing.

[0182] Step 2:

[0183] Transcription of audio data

[0184] Subject: Server

[0185] The server converts the acquired voice data into text using the Google Cloud Speech-to-Text API. Voice data is used as input, and the converted text data is obtained as output. This text data is a written representation of what was said during the online meeting.

[0186] Step 3:

[0187] Summarizing text data

[0188] Subject: Server

[0189] The server sends the converted text data to a natural language processing model (such as BERT or GPT) to extract the main points and key points of the discussion. The model uses the text data as input and obtains summarized text data as output. The model automatically extracts important topics and key points from the meeting.

[0190] Step 4:

[0191] Acquiring and analyzing chat data

[0192] Subject: Server

[0193] The server collects chat data sent during online meetings in real time. It uses the chat data as input and extracts important statements and topics analyzed as output. This data is then analyzed using natural language processing techniques.

[0194] Step 5:

[0195] Sentiment analysis of voice and chat data

[0196] Subject: Server

[0197] The server inputs the acquired voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting participants. It uses voice and chat data as input and obtains emotional states (positive, negative, neutral) as output, thereby identifying the overall mood of the meeting.

[0198] Step 6:

[0199] Integration and visualization of summary information

[0200] Subject: Server

[0201] The server integrates the summarized voice and chat data, as well as the results of the sentiment analysis, and prepares the data for display on the system's user interface. It uses each of the aforementioned data as input and obtains integrated summary information and emotional states as output. This data is then formatted for visualization.

[0202] Step 7:

[0203] Real-time provision to user devices

[0204] Subject: User device

[0205] When a user terminal joins a meeting late, it requests information from the server. The server sends prepared summary information and emotional state to the user terminal in real time. The server receives the request as input and displays the summary information and emotional state as output on the user interface. This process allows the user to quickly understand the content of the discussion and smoothly participate in the meeting.

[0206] 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.

[0207] This system helps late-joining users in online meetings quickly understand the content of the discussion and smoothly participate in the meeting. Processing is mainly carried out between the server, the terminal, and the user. In addition, by combining it with an emotion engine that analyzes the facial expressions and tone of voice of meeting participants, the atmosphere and emotional state of the meeting can be grasped in more detail.

[0208] Overview of program processing

[0209] Audio data processing

[0210] The server confirms that the meeting has started through the online meeting platform's API and obtains streaming audio data. This audio data is converted into text data in real time using a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The converted text data is then sent to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion.

[0211] Processing chat data

[0212] The server collects chat data sent during online meetings in real time. The collected chat data is analyzed using a natural language processing model to extract important statements and topics. This data is then integrated with summary information obtained from the audio data.

[0213] Conducting sentiment analysis

[0214] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting. The analysis results are classified as the mood of the meeting (positive, negative, or neutral).

[0215] Introducing the Emotion Engine

[0216] The server uses an emotion engine to analyze the facial expressions and tone of voice of meeting participants. The emotion engine analyzes facial expression changes from video streaming data to grasp the individual emotional states of meeting participants. In addition, the emotion engine also performs tone analysis of audio data to improve the accuracy of the emotional state. This allows for more detailed identification of the emotional state of the entire meeting.

[0217] Real-time provision

[0218] When a user terminal joins a conference late, it requests the information from the server. This request includes a timestamp of when the user accessed the conference. Upon receiving the user's request, the server transmits a summary of the discussion content, key points, and detailed emotional state to the user terminal in real time. The user terminal displays the received information on a user interface, helping the user quickly grasp the situation of the conference.

[0219] Specific examples

[0220] For example, in a meeting initiated through an online conference platform, the server captures audio streaming and chat data in real time. It uses a speech recognition engine to convert the speech "To discuss the development status of new products" into text. The converted text data is then analyzed using a natural language processing model to extract key points such as "new product development," "progress," and "marketing strategy."

[0221] At the same time, the server also collects chat data and analyzes and extracts important statements, such as "The new product development goal is until March" and "We need to review our marketing strategy." This data is also integrated with summaries of the voice data.

[0222] The emotion engine analyzes video streaming data from meeting participants and identifies their emotional state from their facial expressions and vocal tones. For example, in the middle of a meeting, it can identify a detailed emotional state such as "there was a very positive reaction to the development of a new product, and many participants were smiling."

[0223] When a user joins a meeting late, the user's device sends a request to the server, which immediately sends summary information and emotional state. The user interface displays information such as "Agenda: New product development progress," "Key points: Scheduled for completion by March," and "Meeting atmosphere: Very positive." This allows the user to quickly understand the meeting content and smoothly participate in the discussion.

[0224] In this way, this system provides an optimal means for participants in online meetings to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[0225] The processing flow will be explained below.

[0226] Step 1:

[0227] The server confirms that the meeting has started through the API of the online meeting platform and begins streaming audio data, thereby collecting all spoken audio during the meeting in real time.

[0228] Step 2:

[0229] The server sends the acquired voice data in real time to a speech recognition engine (e.g., Google Cloud Speech-to-Text API) and converts the voice into text data. The speech recognition engine analyzes the voice signal and generates text data such as "To discuss the development status of new products."

[0230] Step 3:

[0231] The server inputs the converted text data into a natural language processing model (e.g., BERT or GPT) and analyzes the text data. Through the analysis, the main points and important points of the discussion, such as "new product development," "progress," and "marketing strategy," are extracted.

[0232] Step 4:

[0233] The server simultaneously collects chat data sent from the online meeting platform. This data includes text messages between participants. These messages are analyzed in real time to extract important statements, such as "The new product development goal is scheduled to be completed by March" and "The marketing strategy needs to be reconsidered."

[0234] Step 5:

[0235] The server integrates the summary information obtained through speech recognition with the important comments obtained through chat analysis, thereby organizing the overall discussion content and creating a single summary.

[0236] Step 6:

[0237] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting participants. As a result of the analysis, the mood of the meeting is classified as positive, negative, or neutral.

[0238] Step 7:

[0239] The server also uses an emotion engine to analyze video streaming data such as participants' facial expressions and tone of voice. The emotion engine analyzes participants' facial expressions during the meeting to identify their specific emotional states (e.g., joy, surprise, anger).

[0240] Step 8:

[0241] The server integrates facial expression analysis and voice tone analysis of the video streaming data to identify the overall emotional state of the meeting in detail, such as "There were many positive reactions in the middle of the meeting" or "Many participants were smiling."

[0242] Step 9:

[0243] When a user terminal joins a conference late, it requests the information from the server, and the request includes a timestamp of when the user accessed the conference.

[0244] Step 10:

[0245] Upon receiving a request from the user terminal, the server transmits the summarized discussion content, key points, and detailed emotional state to the user terminal in real time.

[0246] Step 11:

[0247] The user device displays the received information on the user interface. Specifically, it displays information such as "Agenda: New product development progress," "Important points: Scheduled for completion by March," and "Meeting atmosphere: Very positive." This allows the user to quickly understand the content of the meeting and smoothly participate in the discussion.

[0248] Through this series of processes, our system provides an environment in which users who join online meetings late can efficiently understand the content of the meeting and smoothly participate in the discussion.By using the emotion engine, the overall atmosphere and emotional state of the meeting can be grasped in detail, providing even more useful information.

[0249] Example 2

[0250] 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."

[0251] Late-joining online meetings often face challenges in quickly grasping the content of the meeting and smoothly participating in the discussion. Conventional systems are unable to accurately grasp the atmosphere and emotional state of the meeting, making it difficult to understand the overall emotional trends of all participants. Therefore, there is a need for a system that can not only extract key points and important insights from voice and chat data, but also perform emotion analysis, allowing late-joining users to quickly grasp the meeting situation.

[0252] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0253] In this invention, the server includes means for confirming the start of the online conference platform and acquiring audio data via streaming, speech recognition means for converting the audio data into text data, means for analyzing the converted text data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data using natural language processing to extract important comments, means for analyzing the audio data and chat data using emotion analysis to identify the emotional state of the conference, means for analyzing changes in facial expressions and tone of voice of conference participants to grasp their detailed emotional states, and means for providing the summarized main points, important points, and emotional states of the discussion to users who join late, thereby enabling users who join late to quickly grasp the content and atmosphere of the conference and smoothly participate in the discussion.

[0254] An "online conference platform" is a system that allows multiple users to communicate with each other using audio and video over the Internet.

[0255] "Audio data" refers to digitized data of audio information uttered by participants during a conference.

[0256] "Text data" is sentence data converted from voice data by voice recognition.

[0257] "Speech recognition" is a technology that analyzes voice data and converts its contents into text data.

[0258] "Natural language processing" is a technology that enables computers to understand, analyze, and generate human language.

[0259] "Discussion points" are the core of the important topics or content discussed during the meeting.

[0260] "Key points" are pieces of information or statements that deserve special attention during a meeting.

[0261] "Chat data" refers to text message data sent by participants during an online conference.

[0262] "Sentiment analysis" is a technology that analyzes text data and audio data and identifies emotional states from their content.

[0263] "Emotional state" indicates the current emotional state of the meeting participants (e.g., positive, negative, neutral).

[0264] "Facial expression changes" refers to judging emotions by analyzing the facial movements of meeting participants.

[0265] "Voice tone" is an element used to determine emotions by analyzing the pitch and strength of a speaker's voice.

[0266] A "user" is a person who participates in an online conference.

[0267] A "server" is a computer system that operates an online conference system and processes data.

[0268] A "user terminal" is a device such as a computer or smartphone that a user uses to participate in an online conference.

[0269] "Streaming" is a method of transmitting data continuously in real time.

[0270] This invention is a system that helps late-joining users in online conferences quickly understand the content of the discussion and smoothly participate in the conference. Processing is mainly carried out between the server, the terminal, and the user, and by combining it with an emotion engine that analyzes the facial expressions and tone of voice of conference participants, it is possible to grasp the atmosphere and emotional state of the conference in more detail.

[0271] Audio data processing

[0272] The server confirms that the meeting has started through the online meeting platform's API (e.g., Zoom API, Microsoft Teams API), and obtains streaming audio data. This audio data is converted into text data in real time using a speech recognition engine such as Google Cloud Speech-to-Text API. The converted text data is then sent to a natural language processing model such as BERT or GPT to extract the main points and key points of the discussion.

[0273] Processing chat data

[0274] The server collects chat data sent during online meetings in real time. The collected chat data is analyzed using a natural language processing model to extract important comments and topics. This data is integrated with summary information obtained from the voice data and provided to the user.

[0275] Conducting sentiment analysis

[0276] The server inputs the voice and chat data into emotion analysis models such as VADER and DeepMoji to analyze the emotional state of the meeting. The analysis results are classified as the meeting mood (positive, negative, or neutral), allowing users to quickly understand the situation of the meeting.

[0277] Introducing the Emotion Engine

[0278] The server uses an emotion engine to analyze the facial expressions and tone of voice of meeting participants. The emotion engine analyzes the video streaming data of the meeting to understand the individual emotional state of meeting participants. In addition, the emotion engine also performs tone analysis of the audio data to more precisely identify the overall emotional state of the meeting.

[0279] Real-time provision

[0280] When a user terminal joins a conference late, it requests information from the server. This request includes a timestamp of when the user accessed the conference. Upon receiving the user's request, the server transmits a summary of the discussion, key points, and detailed emotional state to the user terminal in real time. The user terminal displays the received information on a user interface to help the user quickly understand the situation of the conference.

[0281] Specific examples

[0282] For example, in a meeting initiated through an online conference platform, the server captures audio streaming and chat data in real time. A speech recognition engine is used to convert the speech "To discuss the development status of new products" into text. The converted text data is analyzed using a natural language processing model (GPT-4) to extract key points such as "new product development," "progress," and "marketing strategy."

[0283] At the same time, the server also collects chat data and analyzes it using a natural language processing model to extract important statements, such as "The new product development goal is until March" or "We need to review our marketing strategy." This data is then integrated with summary information from the voice data.

[0284] The server uses an emotion engine to analyze the video streaming data of the meeting and identify the emotional state of the meeting participants from their facial expressions and voice tones. For example, in the middle of the meeting, a detailed emotional state can be identified, such as "there was a very positive reaction to the development of the new product, and many of the participants were smiling."

[0285] When a user joins a meeting late, the user device sends a request to the server, and the server quickly sends summary information and emotional state to the user device. The user interface displays "Agenda: New product development progress," "Key points: Scheduled for completion by March," and "Meeting atmosphere: Very positive," allowing the user to quickly understand the meeting content and smoothly participate in the discussion.

[0286] Prompt Sentence Examples

[0287] "Summarize the minutes of the online meeting below and extract the key points. Also, please rate the mood of the meeting as positive, negative, or neutral."

[0288] In this way, this system provides an optimal means for participants in online meetings to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[0289] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0290] Step 1:

[0291] The server uses the API of the online meeting platform to confirm that the meeting has started. Specifically, it uses the Zoom API or Microsoft Teams API to obtain initial data such as the meeting ID and start time, and is then ready to start the audio streaming session.

[0292] Input: Meeting ID and start time obtained from the conference platform API

[0293] Output: Session information for conference streaming

[0294] Step 2:

[0295] The server receives the conference audio data in real time via streaming and converts it into text using the Google Cloud Speech-to-Text API. The audio data is converted in real time and saved as a temporary file.

[0296] Input: Conference streaming audio data

[0297] Output: Text data converted by Google Cloud Speech-to-Text API

[0298] Step 3:

[0299] The server then inputs the converted text data into a natural language processing model (e.g., BERT or GPT) to extract key points and key takeaways from the discussion, identifying the core points discussed during the meeting and storing them in a database.

[0300] Input: Text data converted by the Google Cloud Speech-to-Text API

[0301] Output: Data on gist and key points extracted by the NLP model

[0302] Step 4:

[0303] The server collects chat data using the chat API of the online meeting platform, and inputs the data into a natural language processing model for analysis to extract important comments and topics, thereby understanding the content and importance of the comments.

[0304] Input: Chat data during the meeting

[0305] Output: Data on important utterances and topics extracted by the natural language processing model

[0306] Step 5:

[0307] The server inputs the voice and chat data into a sentiment analysis model such as VADER or DeepMoji to analyze the emotional state of the meeting, classifying the mood of the meeting as positive, negative, or neutral, and storing this information in a database.

[0308] Input: Voice and chat data

[0309] Output: Data on emotional states (positive, negative, neutral) classified by the sentiment analysis model

[0310] Step 6:

[0311] The server inputs video streaming data into an emotion engine to analyze participants' facial expressions and vocal tones, thereby identifying their individual emotional states and providing a more detailed understanding of the overall emotional state of the meeting.

[0312] Input: Video streaming data from a conference

[0313] Output: Data on the individual and overall emotional states of participants analyzed by the emotion engine

[0314] Step 7:

[0315] When a user terminal joins a meeting late, it sends an information request to the server. This request includes a timestamp of the time of joining. The server receives the request and sends a summary of the discussion content, key points, and emotional state to the user terminal in real time. The user terminal displays the received information on a user interface to help the user quickly understand the situation of the meeting.

[0316] Input: Information request from user device (including timestamp)

[0317] Output: Summary information sent from the server, key points, emotional state data, and information displayed in the user interface

[0318] (Application example 2)

[0319] 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."

[0320] With conventional online conference systems, it was difficult for users who joined a meeting late to quickly grasp the content of the discussion. In particular, in store meetings, not being able to immediately grasp the key points and atmosphere of the meeting has a negative impact on work efficiency. Furthermore, there is a lack of means to accurately grasp the emotional state and atmosphere of the meeting, making it even more difficult to grasp the flow of the meeting. To solve these issues, there is a need for a system that can help store managers and staff who join late quickly grasp the content of employee meetings and strategy meetings and effectively apply it to their work.

[0321] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, and means for collecting chat data, analyzing the chat data, and extracting important comments. This makes it possible to analyze the voice data and chat data using sentiment analysis to identify the atmosphere of the meeting. Also included is means for acquiring voice data of a store meeting from a smart device in real time, providing a system that allows users who join a meeting late to quickly understand the situation and smoothly reflect it in their work.

[0322] "Audio data" refers to data that digitally represents information including the sounds of meetings and conversations.

[0323] "Text" is document data obtained by converting voice data into a character string.

[0324] "Natural language processing" is a technology that enables computers to understand, analyze, and generate human language.

[0325] "Key points" refer to the particularly important points or main content in a meeting or discussion.

[0326] "Key points" are those points in a discussion or statement that deserve special attention or priority.

[0327] "Chat data" refers to data used by participants to send and receive text messages during a conference.

[0328] "Sentiment analysis" is a technique for analyzing and identifying a speaker's emotional state from text or audio data.

[0329] "Analysis" is the process of examining and breaking down data to understand its structure and meaning.

[0330] A "server" is a computer system that manages and processes data on a network.

[0331] A "store meeting" is a conference or meeting held by employees and managers at a physical store.

[0332] "Smart devices" are electronic devices with advanced functions, such as smartphones and tablets.

[0333] A "summary" is a concise summary of a longer piece of writing or discussion.

[0334] A "user interface" is a display screen or operating means for interacting with a system.

[0335] The embodiment of this invention is a system that enables users who join an online conference late to quickly understand the conference content and smoothly participate in the discussion. This system is mainly composed of a server, user terminals, and smart devices.

[0336] The server collects voice data and converts it into text. This process uses the Google Cloud Speech-to-Text API as a speech recognition engine. The converted text data is passed through a natural language processing model (e.g., BERT or GPT) to extract the main points and important points of the discussion. The server also collects chat data sent during the meeting and analyzes it with the natural language processing model to extract important comments.

[0337] The server then analyzes the voice and chat data using a sentiment analysis model (e.g., VADER or DeepMoji) to classify the meeting mood as positive, negative, or neutral. This sentiment analysis makes it possible to understand the mood of the meeting and the emotional state of the participants.

[0338] This data is provided in real time to user devices such as smartphones and tablets, which display a summary of the main points of the discussion, key points, and the atmosphere of the meeting on their user interface, allowing late-joining users to quickly understand the current state of the meeting and participate in the discussion efficiently.

[0339] As a concrete example, consider the case where a store manager is late to an important store meeting. The manager opens an application on his smartphone and requests information about the past meeting. The server immediately sends a summary of the meeting content, key points, and emotional state to the manager's smartphone. The user interface displays "Agenda: New product development progress," "Key points: Expected completion by March," and "Meeting atmosphere: Very positive."

[0340] An example prompt might look like this:

[0341] "Please summarize the meeting. What were the key points and the emotional state of the participants?"

[0342] As described above, the system of the present invention supports users who join an online conference late to quickly grasp the content and atmosphere of the discussion at the conference.

[0343] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0344] Step 1:

[0345] The server collects voice data in real time at the start of an online meeting. Specifically, it uses the Google Cloud Speech-to-Text API, a speech recognition engine, to receive the audio of the online meeting as input. This voice data is received in streaming format and converted into text data.

[0346] Step 2:

[0347] The server sends the converted text data from the audio data to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion. This step uses the converted audio data as input. The natural language processing model analyzes the text data, identifies the main points and key points of the discussion, and outputs a summary.

[0348] Step 3:

[0349] The server collects chat data sent during online meetings in real time, receives chat messages as input, and stores their contents as text data.

[0350] Step 4:

[0351] The server analyzes the collected chat data using a natural language processing model to extract important comments. The text data of chat messages is used as input. The model extracts important comments and topics from the text data and integrates them into a summary.

[0352] Step 5:

[0353] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the mood of the meeting. This step uses voice and chat text data as input. The sentiment analysis model analyzes the data and classifies it as positive, negative, or neutral, and outputs the mood of the meeting.

[0354] Step 6:

[0355] When a user terminal joins a conference late, the user terminal sends a request to the server, sending request data including a timestamp when the user accesses the conference as input.

[0356] Step 7:

[0357] The server receives the user's request and transmits the summarized discussion summary, key points, and meeting atmosphere information to the user's terminal in real time. In this step, the server transmits the summarized summary, key points, and meeting atmosphere information as output.

[0358] Step 8:

[0359] The user terminal displays the received information on the user interface. In this step, the information received by the user is displayed in a visually easy-to-identify format, ultimately enabling the user to quickly understand the situation.

[0360] By following the steps above, users who join late can efficiently grasp the content and atmosphere of the meeting and smoothly reflect this in their work.

[0361] 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.

[0362] 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.

[0363] 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.

[0364] [Second embodiment]

[0365] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0366] 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.

[0367] 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).

[0368] 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.

[0369] 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.

[0370] 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).

[0371] 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.

[0372] 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.

[0373] 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.

[0374] 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.

[0375] 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.

[0376] 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."

[0377] The present invention is a system that helps users who join an online conference late quickly understand the content of the discussion and participate in the conference smoothly, and processing is mainly carried out between three parties: a server, a terminal, and the user.

[0378] Overview of program processing

[0379] Audio data processing

[0380] At the start of an online meeting, the server acquires audio data through the conference platform's API. This audio data is streamed in real time and converted into text data by a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The converted text data is sent to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion.

[0381] Processing chat data

[0382] The server collects chat messages sent during online meetings in real time. The collected chat data is analyzed to extract important comments and topics. This analysis is performed using natural language processing technology and is integrated with summary information obtained from the audio data.

[0383] Conducting sentiment analysis

[0384] The server inputs the acquired voice and chat data into an emotion analysis model (such as VADER or DeepMoji) to analyze the emotional state of the meeting participants, which then classifies the overall mood of the meeting as positive, negative, neutral, etc.

[0385] Real-time provision

[0386] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[0387] Specific examples

[0388] For example, in a meeting initiated through an online conferencing platform, the server captures the audio stream in real time and uses a speech recognition engine to convert the speech "To discuss this year's sales targets" into text. The converted text is then analyzed using a natural language processing model to extract key points such as "sales targets," "key dates," and "market strategy."

[0389] At the same time, the server also collects chat data and analyzes and extracts important statements such as "Sales target is 20% higher than last year" and "Market strategy needs to be reviewed." This allows summaries of the voice data and chat data to be integrated.

[0390] The sentiment analysis model analyzes that the atmosphere in the meeting is positive, and information such as "There are many positive opinions regarding the sales target" can be obtained.

[0391] When a request is sent from a user device joining late, the server immediately sends summary information, and the user interface displays "Agenda: This year's sales target," "Important points: Sales target is to increase by 20% from last year," and "Meeting atmosphere: Mostly positive opinions." This allows users to quickly understand the content of the discussion and smoothly participate in the meeting.

[0392] In this way, this system provides an optimal means for online meeting participants to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[0393] The processing flow will be explained below.

[0394] Step 1:

[0395] The server confirms that the meeting has started through the API of the online meeting platform and starts streaming audio data, thereby capturing all spoken audio during the meeting.

[0396] Step 2:

[0397] The server sends the acquired voice data to a speech recognition engine (e.g., Google Cloud Speech-to-Text API) in real time and converts the voice data into text data. During this process, the voice data is processed continuously without interruption.

[0398] Step 3:

[0399] The server inputs the converted text data into a natural language processing model (e.g., BERT or GPT) and performs various analyses to extract the main points and important points of the discussion.

[0400] Step 4:

[0401] The server simultaneously receives chat data from the conference platform and analyzes it in real time, extracting important comments and frequently occurring topics from the chat data and saving them as text data.

[0402] Step 5:

[0403] The server combines the summaries of the voice data and chat data, organizing them so that the flow of the discussion and the main points are clear.

[0404] Step 6:

[0405] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the mood of the meeting (positive, negative, neutral). The results of the sentiment analysis are used to determine the overall mood of the meeting.

[0406] Step 7:

[0407] When a user terminal joins a conference late, it requests the information from the server, and this request includes a timestamp of when the user accessed the conference.

[0408] Step 8:

[0409] Upon receiving a user request, the server sends a summary of the discussion, key points, and the atmosphere of the meeting to the user's terminal in real time, allowing the user to instantly grasp the status of the meeting.

[0410] Step 9:

[0411] The user device displays the received information on the user interface. Specifically, the "agenda," "important points," "meeting atmosphere," etc. are displayed on the screen. Based on this information, the user can quickly join the meeting and make comments and exchange opinions.

[0412] This series of processes provides a system that allows late-joining users to efficiently understand the content of the discussion and contribute to the smooth progress of the meeting.

[0413] Example 1

[0414] 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."

[0415] Users who join online meetings late often have difficulty quickly grasping the main points and key points of the discussion and smoothly participating in the meeting. This problem arises because it takes a lot of time and effort to understand the entire content of the meeting, and because the information they need to know is not provided appropriately. It is also important to understand the atmosphere and emotional state of the meeting, and if this is not possible, users will be unable to participate effectively in the discussion.

[0416] 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.

[0417] In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data, and extracting important comments, means for analyzing the voice data and the chat data using sentiment analysis to identify the atmosphere of the meeting, means for providing the summarized main points and important points of the discussion and the atmosphere of the meeting to participants, means for integrating the main points extracted using natural language processing technology with the summary of the chat data, means for displaying the summarized information on a user interface, means for acquiring voice data in real time from an online conference platform, and means for providing the summarized information in real time in response to a request from a user terminal joining the online conference late, thereby enabling even users who join late to quickly understand the content of the discussion and effectively participate in the conference.

[0418] "Audio data" refers to data that records participants' comments and discussions that occur during an online conference.

[0419] "Text data" is voice data converted into character information, and is data generated by a voice recognition engine.

[0420] "Natural language processing" is a technology for analyzing text data and extracting the main points and important points of an argument.

[0421] "Chat data" refers to text message data sent by participants during an online conference.

[0422] "Sentiment analysis" is a technology that analyzes the emotional state of participants based on voice and chat data to identify the atmosphere of a meeting.

[0423] "Summary information" is information that includes the main points of the discussion, important points, and the atmosphere of the meeting, extracted from the voice data and chat data.

[0424] "User interface" refers to functions such as a screen or operation panel that allow a user to interact with a system.

[0425] An "online conference platform" is a technological infrastructure that allows multiple participants to communicate in real time via the Internet.

[0426] "Real-time" refers to data being generated and processed almost simultaneously, with very little delay.

[0427] A "user terminal" is a device such as a computer or smartphone that a user uses to access the system and obtain information.

[0428] A "request" refers to a request from a user terminal to a server to obtain information.

[0429] The present invention is a system that supports users who join an online conference late by quickly understanding the content of the discussion and allowing them to smoothly participate in the conference. Processing is mainly carried out between three parties: a server, a terminal, and the user.

[0430] Acquiring and converting audio data

[0431] At the start of an online meeting, the server acquires audio data through the API of the meeting platform (e.g., commonly known as a meeting API). The acquired audio data is streamed in real time. The audio data is converted into text data using a speech recognition engine (e.g., commonly known as a speech recognition API). The converted text data is used to extract the main points of the discussion.

[0432] Text summarization and analysis

[0433] The server sends the converted text data to a natural language processing model (e.g., an NLP model), which extracts the main points and key points of the discussion, which are key to quickly understanding the content of the online meeting.

[0434] Chat data collection and analysis

[0435] The server collects chat messages sent during online meetings in real time. The collected chat data is also analyzed using a natural language processing model to extract important statements and topics. This data is then combined with summary information obtained from the audio data.

[0436] Conducting sentiment analysis

[0437] The server inputs the acquired voice and chat data into a sentiment analysis model (e.g., a sentiment analysis model) to analyze the emotional state of the meeting participants, which can then categorize the mood of the meeting into categories such as positive, negative, or neutral.

[0438] Real-time provision

[0439] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[0440] Specific examples

[0441] For example, in a meeting initiated through a conferencing platform (commonly known as a conferencing API), the server captures the audio stream in real time. A speech recognition engine is used to convert the speech "To discuss this year's sales targets" into text. The converted text is then analyzed using a natural language processing model to extract key points such as "sales targets," "key dates," and "market strategy."

[0442] At the same time, the server also collects chat data using the chat API of the conference platform, analyzing and extracting important statements such as "Sales target is 20% higher than last year" and "Market strategy needs to be reviewed." This allows summaries of the voice and chat data to be integrated.

[0443] Through the sentiment analysis model, it is analyzed that the atmosphere of the meeting is positive, and information such as "there are many positive opinions regarding sales targets" can be obtained.

[0444] When a request is sent from a user terminal joining late, the server immediately sends summary information, and the user interface displays "Agenda: This year's sales target," "Important points: Sales target is to increase by 20% from last year," and "Meeting atmosphere: Mostly positive opinions." This allows the user to quickly understand the content of the discussion and smoothly participate in the meeting.

[0445] Prompt Sentence Examples

[0446] The following prompt sentences are examples of inputs to a generative AI model to generate a summary of an online meeting.

[0447] "To discuss this year's sales targets"

[0448] "Sales target is 20% increase from last year"

[0449] "Market strategy needs to be reviewed"

[0450] Sentiment analysis results: "Mostly positive opinions"

[0451] Generated summary: "At the meeting, the sales target for this fiscal year was discussed. The sales target was set at a 20% increase from last year, and it was decided that a review of the market strategy was necessary. The opinions of the meeting participants were positive, with many constructive suggestions being offered."

[0452] As a result, the present invention can support users who join an online conference late to efficiently understand the content of the discussion and effectively participate in the conference.

[0453] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0454] Step 1:

[0455] When an online conference starts, the server acquires audio data through the conference platform's API (for example, the common name conference API). Specifically, it detects the conference's "start event" and acquires the audio stream using the API. The acquired audio data is temporarily stored in a buffer. The input is the audio stream from the conference platform's API, and the output is the audio data stored in the buffer.

[0456] Step 2:

[0457] The server sends the voice data stored in the buffer to a speech recognition engine (for example, a speech recognition API) and converts it into text data. Specifically, it divides the voice data into chunks of a certain size and sends each chunk to the speech recognition engine. The text data returned by the engine is then combined into a single document. The input is the voice data stored in the buffer, and the output is text data.

[0458] Step 3:

[0459] The server sends the converted text data to a natural language processing model (e.g., an NLP model) to extract the main points and key points of the discussion. Specifically, the text data is input into the natural language processing model, which extracts the main points and keywords of the discussion. The input is the text data, and the output is the extracted main points and keywords.

[0460] Step 4:

[0461] The server collects chat messages sent during online meetings in real time through a chat API. Specifically, it monitors the chat's "new message received event" and captures and logs each new message posted. The input is the real-time message from the chat API, and the output is the collected chat data.

[0462] Step 5:

[0463] The server sends the collected chat data to a natural language processing model (NLP model) to extract important comments and topics. The acquired chat data is also analyzed by the NLP model to extract important comments and topics. The input is chat data, and the output is important comments and topics.

[0464] Step 6:

[0465] The server integrates summary information from the voice data and summary information from the chat data. Specifically, it combines each element (main points of the discussion, important keywords, important comments) and formats them into a single summary. The input is summary information from the voice data and summary information from the chat data, and the output is the integrated summary information.

[0466] Step 7:

[0467] The server inputs the integrated summary information into a sentiment analysis model (e.g., a sentiment analysis model) to analyze the emotional states of the meeting participants. Specifically, the text data of the summary information is input into the sentiment analysis model to calculate a positive, negative, or neutral sentiment score. The input is the integrated summary information, and the output is a sentiment score and information about the atmosphere of the meeting.

[0468] Step 8:

[0469] When a user terminal joins a conference late, the user terminal sends an information request to the server. Specifically, when the "join late" button is pressed on the user interface of the user terminal, the information request is sent to the server. The input is a request from the user interface, and the output is a request to the server.

[0470] Step 9:

[0471] The server receives requests from user terminals and transmits the summarized discussion content, key points, and meeting atmosphere to the user terminals in real time. Specifically, it analyzes the requests and transmits the stored summary information to the user terminals in real time. The input is the request from the user terminal, and the output is the transmission of the summary information.

[0472] Step 10:

[0473] The user terminal displays the information received from the server on the user interface. Specifically, the received summary information is displayed at an appropriate position on the screen to allow the user to quickly understand it. The input is the summary information received from the server, and the output is the information displayed on the user interface.

[0474] Through the above processing steps, this system enables even late-joining users to quickly understand the content of the discussion and smoothly participate in the conference.

[0475] (Application example 1)

[0476] 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."

[0477] In online meetings, there is a need for a system that allows late-joining users to quickly understand the content of the discussion and participate smoothly in the meeting. However, current systems make it difficult for late-joining users to immediately understand the content of the discussion, resulting in a decrease in productivity. In addition, there is a lack of means to grasp the overall atmosphere of the meeting, including sentiment analysis. This means that there is an urgent need to develop a system that allows users to participate in the meeting without falling behind.

[0478] 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.

[0479] In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data, and extracting important comments, means for analyzing the voice data and the chat data using sentiment analysis to identify the atmosphere of the meeting, means for providing the summarized main points and important points of the discussion and the atmosphere of the meeting to participants, data acquisition means for acquiring the voice data and the chat data, means for performing on-demand processing in real time using the data acquisition means, and means for visually displaying the processed data on a user interface, thereby enabling users to quickly understand the content of the discussion and smoothly participate in the meeting even if they join late.

[0480] "Audio data" refers to a digital representation of audio signals such as speech captured during an online conference.

[0481] "Text" is information that is created by transcribing audio data and expressing it as text.

[0482] "Natural language processing" is a general term for technology that processes and analyzes human language using computers.

[0483] "Chat data" refers to data of messages sent as text chat during an online conference.

[0484] "Sentiment analysis" is a technique for analyzing and classifying emotional nuances from text and audio data.

[0485] "Data acquisition means" refers to a method or device for acquiring necessary voice data or chat data from a server or user terminal.

[0486] "On-demand processing" is processing that analyzes data in real time as needed and provides results.

[0487] "User interface" is a general term for the screens and input devices that allow a user to interact with a system.

[0488] "Real-time" refers to immediate processing or response with little or no delay.

[0489] "Discussion points" are key points such as main opinions and conclusions discussed during an online meeting.

[0490] "Important points" refer to statements or information that are particularly noteworthy during online meetings.

[0491] "Atmosphere" refers to the atmosphere or mood that reflects the emotions and reactions of all meeting participants.

[0492] This system summarizes the main points and important points of discussions in online meetings, helping participants who join late to quickly understand the situation of the meeting. The process is mainly carried out between the server, the terminal, and the user.

[0493] Audio data processing

[0494] At the start of an online meeting, the server captures real-time audio data through the conference platform's API. This audio data is converted into text using a speech recognition engine such as Google Cloud Speech-to-Text API. This text data is then analyzed using natural language processing models such as BERT and GPT to extract the main points and key points of the discussion.

[0495] Processing chat data

[0496] The server also collects chat data sent during online meetings in real time. This chat data is analyzed using natural language processing technology to extract important comments and topics. This information is then combined with summary information obtained from the audio data.

[0497] Conducting sentiment analysis

[0498] The server inputs the acquired voice and chat data into emotion analysis models such as VADER and DeepMoji to analyze the emotional state of the meeting participants, which then classifies the meeting atmosphere as positive, negative, neutral, etc.

[0499] Real-time provision

[0500] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal visually displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[0501] Specific examples

[0502] For example, in a meeting initiated through an online conference platform, the server captures voice data in real time. It uses a speech recognition engine to convert speech, such as "To discuss this year's sales target," into text. This converted text is then analyzed using a natural language processing model to extract key points. At the same time, the server also collects chat data and analyzes and extracts important statements, such as "The sales target is to be increased by 20% from last year." This process integrates summaries of the voice and chat data. Using a sentiment analysis model, it is determined that the meeting atmosphere is positive, resulting in information such as "There is a lot of positive feedback regarding the sales target." When a request is sent from a user device joining late, the server immediately sends summary information, displaying it on the user interface as "Agenda: This year's sales target," "Key points: The sales target is to be increased by 20% from last year," and "Meeting atmosphere: There is a lot of positive feedback." This allows users to quickly understand the content of the discussion and smoothly participate in the meeting.

[0503] Example prompt sentence:

[0504] Prompts for generating key points

[0505] """

[0506] Meeting audio data: {audio data}

[0507] Chat content: {Chat content}

[0508] Please prepare your abstract in the following format:

[0509] summary:

[0510] 1. Agenda: {Agenda}

[0511] 2. Key Points: {Key Points}

[0512] 3. Meeting atmosphere: {Meeting atmosphere}

[0513] """

[0514] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0515] Step 1:

[0516] Acquiring audio data

[0517] Subject: Server

[0518] The server acquires audio data in real time using the API of the conference platform at the start of the online conference, using the audio stream of the conference platform as input and obtaining audio data as output, which is stored in a digital format and used for subsequent processing.

[0519] Step 2:

[0520] Transcription of audio data

[0521] Subject: Server

[0522] The server converts the acquired voice data into text using the Google Cloud Speech-to-Text API. Voice data is used as input, and the converted text data is obtained as output. This text data is a written representation of what was said during the online meeting.

[0523] Step 3:

[0524] Summarizing text data

[0525] Subject: Server

[0526] The server sends the converted text data to a natural language processing model (such as BERT or GPT) to extract the main points and key points of the discussion. The model uses the text data as input and obtains summarized text data as output. The model automatically extracts important topics and key points from the meeting.

[0527] Step 4:

[0528] Acquiring and analyzing chat data

[0529] Subject: Server

[0530] The server collects chat data sent during online meetings in real time. It uses the chat data as input and extracts important statements and topics analyzed as output. This data is then analyzed using natural language processing techniques.

[0531] Step 5:

[0532] Sentiment analysis of voice and chat data

[0533] Subject: Server

[0534] The server inputs the acquired voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting participants. It uses voice and chat data as input and obtains emotional states (positive, negative, neutral) as output, thereby identifying the overall mood of the meeting.

[0535] Step 6:

[0536] Integration and visualization of summary information

[0537] Subject: Server

[0538] The server integrates the summarized voice and chat data, as well as the results of the sentiment analysis, and prepares the data for display on the system's user interface. It uses each of the aforementioned data as input and obtains integrated summary information and emotional states as output. This data is then formatted for visualization.

[0539] Step 7:

[0540] Real-time provision to user devices

[0541] Subject: User device

[0542] When a user terminal joins a meeting late, it requests information from the server. The server sends prepared summary information and emotional state to the user terminal in real time. The server receives the request as input and displays the summary information and emotional state as output on the user interface. This process allows the user to quickly understand the content of the discussion and smoothly participate in the meeting.

[0543] 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.

[0544] This system helps late-joining users in online meetings quickly understand the content of the discussion and smoothly participate in the meeting. Processing is mainly carried out between the server, the terminal, and the user. In addition, by combining it with an emotion engine that analyzes the facial expressions and tone of voice of meeting participants, the atmosphere and emotional state of the meeting can be grasped in more detail.

[0545] Overview of program processing

[0546] Audio data processing

[0547] The server confirms that the meeting has started through the online meeting platform's API and obtains streaming audio data. This audio data is converted into text data in real time using a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The converted text data is then sent to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion.

[0548] Processing chat data

[0549] The server collects chat data sent during online meetings in real time. The collected chat data is analyzed using a natural language processing model to extract important statements and topics. This data is then integrated with summary information obtained from the audio data.

[0550] Conducting sentiment analysis

[0551] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting. The analysis results are classified as the mood of the meeting (positive, negative, or neutral).

[0552] Introducing the Emotion Engine

[0553] The server uses an emotion engine to analyze the facial expressions and tone of voice of meeting participants. The emotion engine analyzes facial expression changes from video streaming data to grasp the individual emotional states of meeting participants. In addition, the emotion engine also performs tone analysis of audio data to improve the accuracy of the emotional state. This allows for more detailed identification of the emotional state of the entire meeting.

[0554] Real-time provision

[0555] When a user terminal joins a conference late, it requests the information from the server. This request includes a timestamp of when the user accessed the conference. Upon receiving the user's request, the server transmits a summary of the discussion content, key points, and detailed emotional state to the user terminal in real time. The user terminal displays the received information on a user interface, helping the user quickly grasp the situation of the conference.

[0556] Specific examples

[0557] For example, in a meeting initiated through an online conference platform, the server captures audio streaming and chat data in real time. It uses a speech recognition engine to convert the speech "To discuss the development status of new products" into text. The converted text data is then analyzed using a natural language processing model to extract key points such as "new product development," "progress," and "marketing strategy."

[0558] At the same time, the server also collects chat data and analyzes and extracts important statements, such as "The new product development goal is until March" and "We need to review our marketing strategy." This data is also integrated with summaries of the voice data.

[0559] The emotion engine analyzes video streaming data from meeting participants and identifies their emotional state from their facial expressions and vocal tones. For example, in the middle of a meeting, it can identify a detailed emotional state such as "there was a very positive reaction to the development of a new product, and many participants were smiling."

[0560] When a user joins a meeting late, the user's device sends a request to the server, which immediately sends summary information and emotional state. The user interface displays information such as "Agenda: New product development progress," "Key points: Scheduled for completion by March," and "Meeting atmosphere: Very positive." This allows the user to quickly understand the meeting content and smoothly participate in the discussion.

[0561] In this way, this system provides an optimal means for participants in online meetings to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[0562] The processing flow will be explained below.

[0563] Step 1:

[0564] The server confirms that the meeting has started through the API of the online meeting platform and begins streaming audio data, thereby collecting all spoken audio during the meeting in real time.

[0565] Step 2:

[0566] The server sends the acquired voice data in real time to a speech recognition engine (e.g., Google Cloud Speech-to-Text API) and converts the voice into text data. The speech recognition engine analyzes the voice signal and generates text data such as "To discuss the development status of new products."

[0567] Step 3:

[0568] The server inputs the converted text data into a natural language processing model (e.g., BERT or GPT) and analyzes the text data. Through the analysis, the main points and important points of the discussion, such as "new product development," "progress," and "marketing strategy," are extracted.

[0569] Step 4:

[0570] The server simultaneously collects chat data sent from the online meeting platform. This data includes text messages between participants. These messages are analyzed in real time to extract important statements, such as "The new product development goal is scheduled to be completed by March" and "The marketing strategy needs to be reconsidered."

[0571] Step 5:

[0572] The server integrates the summary information obtained through speech recognition with the important comments obtained through chat analysis, thereby organizing the overall discussion content and creating a single summary.

[0573] Step 6:

[0574] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting participants. As a result of the analysis, the mood of the meeting is classified as positive, negative, or neutral.

[0575] Step 7:

[0576] The server also uses an emotion engine to analyze video streaming data such as participants' facial expressions and tone of voice. The emotion engine analyzes participants' facial expressions during the meeting to identify their specific emotional states (e.g., joy, surprise, anger).

[0577] Step 8:

[0578] The server integrates facial expression analysis and voice tone analysis of the video streaming data to identify the overall emotional state of the meeting in detail, such as "There were many positive reactions in the middle of the meeting" or "Many participants were smiling."

[0579] Step 9:

[0580] When a user terminal joins a conference late, it requests the information from the server, and the request includes a timestamp of when the user accessed the conference.

[0581] Step 10:

[0582] Upon receiving a request from the user terminal, the server transmits the summarized discussion content, key points, and detailed emotional state to the user terminal in real time.

[0583] Step 11:

[0584] The user device displays the received information on the user interface. Specifically, it displays information such as "Agenda: New product development progress," "Important points: Scheduled for completion by March," and "Meeting atmosphere: Very positive." This allows the user to quickly understand the content of the meeting and smoothly participate in the discussion.

[0585] Through this series of processes, our system provides an environment in which users who join online meetings late can efficiently understand the content of the meeting and smoothly participate in the discussion.By using the emotion engine, the overall atmosphere and emotional state of the meeting can be grasped in detail, providing even more useful information.

[0586] Example 2

[0587] 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."

[0588] Late-joining online meetings often face challenges in quickly grasping the content of the meeting and smoothly participating in the discussion. Conventional systems are unable to accurately grasp the atmosphere and emotional state of the meeting, making it difficult to understand the overall emotional trends of all participants. Therefore, there is a need for a system that can not only extract key points and important insights from voice and chat data, but also perform emotion analysis, allowing late-joining users to quickly grasp the meeting situation.

[0589] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0590] In this invention, the server includes means for confirming the start of the online conference platform and acquiring audio data via streaming, speech recognition means for converting the audio data into text data, means for analyzing the converted text data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data using natural language processing to extract important comments, means for analyzing the audio data and chat data using emotion analysis to identify the emotional state of the conference, means for analyzing changes in facial expressions and tone of voice of conference participants to grasp their detailed emotional states, and means for providing the summarized main points, important points, and emotional states of the discussion to users who join late, thereby enabling users who join late to quickly grasp the content and atmosphere of the conference and smoothly participate in the discussion.

[0591] An "online conference platform" is a system that allows multiple users to communicate with each other using audio and video over the Internet.

[0592] "Audio data" refers to digitized data of audio information uttered by participants during a conference.

[0593] "Text data" is sentence data converted from voice data by voice recognition.

[0594] "Speech recognition" is a technology that analyzes voice data and converts its contents into text data.

[0595] "Natural language processing" is a technology that enables computers to understand, analyze, and generate human language.

[0596] "Discussion points" are the core of the important topics or content discussed during the meeting.

[0597] "Key points" are pieces of information or statements that deserve special attention during a meeting.

[0598] "Chat data" refers to text message data sent by participants during an online conference.

[0599] "Sentiment analysis" is a technology that analyzes text data and audio data and identifies emotional states from their content.

[0600] "Emotional state" indicates the current emotional state of the meeting participants (e.g., positive, negative, neutral).

[0601] "Facial expression changes" refers to judging emotions by analyzing the facial movements of meeting participants.

[0602] "Voice tone" is an element used to determine emotions by analyzing the pitch and strength of a speaker's voice.

[0603] A "user" is a person who participates in an online conference.

[0604] A "server" is a computer system that operates an online conference system and processes data.

[0605] A "user terminal" is a device such as a computer or smartphone that a user uses to participate in an online conference.

[0606] "Streaming" is a method of transmitting data continuously in real time.

[0607] This invention is a system that helps late-joining users in online conferences quickly understand the content of the discussion and smoothly participate in the conference. Processing is mainly carried out between the server, the terminal, and the user, and by combining it with an emotion engine that analyzes the facial expressions and tone of voice of conference participants, it is possible to grasp the atmosphere and emotional state of the conference in more detail.

[0608] Audio data processing

[0609] The server confirms that the meeting has started through the online meeting platform's API (e.g., Zoom API, Microsoft Teams API), and obtains streaming audio data. This audio data is converted into text data in real time using a speech recognition engine such as Google Cloud Speech-to-Text API. The converted text data is then sent to a natural language processing model such as BERT or GPT to extract the main points and key points of the discussion.

[0610] Processing chat data

[0611] The server collects chat data sent during online meetings in real time. The collected chat data is analyzed using a natural language processing model to extract important comments and topics. This data is integrated with summary information obtained from the voice data and provided to the user.

[0612] Conducting sentiment analysis

[0613] The server inputs the voice and chat data into emotion analysis models such as VADER and DeepMoji to analyze the emotional state of the meeting. The analysis results are classified as the meeting mood (positive, negative, or neutral), allowing users to quickly understand the situation of the meeting.

[0614] Introducing the Emotion Engine

[0615] The server uses an emotion engine to analyze the facial expressions and tone of voice of meeting participants. The emotion engine analyzes the video streaming data of the meeting to understand the individual emotional state of meeting participants. In addition, the emotion engine also performs tone analysis of the audio data to more precisely identify the overall emotional state of the meeting.

[0616] Real-time provision

[0617] When a user terminal joins a conference late, it requests information from the server. This request includes a timestamp of when the user accessed the conference. Upon receiving the user's request, the server transmits a summary of the discussion, key points, and detailed emotional state to the user terminal in real time. The user terminal displays the received information on a user interface to help the user quickly understand the situation of the conference.

[0618] Specific examples

[0619] For example, in a meeting initiated through an online conference platform, the server captures audio streaming and chat data in real time. A speech recognition engine is used to convert the speech "To discuss the development status of new products" into text. The converted text data is analyzed using a natural language processing model (GPT-4) to extract key points such as "new product development," "progress," and "marketing strategy."

[0620] At the same time, the server also collects chat data and analyzes it using a natural language processing model to extract important statements, such as "The new product development goal is until March" or "We need to review our marketing strategy." This data is then integrated with summary information from the voice data.

[0621] The server uses an emotion engine to analyze the video streaming data of the meeting and identify the emotional state of the meeting participants from their facial expressions and voice tones. For example, in the middle of the meeting, a detailed emotional state can be identified, such as "there was a very positive reaction to the development of the new product, and many of the participants were smiling."

[0622] When a user joins a meeting late, the user device sends a request to the server, and the server quickly sends summary information and emotional state to the user device. The user interface displays "Agenda: New product development progress," "Key points: Scheduled for completion by March," and "Meeting atmosphere: Very positive," allowing the user to quickly understand the meeting content and smoothly participate in the discussion.

[0623] Prompt Sentence Examples

[0624] "Summarize the minutes of the online meeting below and extract the key points. Also, please rate the mood of the meeting as positive, negative, or neutral."

[0625] In this way, this system provides an optimal means for participants in online meetings to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[0626] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0627] Step 1:

[0628] The server uses the API of the online meeting platform to confirm that the meeting has started. Specifically, it uses the Zoom API or Microsoft Teams API to obtain initial data such as the meeting ID and start time, and is then ready to start the audio streaming session.

[0629] Input: Meeting ID and start time obtained from the conference platform API

[0630] Output: Session information for conference streaming

[0631] Step 2:

[0632] The server receives the conference audio data in real time via streaming and converts it into text using the Google Cloud Speech-to-Text API. The audio data is converted in real time and saved as a temporary file.

[0633] Input: Conference streaming audio data

[0634] Output: Text data converted by Google Cloud Speech-to-Text API

[0635] Step 3:

[0636] The server then inputs the converted text data into a natural language processing model (e.g., BERT or GPT) to extract key points and key takeaways from the discussion, identifying the core points discussed during the meeting and storing them in a database.

[0637] Input: Text data converted by the Google Cloud Speech-to-Text API

[0638] Output: Data on gist and key points extracted by the NLP model

[0639] Step 4:

[0640] The server collects chat data using the chat API of the online meeting platform, and inputs the data into a natural language processing model for analysis to extract important comments and topics, thereby understanding the content and importance of the comments.

[0641] Input: Chat data during the meeting

[0642] Output: Data on important utterances and topics extracted by the natural language processing model

[0643] Step 5:

[0644] The server inputs the voice and chat data into a sentiment analysis model such as VADER or DeepMoji to analyze the emotional state of the meeting, classifying the mood of the meeting as positive, negative, or neutral, and storing this information in a database.

[0645] Input: Voice and chat data

[0646] Output: Data on emotional states (positive, negative, neutral) classified by the sentiment analysis model

[0647] Step 6:

[0648] The server inputs video streaming data into an emotion engine to analyze participants' facial expressions and vocal tones, thereby identifying their individual emotional states and providing a more detailed understanding of the overall emotional state of the meeting.

[0649] Input: Video streaming data from a conference

[0650] Output: Data on the individual and overall emotional states of participants analyzed by the emotion engine

[0651] Step 7:

[0652] When a user terminal joins a meeting late, it sends an information request to the server. This request includes a timestamp of the time of joining. The server receives the request and sends a summary of the discussion content, key points, and emotional state to the user terminal in real time. The user terminal displays the received information on a user interface to help the user quickly understand the situation of the meeting.

[0653] Input: Information request from user device (including timestamp)

[0654] Output: Summary information sent from the server, key points, emotional state data, and information displayed in the user interface

[0655] (Application example 2)

[0656] 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."

[0657] With conventional online conference systems, it was difficult for users who joined a meeting late to quickly grasp the content of the discussion. In particular, in store meetings, not being able to immediately grasp the key points and atmosphere of the meeting has a negative impact on work efficiency. Furthermore, there is a lack of means to accurately grasp the emotional state and atmosphere of the meeting, making it even more difficult to grasp the flow of the meeting. To solve these issues, there is a need for a system that can help store managers and staff who join late quickly grasp the content of employee meetings and strategy meetings and effectively apply it to their work.

[0658] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, and means for collecting chat data, analyzing the chat data, and extracting important comments. This makes it possible to analyze the voice data and chat data using sentiment analysis to identify the atmosphere of the meeting. Also included is means for acquiring voice data of a store meeting from a smart device in real time, providing a system that allows users who join a meeting late to quickly understand the situation and smoothly reflect it in their work.

[0659] "Audio data" refers to data that digitally represents information including the sounds of meetings and conversations.

[0660] "Text" is document data obtained by converting voice data into a character string.

[0661] "Natural language processing" is a technology that enables computers to understand, analyze, and generate human language.

[0662] "Key points" refer to the particularly important points or main content in a meeting or discussion.

[0663] "Key points" are those points in a discussion or statement that deserve special attention or priority.

[0664] "Chat data" refers to data used by participants to send and receive text messages during a conference.

[0665] "Sentiment analysis" is a technique for analyzing and identifying a speaker's emotional state from text or audio data.

[0666] "Analysis" is the process of examining and breaking down data to understand its structure and meaning.

[0667] A "server" is a computer system that manages and processes data on a network.

[0668] A "store meeting" is a conference or meeting held by employees and managers at a physical store.

[0669] "Smart devices" are electronic devices with advanced functions, such as smartphones and tablets.

[0670] A "summary" is a concise summary of a longer piece of writing or discussion.

[0671] A "user interface" is a display screen or operating means for interacting with a system.

[0672] The embodiment of this invention is a system that enables users who join an online conference late to quickly understand the conference content and smoothly participate in the discussion. This system is mainly composed of a server, user terminals, and smart devices.

[0673] The server collects voice data and converts it into text. This process uses the Google Cloud Speech-to-Text API as a speech recognition engine. The converted text data is passed through a natural language processing model (e.g., BERT or GPT) to extract the main points and important points of the discussion. The server also collects chat data sent during the meeting and analyzes it with the natural language processing model to extract important comments.

[0674] The server then analyzes the voice and chat data using a sentiment analysis model (e.g., VADER or DeepMoji) to classify the meeting mood as positive, negative, or neutral. This sentiment analysis makes it possible to understand the mood of the meeting and the emotional state of the participants.

[0675] This data is provided in real time to user devices such as smartphones and tablets, which display a summary of the main points of the discussion, key points, and the atmosphere of the meeting on their user interface, allowing late-joining users to quickly understand the current state of the meeting and participate in the discussion efficiently.

[0676] As a concrete example, consider the case where a store manager is late to an important store meeting. The manager opens an application on his smartphone and requests information about the past meeting. The server immediately sends a summary of the meeting content, key points, and emotional state to the manager's smartphone. The user interface displays "Agenda: New product development progress," "Key points: Expected completion by March," and "Meeting atmosphere: Very positive."

[0677] An example prompt might look like this:

[0678] "Please summarize the meeting. What were the key points and the emotional state of the participants?"

[0679] As described above, the system of the present invention supports users who join an online conference late to quickly grasp the content and atmosphere of the discussion at the conference.

[0680] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0681] Step 1:

[0682] The server collects voice data in real time at the start of an online meeting. Specifically, it uses the Google Cloud Speech-to-Text API, a speech recognition engine, to receive the audio of the online meeting as input. This voice data is received in streaming format and converted into text data.

[0683] Step 2:

[0684] The server sends the converted text data from the audio data to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion. This step uses the converted audio data as input. The natural language processing model analyzes the text data, identifies the main points and key points of the discussion, and outputs a summary.

[0685] Step 3:

[0686] The server collects chat data sent during online meetings in real time, receives chat messages as input, and stores their contents as text data.

[0687] Step 4:

[0688] The server analyzes the collected chat data using a natural language processing model to extract important comments. The text data of chat messages is used as input. The model extracts important comments and topics from the text data and integrates them into a summary.

[0689] Step 5:

[0690] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the mood of the meeting. This step uses voice and chat text data as input. The sentiment analysis model analyzes the data and classifies it as positive, negative, or neutral, and outputs the mood of the meeting.

[0691] Step 6:

[0692] When a user terminal joins a conference late, the user terminal sends a request to the server, sending request data including a timestamp when the user accesses the conference as input.

[0693] Step 7:

[0694] The server receives the user's request and transmits the summarized discussion summary, key points, and meeting atmosphere information to the user's terminal in real time. In this step, the server transmits the summarized summary, key points, and meeting atmosphere information as output.

[0695] Step 8:

[0696] The user terminal displays the received information on the user interface. In this step, the information received by the user is displayed in a visually easy-to-identify format, ultimately enabling the user to quickly understand the situation.

[0697] By following the steps above, users who join late can efficiently grasp the content and atmosphere of the meeting and smoothly reflect this in their work.

[0698] 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.

[0699] 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.

[0700] 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.

[0701] [Third embodiment]

[0702] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0703] 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.

[0704] 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).

[0705] 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.

[0706] 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.

[0707] 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).

[0708] 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.

[0709] 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.

[0710] 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.

[0711] 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.

[0712] 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.

[0713] 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."

[0714] The present invention is a system that helps users who join an online conference late quickly understand the content of the discussion and participate in the conference smoothly, and processing is mainly carried out between three parties: a server, a terminal, and the user.

[0715] Overview of program processing

[0716] Audio data processing

[0717] At the start of an online meeting, the server acquires audio data through the conference platform's API. This audio data is streamed in real time and converted into text data by a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The converted text data is sent to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion.

[0718] Processing chat data

[0719] The server collects chat messages sent during online meetings in real time. The collected chat data is analyzed to extract important comments and topics. This analysis is performed using natural language processing technology and is integrated with summary information obtained from the audio data.

[0720] Conducting sentiment analysis

[0721] The server inputs the acquired voice and chat data into an emotion analysis model (such as VADER or DeepMoji) to analyze the emotional state of the meeting participants, which then classifies the overall mood of the meeting as positive, negative, neutral, etc.

[0722] Real-time provision

[0723] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[0724] Specific examples

[0725] For example, in a meeting initiated through an online conferencing platform, the server captures the audio stream in real time and uses a speech recognition engine to convert the speech "To discuss this year's sales targets" into text. The converted text is then analyzed using a natural language processing model to extract key points such as "sales targets," "key dates," and "market strategy."

[0726] At the same time, the server also collects chat data and analyzes and extracts important statements such as "Sales target is 20% higher than last year" and "Market strategy needs to be reviewed." This allows summaries of the voice data and chat data to be integrated.

[0727] The sentiment analysis model analyzes that the atmosphere in the meeting is positive, and information such as "There are many positive opinions regarding the sales target" can be obtained.

[0728] When a request is sent from a user device joining late, the server immediately sends summary information, and the user interface displays "Agenda: This year's sales target," "Important points: Sales target is to increase by 20% from last year," and "Meeting atmosphere: Mostly positive opinions." This allows users to quickly understand the content of the discussion and smoothly participate in the meeting.

[0729] In this way, this system provides an optimal means for online meeting participants to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[0730] The processing flow will be explained below.

[0731] Step 1:

[0732] The server confirms that the meeting has started through the API of the online meeting platform and starts streaming audio data, thereby capturing all spoken audio during the meeting.

[0733] Step 2:

[0734] The server sends the acquired voice data to a speech recognition engine (e.g., Google Cloud Speech-to-Text API) in real time and converts the voice data into text data. During this process, the voice data is processed continuously without interruption.

[0735] Step 3:

[0736] The server inputs the converted text data into a natural language processing model (e.g., BERT or GPT) and performs various analyses to extract the main points and important points of the discussion.

[0737] Step 4:

[0738] The server simultaneously receives chat data from the conference platform and analyzes it in real time, extracting important comments and frequently occurring topics from the chat data and saving them as text data.

[0739] Step 5:

[0740] The server combines the summaries of the voice data and chat data, organizing them so that the flow of the discussion and the main points are clear.

[0741] Step 6:

[0742] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the mood of the meeting (positive, negative, neutral). The results of the sentiment analysis are used to determine the overall mood of the meeting.

[0743] Step 7:

[0744] When a user terminal joins a conference late, it requests the information from the server, and this request includes a timestamp of when the user accessed the conference.

[0745] Step 8:

[0746] Upon receiving a user request, the server sends a summary of the discussion, key points, and the atmosphere of the meeting to the user's terminal in real time, allowing the user to instantly grasp the status of the meeting.

[0747] Step 9:

[0748] The user device displays the received information on the user interface. Specifically, the "agenda," "important points," "meeting atmosphere," etc. are displayed on the screen. Based on this information, the user can quickly join the meeting and make comments and exchange opinions.

[0749] This series of processes provides a system that allows late-joining users to efficiently understand the content of the discussion and contribute to the smooth progress of the meeting.

[0750] Example 1

[0751] 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."

[0752] Users who join online meetings late often have difficulty quickly grasping the main points and key points of the discussion and smoothly participating in the meeting. This problem arises because it takes a lot of time and effort to understand the entire content of the meeting, and because the information they need to know is not provided appropriately. It is also important to understand the atmosphere and emotional state of the meeting, and if this is not possible, users will be unable to participate effectively in the discussion.

[0753] 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.

[0754] In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data, and extracting important comments, means for analyzing the voice data and the chat data using sentiment analysis to identify the atmosphere of the meeting, means for providing the summarized main points and important points of the discussion and the atmosphere of the meeting to participants, means for integrating the main points extracted using natural language processing technology with the summary of the chat data, means for displaying the summarized information on a user interface, means for acquiring voice data in real time from an online conference platform, and means for providing the summarized information in real time in response to a request from a user terminal joining the online conference late, thereby enabling even users who join late to quickly understand the content of the discussion and effectively participate in the conference.

[0755] "Audio data" refers to data that records participants' comments and discussions that occur during an online conference.

[0756] "Text data" is voice data converted into character information, and is data generated by a voice recognition engine.

[0757] "Natural language processing" is a technology for analyzing text data and extracting the main points and important points of an argument.

[0758] "Chat data" refers to text message data sent by participants during an online conference.

[0759] "Sentiment analysis" is a technology that analyzes the emotional state of participants based on voice and chat data to identify the atmosphere of a meeting.

[0760] "Summary information" is information that includes the main points of the discussion, important points, and the atmosphere of the meeting, extracted from the voice data and chat data.

[0761] "User interface" refers to functions such as a screen or operation panel that allow a user to interact with a system.

[0762] An "online conference platform" is a technological infrastructure that allows multiple participants to communicate in real time via the Internet.

[0763] "Real-time" refers to data being generated and processed almost simultaneously, with very little delay.

[0764] A "user terminal" is a device such as a computer or smartphone that a user uses to access the system and obtain information.

[0765] A "request" refers to a request from a user terminal to a server to obtain information.

[0766] The present invention is a system that supports users who join an online conference late by quickly understanding the content of the discussion and allowing them to smoothly participate in the conference. Processing is mainly carried out between three parties: a server, a terminal, and the user.

[0767] Acquiring and converting audio data

[0768] At the start of an online meeting, the server acquires audio data through the API of the meeting platform (e.g., commonly known as a meeting API). The acquired audio data is streamed in real time. The audio data is converted into text data using a speech recognition engine (e.g., commonly known as a speech recognition API). The converted text data is used to extract the main points of the discussion.

[0769] Text summarization and analysis

[0770] The server sends the converted text data to a natural language processing model (e.g., an NLP model), which extracts the main points and key points of the discussion, which are key to quickly understanding the content of the online meeting.

[0771] Chat data collection and analysis

[0772] The server collects chat messages sent during online meetings in real time. The collected chat data is also analyzed using a natural language processing model to extract important statements and topics. This data is then combined with summary information obtained from the audio data.

[0773] Conducting sentiment analysis

[0774] The server inputs the acquired voice and chat data into a sentiment analysis model (e.g., a sentiment analysis model) to analyze the emotional state of the meeting participants, which can then categorize the mood of the meeting into categories such as positive, negative, or neutral.

[0775] Real-time provision

[0776] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[0777] Specific examples

[0778] For example, in a meeting initiated through a conferencing platform (commonly known as a conferencing API), the server captures the audio stream in real time. A speech recognition engine is used to convert the speech "To discuss this year's sales targets" into text. The converted text is then analyzed using a natural language processing model to extract key points such as "sales targets," "key dates," and "market strategy."

[0779] At the same time, the server also collects chat data using the chat API of the conference platform, analyzing and extracting important statements such as "Sales target is 20% higher than last year" and "Market strategy needs to be reviewed." This allows summaries of the voice and chat data to be integrated.

[0780] Through the sentiment analysis model, it is analyzed that the atmosphere of the meeting is positive, and information such as "there are many positive opinions regarding sales targets" can be obtained.

[0781] When a request is sent from a user terminal joining late, the server immediately sends summary information, and the user interface displays "Agenda: This year's sales target," "Important points: Sales target is to increase by 20% from last year," and "Meeting atmosphere: Mostly positive opinions." This allows the user to quickly understand the content of the discussion and smoothly participate in the meeting.

[0782] Prompt Sentence Examples

[0783] The following prompt sentences are examples of inputs to a generative AI model to generate a summary of an online meeting.

[0784] "To discuss this year's sales targets"

[0785] "Sales target is 20% increase from last year"

[0786] "Market strategy needs to be reviewed"

[0787] Sentiment analysis results: "Mostly positive opinions"

[0788] Generated summary: "At the meeting, the sales target for this fiscal year was discussed. The sales target was set at a 20% increase from last year, and it was decided that a review of the market strategy was necessary. The opinions of the meeting participants were positive, with many constructive suggestions being offered."

[0789] As a result, the present invention can support users who join an online conference late to efficiently understand the content of the discussion and effectively participate in the conference.

[0790] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0791] Step 1:

[0792] When an online conference starts, the server acquires audio data through the conference platform's API (for example, the common name conference API). Specifically, it detects the conference's "start event" and acquires the audio stream using the API. The acquired audio data is temporarily stored in a buffer. The input is the audio stream from the conference platform's API, and the output is the audio data stored in the buffer.

[0793] Step 2:

[0794] The server sends the voice data stored in the buffer to a speech recognition engine (for example, a speech recognition API) and converts it into text data. Specifically, it divides the voice data into chunks of a certain size and sends each chunk to the speech recognition engine. The text data returned by the engine is then combined into a single document. The input is the voice data stored in the buffer, and the output is text data.

[0795] Step 3:

[0796] The server sends the converted text data to a natural language processing model (e.g., an NLP model) to extract the main points and key points of the discussion. Specifically, the text data is input into the natural language processing model, which extracts the main points and keywords of the discussion. The input is the text data, and the output is the extracted main points and keywords.

[0797] Step 4:

[0798] The server collects chat messages sent during online meetings in real time through a chat API. Specifically, it monitors the chat's "new message received event" and captures and logs each new message posted. The input is the real-time message from the chat API, and the output is the collected chat data.

[0799] Step 5:

[0800] The server sends the collected chat data to a natural language processing model (NLP model) to extract important comments and topics. The acquired chat data is also analyzed by the NLP model to extract important comments and topics. The input is chat data, and the output is important comments and topics.

[0801] Step 6:

[0802] The server integrates summary information from the voice data and summary information from the chat data. Specifically, it combines each element (main points of the discussion, important keywords, important comments) and formats them into a single summary. The input is summary information from the voice data and summary information from the chat data, and the output is the integrated summary information.

[0803] Step 7:

[0804] The server inputs the integrated summary information into a sentiment analysis model (e.g., a sentiment analysis model) to analyze the emotional states of the meeting participants. Specifically, the text data of the summary information is input into the sentiment analysis model to calculate a positive, negative, or neutral sentiment score. The input is the integrated summary information, and the output is a sentiment score and information about the atmosphere of the meeting.

[0805] Step 8:

[0806] When a user terminal joins a conference late, the user terminal sends an information request to the server. Specifically, when the "join late" button is pressed on the user interface of the user terminal, the information request is sent to the server. The input is a request from the user interface, and the output is a request to the server.

[0807] Step 9:

[0808] The server receives requests from user terminals and transmits the summarized discussion content, key points, and meeting atmosphere to the user terminals in real time. Specifically, it analyzes the requests and transmits the stored summary information to the user terminals in real time. The input is the request from the user terminal, and the output is the transmission of the summary information.

[0809] Step 10:

[0810] The user terminal displays the information received from the server on the user interface. Specifically, the received summary information is displayed at an appropriate position on the screen to allow the user to quickly understand it. The input is the summary information received from the server, and the output is the information displayed on the user interface.

[0811] Through the above processing steps, this system enables even late-joining users to quickly understand the content of the discussion and smoothly participate in the conference.

[0812] (Application example 1)

[0813] 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."

[0814] In online meetings, there is a need for a system that allows late-joining users to quickly understand the content of the discussion and participate smoothly in the meeting. However, current systems make it difficult for late-joining users to immediately understand the content of the discussion, resulting in a decrease in productivity. In addition, there is a lack of means to grasp the overall atmosphere of the meeting, including sentiment analysis. This means that there is an urgent need to develop a system that allows users to participate in the meeting without falling behind.

[0815] 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.

[0816] In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data, and extracting important comments, means for analyzing the voice data and the chat data using sentiment analysis to identify the atmosphere of the meeting, means for providing the summarized main points and important points of the discussion and the atmosphere of the meeting to participants, data acquisition means for acquiring the voice data and the chat data, means for performing on-demand processing in real time using the data acquisition means, and means for visually displaying the processed data on a user interface, thereby enabling users to quickly understand the content of the discussion and smoothly participate in the meeting even if they join late.

[0817] "Audio data" refers to a digital representation of audio signals such as speech captured during an online conference.

[0818] "Text" is information that is created by transcribing audio data and expressing it as text.

[0819] "Natural language processing" is a general term for technology that processes and analyzes human language using computers.

[0820] "Chat data" refers to data of messages sent as text chat during an online conference.

[0821] "Sentiment analysis" is a technique for analyzing and classifying emotional nuances from text and audio data.

[0822] "Data acquisition means" refers to a method or device for acquiring necessary voice data or chat data from a server or user terminal.

[0823] "On-demand processing" is processing that analyzes data in real time as needed and provides results.

[0824] "User interface" is a general term for the screens and input devices that allow a user to interact with a system.

[0825] "Real-time" refers to immediate processing or response with little or no delay.

[0826] "Discussion points" are key points such as main opinions and conclusions discussed during an online meeting.

[0827] "Important points" refer to statements or information that are particularly noteworthy during online meetings.

[0828] "Atmosphere" refers to the atmosphere or mood that reflects the emotions and reactions of all meeting participants.

[0829] This system summarizes the main points and important points of discussions in online meetings, helping participants who join late to quickly understand the situation of the meeting. The process is mainly carried out between the server, the terminal, and the user.

[0830] Audio data processing

[0831] At the start of an online meeting, the server captures real-time audio data through the conference platform's API. This audio data is converted into text using a speech recognition engine such as Google Cloud Speech-to-Text API. This text data is then analyzed using natural language processing models such as BERT and GPT to extract the main points and key points of the discussion.

[0832] Processing chat data

[0833] The server also collects chat data sent during online meetings in real time. This chat data is analyzed using natural language processing technology to extract important comments and topics. This information is then combined with summary information obtained from the audio data.

[0834] Conducting sentiment analysis

[0835] The server inputs the acquired voice and chat data into emotion analysis models such as VADER and DeepMoji to analyze the emotional state of the meeting participants, which then classifies the meeting atmosphere as positive, negative, neutral, etc.

[0836] Real-time provision

[0837] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal visually displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[0838] Specific examples

[0839] For example, in a meeting initiated through an online conference platform, the server captures voice data in real time. It uses a speech recognition engine to convert speech, such as "To discuss this year's sales target," into text. This converted text is then analyzed using a natural language processing model to extract key points. At the same time, the server also collects chat data and analyzes and extracts important statements, such as "The sales target is to be increased by 20% from last year." This process integrates summaries of the voice and chat data. Using a sentiment analysis model, it is determined that the meeting atmosphere is positive, resulting in information such as "There is a lot of positive feedback regarding the sales target." When a request is sent from a user device joining late, the server immediately sends summary information, displaying it on the user interface as "Agenda: This year's sales target," "Key points: The sales target is to be increased by 20% from last year," and "Meeting atmosphere: There is a lot of positive feedback." This allows users to quickly understand the content of the discussion and smoothly participate in the meeting.

[0840] Example prompt sentence:

[0841] Prompts for generating key points

[0842] """

[0843] Meeting audio data: {audio data}

[0844] Chat content: {Chat content}

[0845] Please prepare your abstract in the following format:

[0846] summary:

[0847] 1. Agenda: {Agenda}

[0848] 2. Key Points: {Key Points}

[0849] 3. Meeting atmosphere: {Meeting atmosphere}

[0850] """

[0851] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0852] Step 1:

[0853] Acquiring audio data

[0854] Subject: Server

[0855] The server acquires audio data in real time using the API of the conference platform at the start of the online conference, using the audio stream of the conference platform as input and obtaining audio data as output, which is stored in a digital format and used for subsequent processing.

[0856] Step 2:

[0857] Transcription of audio data

[0858] Subject: Server

[0859] The server converts the acquired voice data into text using the Google Cloud Speech-to-Text API. Voice data is used as input, and the converted text data is obtained as output. This text data is a written representation of what was said during the online meeting.

[0860] Step 3:

[0861] Summarizing text data

[0862] Subject: Server

[0863] The server sends the converted text data to a natural language processing model (such as BERT or GPT) to extract the main points and key points of the discussion. The model uses the text data as input and obtains summarized text data as output. The model automatically extracts important topics and key points from the meeting.

[0864] Step 4:

[0865] Acquiring and analyzing chat data

[0866] Subject: Server

[0867] The server collects chat data sent during online meetings in real time. It uses the chat data as input and extracts important statements and topics analyzed as output. This data is then analyzed using natural language processing techniques.

[0868] Step 5:

[0869] Sentiment analysis of voice and chat data

[0870] Subject: Server

[0871] The server inputs the acquired voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting participants. It uses voice and chat data as input and obtains emotional states (positive, negative, neutral) as output, thereby identifying the overall mood of the meeting.

[0872] Step 6:

[0873] Integration and visualization of summary information

[0874] Subject: Server

[0875] The server integrates the summarized voice and chat data, as well as the results of the sentiment analysis, and prepares the data for display on the system's user interface. It uses each of the aforementioned data as input and obtains integrated summary information and emotional states as output. This data is then formatted for visualization.

[0876] Step 7:

[0877] Real-time provision to user devices

[0878] Subject: User device

[0879] When a user terminal joins a meeting late, it requests information from the server. The server sends prepared summary information and emotional state to the user terminal in real time. The server receives the request as input and displays the summary information and emotional state as output on the user interface. This process allows the user to quickly understand the content of the discussion and smoothly participate in the meeting.

[0880] 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.

[0881] This system helps late-joining users in online meetings quickly understand the content of the discussion and smoothly participate in the meeting. Processing is mainly carried out between the server, the terminal, and the user. In addition, by combining it with an emotion engine that analyzes the facial expressions and tone of voice of meeting participants, the atmosphere and emotional state of the meeting can be grasped in more detail.

[0882] Overview of program processing

[0883] Audio data processing

[0884] The server confirms that the meeting has started through the online meeting platform's API and obtains streaming audio data. This audio data is converted into text data in real time using a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The converted text data is then sent to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion.

[0885] Processing chat data

[0886] The server collects chat data sent during online meetings in real time. The collected chat data is analyzed using a natural language processing model to extract important statements and topics. This data is then integrated with summary information obtained from the audio data.

[0887] Conducting sentiment analysis

[0888] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting. The analysis results are classified as the mood of the meeting (positive, negative, or neutral).

[0889] Introducing the Emotion Engine

[0890] The server uses an emotion engine to analyze the facial expressions and tone of voice of meeting participants. The emotion engine analyzes facial expression changes from video streaming data to grasp the individual emotional states of meeting participants. In addition, the emotion engine also performs tone analysis of audio data to improve the accuracy of the emotional state. This allows for more detailed identification of the emotional state of the entire meeting.

[0891] Real-time provision

[0892] When a user terminal joins a conference late, it requests the information from the server. This request includes a timestamp of when the user accessed the conference. Upon receiving the user's request, the server transmits a summary of the discussion content, key points, and detailed emotional state to the user terminal in real time. The user terminal displays the received information on a user interface, helping the user quickly grasp the situation of the conference.

[0893] Specific examples

[0894] For example, in a meeting initiated through an online conference platform, the server captures audio streaming and chat data in real time. It uses a speech recognition engine to convert the speech "To discuss the development status of new products" into text. The converted text data is then analyzed using a natural language processing model to extract key points such as "new product development," "progress," and "marketing strategy."

[0895] At the same time, the server also collects chat data and analyzes and extracts important statements, such as "The new product development goal is until March" and "We need to review our marketing strategy." This data is also integrated with summaries of the voice data.

[0896] The emotion engine analyzes video streaming data from meeting participants and identifies their emotional state from their facial expressions and vocal tones. For example, in the middle of a meeting, it can identify a detailed emotional state such as "there was a very positive reaction to the development of a new product, and many participants were smiling."

[0897] When a user joins a meeting late, the user's device sends a request to the server, which immediately sends summary information and emotional state. The user interface displays information such as "Agenda: New product development progress," "Key points: Scheduled for completion by March," and "Meeting atmosphere: Very positive." This allows the user to quickly understand the meeting content and smoothly participate in the discussion.

[0898] In this way, this system provides an optimal means for participants in online meetings to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[0899] The processing flow will be explained below.

[0900] Step 1:

[0901] The server confirms that the meeting has started through the API of the online meeting platform and begins streaming audio data, thereby collecting all spoken audio during the meeting in real time.

[0902] Step 2:

[0903] The server sends the acquired voice data in real time to a speech recognition engine (e.g., Google Cloud Speech-to-Text API) and converts the voice into text data. The speech recognition engine analyzes the voice signal and generates text data such as "To discuss the development status of new products."

[0904] Step 3:

[0905] The server inputs the converted text data into a natural language processing model (e.g., BERT or GPT) and analyzes the text data. Through the analysis, the main points and important points of the discussion, such as "new product development," "progress," and "marketing strategy," are extracted.

[0906] Step 4:

[0907] The server simultaneously collects chat data sent from the online meeting platform. This data includes text messages between participants. These messages are analyzed in real time to extract important statements, such as "The new product development goal is scheduled to be completed by March" and "The marketing strategy needs to be reconsidered."

[0908] Step 5:

[0909] The server integrates the summary information obtained through speech recognition with the important comments obtained through chat analysis, thereby organizing the overall discussion content and creating a single summary.

[0910] Step 6:

[0911] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting participants. As a result of the analysis, the mood of the meeting is classified as positive, negative, or neutral.

[0912] Step 7:

[0913] The server also uses an emotion engine to analyze video streaming data such as participants' facial expressions and tone of voice. The emotion engine analyzes participants' facial expressions during the meeting to identify their specific emotional states (e.g., joy, surprise, anger).

[0914] Step 8:

[0915] The server integrates facial expression analysis and voice tone analysis of the video streaming data to identify the overall emotional state of the meeting in detail, such as "There were many positive reactions in the middle of the meeting" or "Many participants were smiling."

[0916] Step 9:

[0917] When a user terminal joins a conference late, it requests the information from the server, and the request includes a timestamp of when the user accessed the conference.

[0918] Step 10:

[0919] Upon receiving a request from the user terminal, the server transmits the summarized discussion content, key points, and detailed emotional state to the user terminal in real time.

[0920] Step 11:

[0921] The user device displays the received information on the user interface. Specifically, it displays information such as "Agenda: New product development progress," "Important points: Scheduled for completion by March," and "Meeting atmosphere: Very positive." This allows the user to quickly understand the content of the meeting and smoothly participate in the discussion.

[0922] Through this series of processes, our system provides an environment in which users who join online meetings late can efficiently understand the content of the meeting and smoothly participate in the discussion.By using the emotion engine, the overall atmosphere and emotional state of the meeting can be grasped in detail, providing even more useful information.

[0923] Example 2

[0924] 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."

[0925] Late-joining online meetings often face challenges in quickly grasping the content of the meeting and smoothly participating in the discussion. Conventional systems are unable to accurately grasp the atmosphere and emotional state of the meeting, making it difficult to understand the overall emotional trends of all participants. Therefore, there is a need for a system that can not only extract key points and important insights from voice and chat data, but also perform emotion analysis, allowing late-joining users to quickly grasp the meeting situation.

[0926] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0927] In this invention, the server includes means for confirming the start of the online conference platform and acquiring audio data via streaming, speech recognition means for converting the audio data into text data, means for analyzing the converted text data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data using natural language processing to extract important comments, means for analyzing the audio data and chat data using emotion analysis to identify the emotional state of the conference, means for analyzing changes in facial expressions and tone of voice of conference participants to grasp their detailed emotional states, and means for providing the summarized main points, important points, and emotional states of the discussion to users who join late, thereby enabling users who join late to quickly grasp the content and atmosphere of the conference and smoothly participate in the discussion.

[0928] An "online conference platform" is a system that allows multiple users to communicate with each other using audio and video over the Internet.

[0929] "Audio data" refers to digitized data of audio information uttered by participants during a conference.

[0930] "Text data" is sentence data converted from voice data by voice recognition.

[0931] "Speech recognition" is a technology that analyzes voice data and converts its contents into text data.

[0932] "Natural language processing" is a technology that enables computers to understand, analyze, and generate human language.

[0933] "Discussion points" are the core of the important topics or content discussed during the meeting.

[0934] "Key points" are pieces of information or statements that deserve special attention during a meeting.

[0935] "Chat data" refers to text message data sent by participants during an online conference.

[0936] "Sentiment analysis" is a technology that analyzes text data and audio data and identifies emotional states from their content.

[0937] "Emotional state" indicates the current emotional state of the meeting participants (e.g., positive, negative, neutral).

[0938] "Facial expression changes" refers to judging emotions by analyzing the facial movements of meeting participants.

[0939] "Voice tone" is an element used to determine emotions by analyzing the pitch and strength of a speaker's voice.

[0940] A "user" is a person who participates in an online conference.

[0941] A "server" is a computer system that operates an online conference system and processes data.

[0942] A "user terminal" is a device such as a computer or smartphone that a user uses to participate in an online conference.

[0943] "Streaming" is a method of transmitting data continuously in real time.

[0944] This invention is a system that helps late-joining users in online conferences quickly understand the content of the discussion and smoothly participate in the conference. Processing is mainly carried out between the server, the terminal, and the user, and by combining it with an emotion engine that analyzes the facial expressions and tone of voice of conference participants, it is possible to grasp the atmosphere and emotional state of the conference in more detail.

[0945] Audio data processing

[0946] The server confirms that the meeting has started through the online meeting platform's API (e.g., Zoom API, Microsoft Teams API), and obtains streaming audio data. This audio data is converted into text data in real time using a speech recognition engine such as Google Cloud Speech-to-Text API. The converted text data is then sent to a natural language processing model such as BERT or GPT to extract the main points and key points of the discussion.

[0947] Processing chat data

[0948] The server collects chat data sent during online meetings in real time. The collected chat data is analyzed using a natural language processing model to extract important comments and topics. This data is integrated with summary information obtained from the voice data and provided to the user.

[0949] Conducting sentiment analysis

[0950] The server inputs the voice and chat data into emotion analysis models such as VADER and DeepMoji to analyze the emotional state of the meeting. The analysis results are classified as the meeting mood (positive, negative, or neutral), allowing users to quickly understand the situation of the meeting.

[0951] Introducing the Emotion Engine

[0952] The server uses an emotion engine to analyze the facial expressions and tone of voice of meeting participants. The emotion engine analyzes the video streaming data of the meeting to understand the individual emotional state of meeting participants. In addition, the emotion engine also performs tone analysis of the audio data to more precisely identify the overall emotional state of the meeting.

[0953] Real-time provision

[0954] When a user terminal joins a conference late, it requests information from the server. This request includes a timestamp of when the user accessed the conference. Upon receiving the user's request, the server transmits a summary of the discussion, key points, and detailed emotional state to the user terminal in real time. The user terminal displays the received information on a user interface to help the user quickly understand the situation of the conference.

[0955] Specific examples

[0956] For example, in a meeting initiated through an online conference platform, the server captures audio streaming and chat data in real time. A speech recognition engine is used to convert the speech "To discuss the development status of new products" into text. The converted text data is analyzed using a natural language processing model (GPT-4) to extract key points such as "new product development," "progress," and "marketing strategy."

[0957] At the same time, the server also collects chat data and analyzes it using a natural language processing model to extract important statements, such as "The new product development goal is until March" or "We need to review our marketing strategy." This data is then integrated with summary information from the voice data.

[0958] The server uses an emotion engine to analyze the video streaming data of the meeting and identify the emotional state of the meeting participants from their facial expressions and voice tones. For example, in the middle of the meeting, a detailed emotional state can be identified, such as "there was a very positive reaction to the development of the new product, and many of the participants were smiling."

[0959] When a user joins a meeting late, the user device sends a request to the server, and the server quickly sends summary information and emotional state to the user device. The user interface displays "Agenda: New product development progress," "Key points: Scheduled for completion by March," and "Meeting atmosphere: Very positive," allowing the user to quickly understand the meeting content and smoothly participate in the discussion.

[0960] Prompt Sentence Examples

[0961] "Summarize the minutes of the online meeting below and extract the key points. Also, please rate the mood of the meeting as positive, negative, or neutral."

[0962] In this way, this system provides an optimal means for participants in online meetings to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[0963] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0964] Step 1:

[0965] The server uses the API of the online meeting platform to confirm that the meeting has started. Specifically, it uses the Zoom API or Microsoft Teams API to obtain initial data such as the meeting ID and start time, and is then ready to start the audio streaming session.

[0966] Input: Meeting ID and start time obtained from the conference platform API

[0967] Output: Session information for conference streaming

[0968] Step 2:

[0969] The server receives the conference audio data in real time via streaming and converts it into text using the Google Cloud Speech-to-Text API. The audio data is converted in real time and saved as a temporary file.

[0970] Input: Conference streaming audio data

[0971] Output: Text data converted by Google Cloud Speech-to-Text API

[0972] Step 3:

[0973] The server then inputs the converted text data into a natural language processing model (e.g., BERT or GPT) to extract key points and key takeaways from the discussion, identifying the core points discussed during the meeting and storing them in a database.

[0974] Input: Text data converted by the Google Cloud Speech-to-Text API

[0975] Output: Data on gist and key points extracted by the NLP model

[0976] Step 4:

[0977] The server collects chat data using the chat API of the online meeting platform, and inputs the data into a natural language processing model for analysis to extract important comments and topics, thereby understanding the content and importance of the comments.

[0978] Input: Chat data during the meeting

[0979] Output: Data on important utterances and topics extracted by the natural language processing model

[0980] Step 5:

[0981] The server inputs the voice and chat data into a sentiment analysis model such as VADER or DeepMoji to analyze the emotional state of the meeting, classifying the mood of the meeting as positive, negative, or neutral, and storing this information in a database.

[0982] Input: Voice and chat data

[0983] Output: Data on emotional states (positive, negative, neutral) classified by the sentiment analysis model

[0984] Step 6:

[0985] The server inputs video streaming data into an emotion engine to analyze participants' facial expressions and vocal tones, thereby identifying their individual emotional states and providing a more detailed understanding of the overall emotional state of the meeting.

[0986] Input: Video streaming data from a conference

[0987] Output: Data on the individual and overall emotional states of participants analyzed by the emotion engine

[0988] Step 7:

[0989] When a user terminal joins a meeting late, it sends an information request to the server. This request includes a timestamp of the time of joining. The server receives the request and sends a summary of the discussion content, key points, and emotional state to the user terminal in real time. The user terminal displays the received information on a user interface to help the user quickly understand the situation of the meeting.

[0990] Input: Information request from user device (including timestamp)

[0991] Output: Summary information sent from the server, key points, emotional state data, and information displayed in the user interface

[0992] (Application example 2)

[0993] 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."

[0994] With conventional online conference systems, it was difficult for users who joined a meeting late to quickly grasp the content of the discussion. In particular, in store meetings, not being able to immediately grasp the key points and atmosphere of the meeting has a negative impact on work efficiency. Furthermore, there is a lack of means to accurately grasp the emotional state and atmosphere of the meeting, making it even more difficult to grasp the flow of the meeting. To solve these issues, there is a need for a system that can help store managers and staff who join late quickly grasp the content of employee meetings and strategy meetings and effectively apply it to their work.

[0995] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, and means for collecting chat data, analyzing the chat data, and extracting important comments. This makes it possible to analyze the voice data and chat data using sentiment analysis to identify the atmosphere of the meeting. Also included is means for acquiring voice data of a store meeting from a smart device in real time, providing a system that allows users who join a meeting late to quickly understand the situation and smoothly reflect it in their work.

[0996] "Audio data" refers to data that digitally represents information including the sounds of meetings and conversations.

[0997] "Text" is document data obtained by converting voice data into a character string.

[0998] "Natural language processing" is a technology that enables computers to understand, analyze, and generate human language.

[0999] "Key points" refer to the particularly important points or main content in a meeting or discussion.

[1000] "Key points" are those points in a discussion or statement that deserve special attention or priority.

[1001] "Chat data" refers to data used by participants to send and receive text messages during a conference.

[1002] "Sentiment analysis" is a technique for analyzing and identifying a speaker's emotional state from text or audio data.

[1003] "Analysis" is the process of examining and breaking down data to understand its structure and meaning.

[1004] A "server" is a computer system that manages and processes data on a network.

[1005] A "store meeting" is a conference or meeting held by employees and managers at a physical store.

[1006] "Smart devices" are electronic devices with advanced functions, such as smartphones and tablets.

[1007] A "summary" is a concise summary of a longer piece of writing or discussion.

[1008] A "user interface" is a display screen or operating means for interacting with a system.

[1009] The embodiment of this invention is a system that enables users who join an online conference late to quickly understand the conference content and smoothly participate in the discussion. This system is mainly composed of a server, user terminals, and smart devices.

[1010] The server collects voice data and converts it into text. This process uses the Google Cloud Speech-to-Text API as a speech recognition engine. The converted text data is passed through a natural language processing model (e.g., BERT or GPT) to extract the main points and important points of the discussion. The server also collects chat data sent during the meeting and analyzes it with the natural language processing model to extract important comments.

[1011] The server then analyzes the voice and chat data using a sentiment analysis model (e.g., VADER or DeepMoji) to classify the meeting mood as positive, negative, or neutral. This sentiment analysis makes it possible to understand the mood of the meeting and the emotional state of the participants.

[1012] This data is provided in real time to user devices such as smartphones and tablets, which display a summary of the main points of the discussion, key points, and the atmosphere of the meeting on their user interface, allowing late-joining users to quickly understand the current state of the meeting and participate in the discussion efficiently.

[1013] As a concrete example, consider the case where a store manager is late to an important store meeting. The manager opens an application on his smartphone and requests information about the past meeting. The server immediately sends a summary of the meeting content, key points, and emotional state to the manager's smartphone. The user interface displays "Agenda: New product development progress," "Key points: Expected completion by March," and "Meeting atmosphere: Very positive."

[1014] An example prompt might look like this:

[1015] "Please summarize the meeting. What were the key points and the emotional state of the participants?"

[1016] As described above, the system of the present invention supports users who join an online conference late to quickly grasp the content and atmosphere of the discussion at the conference.

[1017] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1018] Step 1:

[1019] The server collects voice data in real time at the start of an online meeting. Specifically, it uses the Google Cloud Speech-to-Text API, a speech recognition engine, to receive the audio of the online meeting as input. This voice data is received in streaming format and converted into text data.

[1020] Step 2:

[1021] The server sends the converted text data from the audio data to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion. This step uses the converted audio data as input. The natural language processing model analyzes the text data, identifies the main points and key points of the discussion, and outputs a summary.

[1022] Step 3:

[1023] The server collects chat data sent during online meetings in real time, receives chat messages as input, and stores their contents as text data.

[1024] Step 4:

[1025] The server analyzes the collected chat data using a natural language processing model to extract important comments. The text data of chat messages is used as input. The model extracts important comments and topics from the text data and integrates them into a summary.

[1026] Step 5:

[1027] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the mood of the meeting. This step uses voice and chat text data as input. The sentiment analysis model analyzes the data and classifies it as positive, negative, or neutral, and outputs the mood of the meeting.

[1028] Step 6:

[1029] When a user terminal joins a conference late, the user terminal sends a request to the server, sending request data including a timestamp when the user accesses the conference as input.

[1030] Step 7:

[1031] The server receives the user's request and transmits the summarized discussion summary, key points, and meeting atmosphere information to the user's terminal in real time. In this step, the server transmits the summarized summary, key points, and meeting atmosphere information as output.

[1032] Step 8:

[1033] The user terminal displays the received information on the user interface. In this step, the information received by the user is displayed in a visually easy-to-identify format, ultimately enabling the user to quickly understand the situation.

[1034] By following the steps above, users who join late can efficiently grasp the content and atmosphere of the meeting and smoothly reflect this in their work.

[1035] 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.

[1036] 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.

[1037] 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.

[1038] [Fourth embodiment]

[1039] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1040] 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.

[1041] 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).

[1042] 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.

[1043] 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.

[1044] 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).

[1045] 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.

[1046] 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.

[1047] 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.

[1048] 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.

[1049] 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.

[1050] 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.

[1051] 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."

[1052] The present invention is a system that helps users who join an online conference late quickly understand the content of the discussion and participate in the conference smoothly, and processing is mainly carried out between three parties: a server, a terminal, and the user.

[1053] Overview of program processing

[1054] Audio data processing

[1055] At the start of an online meeting, the server acquires audio data through the conference platform's API. This audio data is streamed in real time and converted into text data by a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The converted text data is sent to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion.

[1056] Processing chat data

[1057] The server collects chat messages sent during online meetings in real time. The collected chat data is analyzed to extract important comments and topics. This analysis is performed using natural language processing technology and is integrated with summary information obtained from the audio data.

[1058] Conducting sentiment analysis

[1059] The server inputs the acquired voice and chat data into an emotion analysis model (such as VADER or DeepMoji) to analyze the emotional state of the meeting participants, which then classifies the overall mood of the meeting as positive, negative, neutral, etc.

[1060] Real-time provision

[1061] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[1062] Specific examples

[1063] For example, in a meeting initiated through an online conferencing platform, the server captures the audio stream in real time and uses a speech recognition engine to convert the speech "To discuss this year's sales targets" into text. The converted text is then analyzed using a natural language processing model to extract key points such as "sales targets," "key dates," and "market strategy."

[1064] At the same time, the server also collects chat data and analyzes and extracts important statements such as "Sales target is 20% higher than last year" and "Market strategy needs to be reviewed." This allows summaries of the voice data and chat data to be integrated.

[1065] The sentiment analysis model analyzes that the atmosphere in the meeting is positive, and information such as "There are many positive opinions regarding the sales target" can be obtained.

[1066] When a request is sent from a user device joining late, the server immediately sends summary information, and the user interface displays "Agenda: This year's sales target," "Important points: Sales target is to increase by 20% from last year," and "Meeting atmosphere: Mostly positive opinions." This allows users to quickly understand the content of the discussion and smoothly participate in the meeting.

[1067] In this way, this system provides an optimal means for online meeting participants to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[1068] The processing flow will be explained below.

[1069] Step 1:

[1070] The server confirms that the meeting has started through the API of the online meeting platform and starts streaming audio data, thereby capturing all spoken audio during the meeting.

[1071] Step 2:

[1072] The server sends the acquired voice data to a speech recognition engine (e.g., Google Cloud Speech-to-Text API) in real time and converts the voice data into text data. During this process, the voice data is processed continuously without interruption.

[1073] Step 3:

[1074] The server inputs the converted text data into a natural language processing model (e.g., BERT or GPT) and performs various analyses to extract the main points and important points of the discussion.

[1075] Step 4:

[1076] The server simultaneously receives chat data from the conference platform and analyzes it in real time, extracting important comments and frequently occurring topics from the chat data and saving them as text data.

[1077] Step 5:

[1078] The server combines the summaries of the voice data and chat data, organizing them so that the flow of the discussion and the main points are clear.

[1079] Step 6:

[1080] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the mood of the meeting (positive, negative, neutral). The results of the sentiment analysis are used to determine the overall mood of the meeting.

[1081] Step 7:

[1082] When a user terminal joins a conference late, it requests the information from the server, and this request includes a timestamp of when the user accessed the conference.

[1083] Step 8:

[1084] Upon receiving a user request, the server sends a summary of the discussion, key points, and the atmosphere of the meeting to the user's terminal in real time, allowing the user to instantly grasp the status of the meeting.

[1085] Step 9:

[1086] The user device displays the received information on the user interface. Specifically, the "agenda," "important points," "meeting atmosphere," etc. are displayed on the screen. Based on this information, the user can quickly join the meeting and make comments and exchange opinions.

[1087] This series of processes provides a system that allows late-joining users to efficiently understand the content of the discussion and contribute to the smooth progress of the meeting.

[1088] Example 1

[1089] 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."

[1090] Users who join online meetings late often have difficulty quickly grasping the main points and key points of the discussion and smoothly participating in the meeting. This problem arises because it takes a lot of time and effort to understand the entire content of the meeting, and because the information they need to know is not provided appropriately. It is also important to understand the atmosphere and emotional state of the meeting, and if this is not possible, users will be unable to participate effectively in the discussion.

[1091] 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.

[1092] In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data, and extracting important comments, means for analyzing the voice data and the chat data using sentiment analysis to identify the atmosphere of the meeting, means for providing the summarized main points and important points of the discussion and the atmosphere of the meeting to participants, means for integrating the main points extracted using natural language processing technology with the summary of the chat data, means for displaying the summarized information on a user interface, means for acquiring voice data in real time from an online conference platform, and means for providing the summarized information in real time in response to a request from a user terminal joining the online conference late, thereby enabling even users who join late to quickly understand the content of the discussion and effectively participate in the conference.

[1093] "Audio data" refers to data that records participants' comments and discussions that occur during an online conference.

[1094] "Text data" is voice data converted into character information, and is data generated by a voice recognition engine.

[1095] "Natural language processing" is a technology for analyzing text data and extracting the main points and important points of an argument.

[1096] "Chat data" refers to text message data sent by participants during an online conference.

[1097] "Sentiment analysis" is a technology that analyzes the emotional state of participants based on voice and chat data to identify the atmosphere of a meeting.

[1098] "Summary information" is information that includes the main points of the discussion, important points, and the atmosphere of the meeting, extracted from the voice data and chat data.

[1099] "User interface" refers to functions such as a screen or operation panel that allow a user to interact with a system.

[1100] An "online conference platform" is a technological infrastructure that allows multiple participants to communicate in real time via the Internet.

[1101] "Real-time" refers to data being generated and processed almost simultaneously, with very little delay.

[1102] A "user terminal" is a device such as a computer or smartphone that a user uses to access the system and obtain information.

[1103] A "request" refers to a request from a user terminal to a server to obtain information.

[1104] The present invention is a system that supports users who join an online conference late by quickly understanding the content of the discussion and allowing them to smoothly participate in the conference. Processing is mainly carried out between three parties: a server, a terminal, and the user.

[1105] Acquiring and converting audio data

[1106] At the start of an online meeting, the server acquires audio data through the API of the meeting platform (e.g., commonly known as a meeting API). The acquired audio data is streamed in real time. The audio data is converted into text data using a speech recognition engine (e.g., commonly known as a speech recognition API). The converted text data is used to extract the main points of the discussion.

[1107] Text summarization and analysis

[1108] The server sends the converted text data to a natural language processing model (e.g., an NLP model), which extracts the main points and key points of the discussion, which are key to quickly understanding the content of the online meeting.

[1109] Chat data collection and analysis

[1110] The server collects chat messages sent during online meetings in real time. The collected chat data is also analyzed using a natural language processing model to extract important statements and topics. This data is then combined with summary information obtained from the audio data.

[1111] Conducting sentiment analysis

[1112] The server inputs the acquired voice and chat data into a sentiment analysis model (e.g., a sentiment analysis model) to analyze the emotional state of the meeting participants, which can then categorize the mood of the meeting into categories such as positive, negative, or neutral.

[1113] Real-time provision

[1114] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[1115] Specific examples

[1116] For example, in a meeting initiated through a conferencing platform (commonly known as a conferencing API), the server captures the audio stream in real time. A speech recognition engine is used to convert the speech "To discuss this year's sales targets" into text. The converted text is then analyzed using a natural language processing model to extract key points such as "sales targets," "key dates," and "market strategy."

[1117] At the same time, the server also collects chat data using the chat API of the conference platform, analyzing and extracting important statements such as "Sales target is 20% higher than last year" and "Market strategy needs to be reviewed." This allows summaries of the voice and chat data to be integrated.

[1118] Through the sentiment analysis model, it is analyzed that the atmosphere of the meeting is positive, and information such as "there are many positive opinions regarding sales targets" can be obtained.

[1119] When a request is sent from a user terminal joining late, the server immediately sends summary information, and the user interface displays "Agenda: This year's sales target," "Important points: Sales target is to increase by 20% from last year," and "Meeting atmosphere: Mostly positive opinions." This allows the user to quickly understand the content of the discussion and smoothly participate in the meeting.

[1120] Prompt Sentence Examples

[1121] The following prompt sentences are examples of inputs to a generative AI model to generate a summary of an online meeting.

[1122] "To discuss this year's sales targets"

[1123] "Sales target is 20% increase from last year"

[1124] "Market strategy needs to be reviewed"

[1125] Sentiment analysis results: "Mostly positive opinions"

[1126] Generated summary: "At the meeting, the sales target for this fiscal year was discussed. The sales target was set at a 20% increase from last year, and it was decided that a review of the market strategy was necessary. The opinions of the meeting participants were positive, with many constructive suggestions being offered."

[1127] As a result, the present invention can support users who join an online conference late to efficiently understand the content of the discussion and effectively participate in the conference.

[1128] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1129] Step 1:

[1130] When an online conference starts, the server acquires audio data through the conference platform's API (for example, the common name conference API). Specifically, it detects the conference's "start event" and acquires the audio stream using the API. The acquired audio data is temporarily stored in a buffer. The input is the audio stream from the conference platform's API, and the output is the audio data stored in the buffer.

[1131] Step 2:

[1132] The server sends the voice data stored in the buffer to a speech recognition engine (for example, a speech recognition API) and converts it into text data. Specifically, it divides the voice data into chunks of a certain size and sends each chunk to the speech recognition engine. The text data returned by the engine is then combined into a single document. The input is the voice data stored in the buffer, and the output is text data.

[1133] Step 3:

[1134] The server sends the converted text data to a natural language processing model (e.g., an NLP model) to extract the main points and key points of the discussion. Specifically, the text data is input into the natural language processing model, which extracts the main points and keywords of the discussion. The input is the text data, and the output is the extracted main points and keywords.

[1135] Step 4:

[1136] The server collects chat messages sent during online meetings in real time through a chat API. Specifically, it monitors the chat's "new message received event" and captures and logs each new message posted. The input is the real-time message from the chat API, and the output is the collected chat data.

[1137] Step 5:

[1138] The server sends the collected chat data to a natural language processing model (NLP model) to extract important comments and topics. The acquired chat data is also analyzed by the NLP model to extract important comments and topics. The input is chat data, and the output is important comments and topics.

[1139] Step 6:

[1140] The server integrates summary information from the voice data and summary information from the chat data. Specifically, it combines each element (main points of the discussion, important keywords, important comments) and formats them into a single summary. The input is summary information from the voice data and summary information from the chat data, and the output is the integrated summary information.

[1141] Step 7:

[1142] The server inputs the integrated summary information into a sentiment analysis model (e.g., a sentiment analysis model) to analyze the emotional states of the meeting participants. Specifically, the text data of the summary information is input into the sentiment analysis model to calculate a positive, negative, or neutral sentiment score. The input is the integrated summary information, and the output is a sentiment score and information about the atmosphere of the meeting.

[1143] Step 8:

[1144] When a user terminal joins a conference late, the user terminal sends an information request to the server. Specifically, when the "join late" button is pressed on the user interface of the user terminal, the information request is sent to the server. The input is a request from the user interface, and the output is a request to the server.

[1145] Step 9:

[1146] The server receives requests from user terminals and transmits the summarized discussion content, key points, and meeting atmosphere to the user terminals in real time. Specifically, it analyzes the requests and transmits the stored summary information to the user terminals in real time. The input is the request from the user terminal, and the output is the transmission of the summary information.

[1147] Step 10:

[1148] The user terminal displays the information received from the server on the user interface. Specifically, the received summary information is displayed at an appropriate position on the screen to allow the user to quickly understand it. The input is the summary information received from the server, and the output is the information displayed on the user interface.

[1149] Through the above processing steps, this system enables even late-joining users to quickly understand the content of the discussion and smoothly participate in the conference.

[1150] (Application example 1)

[1151] 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."

[1152] In online meetings, there is a need for a system that allows late-joining users to quickly understand the content of the discussion and participate smoothly in the meeting. However, current systems make it difficult for late-joining users to immediately understand the content of the discussion, resulting in a decrease in productivity. In addition, there is a lack of means to grasp the overall atmosphere of the meeting, including sentiment analysis. This means that there is an urgent need to develop a system that allows users to participate in the meeting without falling behind.

[1153] 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.

[1154] In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data, and extracting important comments, means for analyzing the voice data and the chat data using sentiment analysis to identify the atmosphere of the meeting, means for providing the summarized main points and important points of the discussion and the atmosphere of the meeting to participants, data acquisition means for acquiring the voice data and the chat data, means for performing on-demand processing in real time using the data acquisition means, and means for visually displaying the processed data on a user interface, thereby enabling users to quickly understand the content of the discussion and smoothly participate in the meeting even if they join late.

[1155] "Audio data" refers to a digital representation of audio signals such as speech captured during an online conference.

[1156] "Text" is information that is created by transcribing audio data and expressing it as text.

[1157] "Natural language processing" is a general term for technology that processes and analyzes human language using computers.

[1158] "Chat data" refers to data of messages sent as text chat during an online conference.

[1159] "Sentiment analysis" is a technique for analyzing and classifying emotional nuances from text and audio data.

[1160] "Data acquisition means" refers to a method or device for acquiring necessary voice data or chat data from a server or user terminal.

[1161] "On-demand processing" is processing that analyzes data in real time as needed and provides results.

[1162] "User interface" is a general term for the screens and input devices that allow a user to interact with a system.

[1163] "Real-time" refers to immediate processing or response with little or no delay.

[1164] "Discussion points" are key points such as main opinions and conclusions discussed during an online meeting.

[1165] "Important points" refer to statements or information that are particularly noteworthy during online meetings.

[1166] "Atmosphere" refers to the atmosphere or mood that reflects the emotions and reactions of all meeting participants.

[1167] This system summarizes the main points and important points of discussions in online meetings, helping participants who join late to quickly understand the situation of the meeting. The process is mainly carried out between the server, the terminal, and the user.

[1168] Audio data processing

[1169] At the start of an online meeting, the server captures real-time audio data through the conference platform's API. This audio data is converted into text using a speech recognition engine such as Google Cloud Speech-to-Text API. This text data is then analyzed using natural language processing models such as BERT and GPT to extract the main points and key points of the discussion.

[1170] Processing chat data

[1171] The server also collects chat data sent during online meetings in real time. This chat data is analyzed using natural language processing technology to extract important comments and topics. This information is then combined with summary information obtained from the audio data.

[1172] Conducting sentiment analysis

[1173] The server inputs the acquired voice and chat data into emotion analysis models such as VADER and DeepMoji to analyze the emotional state of the meeting participants, which then classifies the meeting atmosphere as positive, negative, neutral, etc.

[1174] Real-time provision

[1175] When a user terminal joins a meeting late, it requests the information from the server. The server transmits a summary of the discussion, key points, and the atmosphere of the meeting to the user terminal in real time. The user terminal visually displays the received information on the user interface, helping the user quickly understand the situation of the meeting.

[1176] Specific examples

[1177] For example, in a meeting initiated through an online conference platform, the server captures voice data in real time. It uses a speech recognition engine to convert speech, such as "To discuss this year's sales target," into text. This converted text is then analyzed using a natural language processing model to extract key points. At the same time, the server also collects chat data and analyzes and extracts important statements, such as "The sales target is to be increased by 20% from last year." This process integrates summaries of the voice and chat data. Using a sentiment analysis model, it is determined that the meeting atmosphere is positive, resulting in information such as "There is a lot of positive feedback regarding the sales target." When a request is sent from a user device joining late, the server immediately sends summary information, displaying it on the user interface as "Agenda: This year's sales target," "Key points: The sales target is to be increased by 20% from last year," and "Meeting atmosphere: There is a lot of positive feedback." This allows users to quickly understand the content of the discussion and smoothly participate in the meeting.

[1178] Example prompt sentence:

[1179] Prompts for generating key points

[1180] """

[1181] Meeting audio data: {audio data}

[1182] Chat content: {Chat content}

[1183] Please prepare your abstract in the following format:

[1184] summary:

[1185] 1. Agenda: {Agenda}

[1186] 2. Key Points: {Key Points}

[1187] 3. Meeting atmosphere: {Meeting atmosphere}

[1188] """

[1189] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1190] Step 1:

[1191] Acquiring audio data

[1192] Subject: Server

[1193] The server acquires audio data in real time using the API of the conference platform at the start of the online conference, using the audio stream of the conference platform as input and obtaining audio data as output, which is stored in a digital format and used for subsequent processing.

[1194] Step 2:

[1195] Transcription of audio data

[1196] Subject: Server

[1197] The server converts the acquired voice data into text using the Google Cloud Speech-to-Text API. Voice data is used as input, and the converted text data is obtained as output. This text data is a written representation of what was said during the online meeting.

[1198] Step 3:

[1199] Summarizing text data

[1200] Subject: Server

[1201] The server sends the converted text data to a natural language processing model (such as BERT or GPT) to extract the main points and key points of the discussion. The model uses the text data as input and obtains summarized text data as output. The model automatically extracts important topics and key points from the meeting.

[1202] Step 4:

[1203] Acquiring and analyzing chat data

[1204] Subject: Server

[1205] The server collects chat data sent during online meetings in real time. It uses the chat data as input and extracts important statements and topics analyzed as output. This data is then analyzed using natural language processing techniques.

[1206] Step 5:

[1207] Sentiment analysis of voice and chat data

[1208] Subject: Server

[1209] The server inputs the acquired voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting participants. It uses voice and chat data as input and obtains emotional states (positive, negative, neutral) as output, thereby identifying the overall mood of the meeting.

[1210] Step 6:

[1211] Integration and visualization of summary information

[1212] Subject: Server

[1213] The server integrates the summarized voice and chat data, as well as the results of the sentiment analysis, and prepares the data for display on the system's user interface. It uses each of the aforementioned data as input and obtains integrated summary information and emotional states as output. This data is then formatted for visualization.

[1214] Step 7:

[1215] Real-time provision to user devices

[1216] Subject: User device

[1217] When a user terminal joins a meeting late, it requests information from the server. The server sends prepared summary information and emotional state to the user terminal in real time. The server receives the request as input and displays the summary information and emotional state as output on the user interface. This process allows the user to quickly understand the content of the discussion and smoothly participate in the meeting.

[1218] 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.

[1219] This system helps late-joining users in online meetings quickly understand the content of the discussion and smoothly participate in the meeting. Processing is mainly carried out between the server, the terminal, and the user. In addition, by combining it with an emotion engine that analyzes the facial expressions and tone of voice of meeting participants, the atmosphere and emotional state of the meeting can be grasped in more detail.

[1220] Overview of program processing

[1221] Audio data processing

[1222] The server confirms that the meeting has started through the online meeting platform's API and obtains streaming audio data. This audio data is converted into text data in real time using a speech recognition engine (e.g., Google Cloud Speech-to-Text API). The converted text data is then sent to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion.

[1223] Processing chat data

[1224] The server collects chat data sent during online meetings in real time. The collected chat data is analyzed using a natural language processing model to extract important statements and topics. This data is then integrated with summary information obtained from the audio data.

[1225] Conducting sentiment analysis

[1226] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting. The analysis results are classified as the mood of the meeting (positive, negative, or neutral).

[1227] Introducing the Emotion Engine

[1228] The server uses an emotion engine to analyze the facial expressions and tone of voice of meeting participants. The emotion engine analyzes facial expression changes from video streaming data to grasp the individual emotional states of meeting participants. In addition, the emotion engine also performs tone analysis of audio data to improve the accuracy of the emotional state. This allows for more detailed identification of the emotional state of the entire meeting.

[1229] Real-time provision

[1230] When a user terminal joins a conference late, it requests the information from the server. This request includes a timestamp of when the user accessed the conference. Upon receiving the user's request, the server transmits a summary of the discussion content, key points, and detailed emotional state to the user terminal in real time. The user terminal displays the received information on a user interface, helping the user quickly grasp the situation of the conference.

[1231] Specific examples

[1232] For example, in a meeting initiated through an online conference platform, the server captures audio streaming and chat data in real time. It uses a speech recognition engine to convert the speech "To discuss the development status of new products" into text. The converted text data is then analyzed using a natural language processing model to extract key points such as "new product development," "progress," and "marketing strategy."

[1233] At the same time, the server also collects chat data and analyzes and extracts important statements, such as "The new product development goal is until March" and "We need to review our marketing strategy." This data is also integrated with summaries of the voice data.

[1234] The emotion engine analyzes video streaming data from meeting participants and identifies their emotional state from their facial expressions and vocal tones. For example, in the middle of a meeting, it can identify a detailed emotional state such as "there was a very positive reaction to the development of a new product, and many participants were smiling."

[1235] When a user joins a meeting late, the user's device sends a request to the server, which immediately sends summary information and emotional state. The user interface displays information such as "Agenda: New product development progress," "Key points: Scheduled for completion by March," and "Meeting atmosphere: Very positive." This allows the user to quickly understand the meeting content and smoothly participate in the discussion.

[1236] In this way, this system provides an optimal means for participants in online meetings to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[1237] The processing flow will be explained below.

[1238] Step 1:

[1239] The server confirms that the meeting has started through the API of the online meeting platform and begins streaming audio data, thereby collecting all spoken audio during the meeting in real time.

[1240] Step 2:

[1241] The server sends the acquired voice data in real time to a speech recognition engine (e.g., Google Cloud Speech-to-Text API) and converts the voice into text data. The speech recognition engine analyzes the voice signal and generates text data such as "To discuss the development status of new products."

[1242] Step 3:

[1243] The server inputs the converted text data into a natural language processing model (e.g., BERT or GPT) and analyzes the text data. Through the analysis, the main points and important points of the discussion, such as "new product development," "progress," and "marketing strategy," are extracted.

[1244] Step 4:

[1245] The server simultaneously collects chat data sent from the online meeting platform. This data includes text messages between participants. These messages are analyzed in real time to extract important statements, such as "The new product development goal is scheduled to be completed by March" and "The marketing strategy needs to be reconsidered."

[1246] Step 5:

[1247] The server integrates the summary information obtained through speech recognition with the important comments obtained through chat analysis, thereby organizing the overall discussion content and creating a single summary.

[1248] Step 6:

[1249] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the emotional state of the meeting participants. As a result of the analysis, the mood of the meeting is classified as positive, negative, or neutral.

[1250] Step 7:

[1251] The server also uses an emotion engine to analyze video streaming data such as participants' facial expressions and tone of voice. The emotion engine analyzes participants' facial expressions during the meeting to identify their specific emotional states (e.g., joy, surprise, anger).

[1252] Step 8:

[1253] The server integrates facial expression analysis and voice tone analysis of the video streaming data to identify the overall emotional state of the meeting in detail, such as "There were many positive reactions in the middle of the meeting" or "Many participants were smiling."

[1254] Step 9:

[1255] When a user terminal joins a conference late, it requests the information from the server, and the request includes a timestamp of when the user accessed the conference.

[1256] Step 10:

[1257] Upon receiving a request from the user terminal, the server transmits the summarized discussion content, key points, and detailed emotional state to the user terminal in real time.

[1258] Step 11:

[1259] The user device displays the received information on the user interface. Specifically, it displays information such as "Agenda: New product development progress," "Important points: Scheduled for completion by March," and "Meeting atmosphere: Very positive." This allows the user to quickly understand the content of the meeting and smoothly participate in the discussion.

[1260] Through this series of processes, our system provides an environment in which users who join online meetings late can efficiently understand the content of the meeting and smoothly participate in the discussion.By using the emotion engine, the overall atmosphere and emotional state of the meeting can be grasped in detail, providing even more useful information.

[1261] Example 2

[1262] 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."

[1263] Late-joining online meetings often face challenges in quickly grasping the content of the meeting and smoothly participating in the discussion. Conventional systems are unable to accurately grasp the atmosphere and emotional state of the meeting, making it difficult to understand the overall emotional trends of all participants. Therefore, there is a need for a system that can not only extract key points and important insights from voice and chat data, but also perform emotion analysis, allowing late-joining users to quickly grasp the meeting situation.

[1264] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1265] In this invention, the server includes means for confirming the start of the online conference platform and acquiring audio data via streaming, speech recognition means for converting the audio data into text data, means for analyzing the converted text data using natural language processing to extract the main points and important points of the discussion, means for collecting chat data, analyzing the chat data using natural language processing to extract important comments, means for analyzing the audio data and chat data using emotion analysis to identify the emotional state of the conference, means for analyzing changes in facial expressions and tone of voice of conference participants to grasp their detailed emotional states, and means for providing the summarized main points, important points, and emotional states of the discussion to users who join late, thereby enabling users who join late to quickly grasp the content and atmosphere of the conference and smoothly participate in the discussion.

[1266] An "online conference platform" is a system that allows multiple users to communicate with each other using audio and video over the Internet.

[1267] "Audio data" refers to digitized data of audio information uttered by participants during a conference.

[1268] "Text data" is sentence data converted from voice data by voice recognition.

[1269] "Speech recognition" is a technology that analyzes voice data and converts its contents into text data.

[1270] "Natural language processing" is a technology that enables computers to understand, analyze, and generate human language.

[1271] "Discussion points" are the core of the important topics or content discussed during the meeting.

[1272] "Key points" are pieces of information or statements that deserve special attention during a meeting.

[1273] "Chat data" refers to text message data sent by participants during an online conference.

[1274] "Sentiment analysis" is a technology that analyzes text data and audio data and identifies emotional states from their content.

[1275] "Emotional state" indicates the current emotional state of the meeting participants (e.g., positive, negative, neutral).

[1276] "Facial expression changes" refers to judging emotions by analyzing the facial movements of meeting participants.

[1277] "Voice tone" is an element used to determine emotions by analyzing the pitch and strength of a speaker's voice.

[1278] A "user" is a person who participates in an online conference.

[1279] A "server" is a computer system that operates an online conference system and processes data.

[1280] A "user terminal" is a device such as a computer or smartphone that a user uses to participate in an online conference.

[1281] "Streaming" is a method of transmitting data continuously in real time.

[1282] This invention is a system that helps late-joining users in online conferences quickly understand the content of the discussion and smoothly participate in the conference. Processing is mainly carried out between the server, the terminal, and the user, and by combining it with an emotion engine that analyzes the facial expressions and tone of voice of conference participants, it is possible to grasp the atmosphere and emotional state of the conference in more detail.

[1283] Audio data processing

[1284] The server confirms that the meeting has started through the online meeting platform's API (e.g., Zoom API, Microsoft Teams API), and obtains streaming audio data. This audio data is converted into text data in real time using a speech recognition engine such as Google Cloud Speech-to-Text API. The converted text data is then sent to a natural language processing model such as BERT or GPT to extract the main points and key points of the discussion.

[1285] Processing chat data

[1286] The server collects chat data sent during online meetings in real time. The collected chat data is analyzed using a natural language processing model to extract important comments and topics. This data is integrated with summary information obtained from the voice data and provided to the user.

[1287] Conducting sentiment analysis

[1288] The server inputs the voice and chat data into emotion analysis models such as VADER and DeepMoji to analyze the emotional state of the meeting. The analysis results are classified as the meeting mood (positive, negative, or neutral), allowing users to quickly understand the situation of the meeting.

[1289] Introducing the Emotion Engine

[1290] The server uses an emotion engine to analyze the facial expressions and tone of voice of meeting participants. The emotion engine analyzes the video streaming data of the meeting to understand the individual emotional state of meeting participants. In addition, the emotion engine also performs tone analysis of the audio data to more precisely identify the overall emotional state of the meeting.

[1291] Real-time provision

[1292] When a user terminal joins a conference late, it requests information from the server. This request includes a timestamp of when the user accessed the conference. Upon receiving the user's request, the server transmits a summary of the discussion, key points, and detailed emotional state to the user terminal in real time. The user terminal displays the received information on a user interface to help the user quickly understand the situation of the conference.

[1293] Specific examples

[1294] For example, in a meeting initiated through an online conference platform, the server captures audio streaming and chat data in real time. A speech recognition engine is used to convert the speech "To discuss the development status of new products" into text. The converted text data is analyzed using a natural language processing model (GPT-4) to extract key points such as "new product development," "progress," and "marketing strategy."

[1295] At the same time, the server also collects chat data and analyzes it using a natural language processing model to extract important statements, such as "The new product development goal is until March" or "We need to review our marketing strategy." This data is then integrated with summary information from the voice data.

[1296] The server uses an emotion engine to analyze the video streaming data of the meeting and identify the emotional state of the meeting participants from their facial expressions and voice tones. For example, in the middle of the meeting, a detailed emotional state can be identified, such as "there was a very positive reaction to the development of the new product, and many of the participants were smiling."

[1297] When a user joins a meeting late, the user device sends a request to the server, and the server quickly sends summary information and emotional state to the user device. The user interface displays "Agenda: New product development progress," "Key points: Scheduled for completion by March," and "Meeting atmosphere: Very positive," allowing the user to quickly understand the meeting content and smoothly participate in the discussion.

[1298] Prompt Sentence Examples

[1299] "Summarize the minutes of the online meeting below and extract the key points. Also, please rate the mood of the meeting as positive, negative, or neutral."

[1300] In this way, this system provides an optimal means for participants in online meetings to efficiently understand the content of the discussion even if they are late, thereby improving productivity.

[1301] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1302] Step 1:

[1303] The server uses the API of the online meeting platform to confirm that the meeting has started. Specifically, it uses the Zoom API or Microsoft Teams API to obtain initial data such as the meeting ID and start time, and is then ready to start the audio streaming session.

[1304] Input: Meeting ID and start time obtained from the conference platform API

[1305] Output: Session information for conference streaming

[1306] Step 2:

[1307] The server receives the conference audio data in real time via streaming and converts it into text using the Google Cloud Speech-to-Text API. The audio data is converted in real time and saved as a temporary file.

[1308] Input: Conference streaming audio data

[1309] Output: Text data converted by Google Cloud Speech-to-Text API

[1310] Step 3:

[1311] The server then inputs the converted text data into a natural language processing model (e.g., BERT or GPT) to extract key points and key takeaways from the discussion, identifying the core points discussed during the meeting and storing them in a database.

[1312] Input: Text data converted by the Google Cloud Speech-to-Text API

[1313] Output: Data on gist and key points extracted by the NLP model

[1314] Step 4:

[1315] The server collects chat data using the chat API of the online meeting platform, and inputs the data into a natural language processing model for analysis to extract important comments and topics, thereby understanding the content and importance of the comments.

[1316] Input: Chat data during the meeting

[1317] Output: Data on important utterances and topics extracted by the natural language processing model

[1318] Step 5:

[1319] The server inputs the voice and chat data into a sentiment analysis model such as VADER or DeepMoji to analyze the emotional state of the meeting, classifying the mood of the meeting as positive, negative, or neutral, and storing this information in a database.

[1320] Input: Voice and chat data

[1321] Output: Data on emotional states (positive, negative, neutral) classified by the sentiment analysis model

[1322] Step 6:

[1323] The server inputs video streaming data into an emotion engine to analyze participants' facial expressions and vocal tones, thereby identifying their individual emotional states and providing a more detailed understanding of the overall emotional state of the meeting.

[1324] Input: Video streaming data from a conference

[1325] Output: Data on the individual and overall emotional states of participants analyzed by the emotion engine

[1326] Step 7:

[1327] When a user terminal joins a meeting late, it sends an information request to the server. This request includes a timestamp of the time of joining. The server receives the request and sends a summary of the discussion content, key points, and emotional state to the user terminal in real time. The user terminal displays the received information on a user interface to help the user quickly understand the situation of the meeting.

[1328] Input: Information request from user device (including timestamp)

[1329] Output: Summary information sent from the server, key points, emotional state data, and information displayed in the user interface

[1330] (Application example 2)

[1331] 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."

[1332] With conventional online conference systems, it was difficult for users who joined a meeting late to quickly grasp the content of the discussion. In particular, in store meetings, not being able to immediately grasp the key points and atmosphere of the meeting has a negative impact on work efficiency. Furthermore, there is a lack of means to accurately grasp the emotional state and atmosphere of the meeting, making it even more difficult to grasp the flow of the meeting. To solve these issues, there is a need for a system that can help store managers and staff who join late quickly grasp the content of employee meetings and strategy meetings and effectively apply it to their work.

[1333] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for collecting voice data and converting the voice data into text, means for analyzing the text data converted from the voice data using natural language processing to extract the main points and important points of the discussion, and means for collecting chat data, analyzing the chat data, and extracting important comments. This makes it possible to analyze the voice data and chat data using sentiment analysis to identify the atmosphere of the meeting. Also included is means for acquiring voice data of a store meeting from a smart device in real time, providing a system that allows users who join a meeting late to quickly understand the situation and smoothly reflect it in their work.

[1334] "Audio data" refers to data that digitally represents information including the sounds of meetings and conversations.

[1335] "Text" is document data obtained by converting voice data into a character string.

[1336] "Natural language processing" is a technology that enables computers to understand, analyze, and generate human language.

[1337] "Key points" refer to the particularly important points or main content in a meeting or discussion.

[1338] "Key points" are those points in a discussion or statement that deserve special attention or priority.

[1339] "Chat data" refers to data used by participants to send and receive text messages during a conference.

[1340] "Sentiment analysis" is a technique for analyzing and identifying a speaker's emotional state from text or audio data.

[1341] "Analysis" is the process of examining and breaking down data to understand its structure and meaning.

[1342] A "server" is a computer system that manages and processes data on a network.

[1343] A "store meeting" is a conference or meeting held by employees and managers at a physical store.

[1344] "Smart devices" are electronic devices with advanced functions, such as smartphones and tablets.

[1345] A "summary" is a concise summary of a longer piece of writing or discussion.

[1346] A "user interface" is a display screen or operating means for interacting with a system.

[1347] The embodiment of this invention is a system that enables users who join an online conference late to quickly understand the conference content and smoothly participate in the discussion. This system is mainly composed of a server, user terminals, and smart devices.

[1348] The server collects voice data and converts it into text. This process uses the Google Cloud Speech-to-Text API as a speech recognition engine. The converted text data is passed through a natural language processing model (e.g., BERT or GPT) to extract the main points and important points of the discussion. The server also collects chat data sent during the meeting and analyzes it with the natural language processing model to extract important comments.

[1349] The server then analyzes the voice and chat data using a sentiment analysis model (e.g., VADER or DeepMoji) to classify the meeting mood as positive, negative, or neutral. This sentiment analysis makes it possible to understand the mood of the meeting and the emotional state of the participants.

[1350] This data is provided in real time to user devices such as smartphones and tablets, which display a summary of the main points of the discussion, key points, and the atmosphere of the meeting on their user interface, allowing late-joining users to quickly understand the current state of the meeting and participate in the discussion efficiently.

[1351] As a concrete example, consider the case where a store manager is late to an important store meeting. The manager opens an application on his smartphone and requests information about the past meeting. The server immediately sends a summary of the meeting content, key points, and emotional state to the manager's smartphone. The user interface displays "Agenda: New product development progress," "Key points: Expected completion by March," and "Meeting atmosphere: Very positive."

[1352] An example prompt might look like this:

[1353] "Please summarize the meeting. What were the key points and the emotional state of the participants?"

[1354] As described above, the system of the present invention supports users who join an online conference late to quickly grasp the content and atmosphere of the discussion at the conference.

[1355] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1356] Step 1:

[1357] The server collects voice data in real time at the start of an online meeting. Specifically, it uses the Google Cloud Speech-to-Text API, a speech recognition engine, to receive the audio of the online meeting as input. This voice data is received in streaming format and converted into text data.

[1358] Step 2:

[1359] The server sends the converted text data from the audio data to a natural language processing model (e.g., BERT or GPT) to extract the main points and key points of the discussion. This step uses the converted audio data as input. The natural language processing model analyzes the text data, identifies the main points and key points of the discussion, and outputs a summary.

[1360] Step 3:

[1361] The server collects chat data sent during online meetings in real time, receives chat messages as input, and stores their contents as text data.

[1362] Step 4:

[1363] The server analyzes the collected chat data using a natural language processing model to extract important comments. The text data of chat messages is used as input. The model extracts important comments and topics from the text data and integrates them into a summary.

[1364] Step 5:

[1365] The server inputs the voice and chat data into a sentiment analysis model (e.g., VADER or DeepMoji) to analyze the mood of the meeting. This step uses voice and chat text data as input. The sentiment analysis model analyzes the data and classifies it as positive, negative, or neutral, and outputs the mood of the meeting.

[1366] Step 6:

[1367] When a user terminal joins a conference late, the user terminal sends a request to the server, sending request data including a timestamp when the user accesses the conference as input.

[1368] Step 7:

[1369] The server receives the user's request and transmits the summarized discussion summary, key points, and meeting atmosphere information to the user's terminal in real time. In this step, the server transmits the summarized summary, key points, and meeting atmosphere information as output.

[1370] Step 8:

[1371] The user terminal displays the received information on the user interface. In this step, the information received by the user is displayed in a visually easy-to-identify format, ultimately enabling the user to quickly understand the situation.

[1372] By following the steps above, users who join late can efficiently grasp the content and atmosphere of the meeting and smoothly reflect this in their work.

[1373] 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.

[1374] 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.

[1375] 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.

[1376] 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.

[1377] 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.

[1378] 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.

[1379] 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).

[1380] 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.

[1381] 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."

[1382] 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.

[1383] 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).

[1384] 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.

[1385] 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.

[1386] 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.

[1387] 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.

[1388] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1389] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1390] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1391] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1392] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1393] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1394] The following is further disclosed regarding the above embodiment.

[1395] (Claim 1)

[1396] To summarize the main points and key points of discussions in online meetings when joining late,

[1397] means for collecting voice data and converting said voice data into text;

[1398] a means for analyzing the text data converted from the voice data by natural language processing and extracting the main points and important points of the discussion;

[1399] A means for collecting chat data, analyzing the chat data, and extracting important comments;

[1400] means for analyzing the voice data and the chat data by sentiment analysis to identify the atmosphere of the conference;

[1401] and means for providing the summarized discussion points, key points, and atmosphere of the meeting to participants.

[1402] (Claim 2)

[1403] 2. The system according to claim 1, wherein the audio data is collected in real time at the start of an online conference.

[1404] (Claim 3)

[1405] 2. The system according to claim 1, wherein the text data analyzed by the natural language processing is integrated with a summary of the chat data.

[1406] "Example 1"

[1407] (Claim 1)

[1408] To summarize the main points and key points of discussions in online meetings when joining late,

[1409] means for collecting voice data and converting said voice data into text;

[1410] a means for analyzing the text data converted from the voice data by natural language processing and extracting the main points and important points of the discussion;

[1411] A means for collecting chat data, analyzing the chat data, and extracting important comments;

[1412] means for analyzing the voice data and the chat data by sentiment analysis to identify the atmosphere of the conference;

[1413] a means for providing participants with the summarized discussion points, key points, and atmosphere of the meeting;

[1414] a means for integrating the summary of chat data with the key points extracted by natural language processing technology;

[1415] means for displaying the summarized information in a user interface;

[1416] A means for acquiring audio data in real time from an online conference platform;

[1417] means for providing the summarized information in real time in response to a request from a user terminal that joins the online conference late;

[1418] A system including:

[1419] (Claim 2)

[1420] 2. The system according to claim 1, wherein the audio data is collected in real time at the start of an online conference.

[1421] (Claim 3)

[1422] 2. The system according to claim 1, wherein the text data analyzed by the natural language processing is integrated with a summary of the chat data.

[1423] "Application Example 1"

[1424] (Claim 1)

[1425] To summarize the main points and key points of discussions in online meetings when joining late,

[1426] means for collecting voice data and converting said voice data into text;

[1427] a means for analyzing the text data converted from the voice data by natural language processing and extracting the main points and important points of the discussion;

[1428] A means for collecting chat data, analyzing the chat data, and extracting important comments;

[1429] means for analyzing the voice data and the chat data by sentiment analysis to identify the atmosphere of the conference;

[1430] a means for providing participants with the summarized discussion points, key points, and atmosphere of the meeting;

[1431] data acquisition means for acquiring the voice data and the chat data;

[1432] means for performing on-demand processing in real time using the data acquisition means;

[1433] means for visually displaying the processed data on a user interface;

[1434] A system including:

[1435] (Claim 2)

[1436] 2. The system according to claim 1, wherein the audio data is collected in real time at the start of an online conference.

[1437] (Claim 3)

[1438] 2. The system according to claim 1, wherein the text data analyzed by the natural language processing and a summary of the chat data are integrated, and the results of the sentiment analysis are visually displayed.

[1439] "Example 2: Combining Emotion Engines"

[1440] (Claim 1)

[1441] A means to confirm the start of the online conference platform and obtain streaming audio data,

[1442] a voice recognition means for converting the voice data into text data;

[1443] A means for analyzing the converted text data using natural language processing and extracting the main points and important points of the discussion;

[1444] a means for collecting chat data, analyzing the chat data by natural language processing, and extracting important comments;

[1445] means for analyzing the voice data and the chat data by sentiment analysis to identify the emotional state of the conference;

[1446] A means to analyze changes in facial expressions and tone of voice of meeting participants to grasp their detailed emotional state,

[1447] and means for providing the summarized discussion points, key points, and emotional state to late joining users.

[1448] (Claim 2)

[1449] 2. The system according to claim 1, wherein the audio data is collected in real time at the start of an online conference.

[1450] (Claim 3)

[1451] 2. The system according to claim 1, wherein the text data analyzed by the natural language processing is integrated with a summary of the chat data.

[1452] "Application example 2 when combining emotion engines"

[1453] (Claim 1)

[1454] To summarize the main points and key points of discussions in online meetings when joining late,

[1455] means for collecting voice data and converting said voice data into text;

[1456] a means for analyzing the text data converted from the voice data by natural language processing and extracting the main points and important points of the discussion;

[1457] A means for collecting chat data, analyzing the chat data, and extracting important comments;

[1458] means for analyzing the voice data and the chat data by sentiment analysis to identify the atmosphere of the conference;

[1459] A means of acquiring store meeting audio data from a smart device in real time;

[1460] and means for providing the summarized discussion points, key points, and atmosphere of the meeting to participants.

[1461] (Claim 2)

[1462] 2. The system according to claim 1, wherein the audio data is collected in real time at the start of an online conference.

[1463] (Claim 3)

[1464] 2. The system according to claim 1, wherein the text data analyzed by the natural language processing is integrated with a summary of the chat data. [Explanation of symbols]

[1465] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. To summarize the main points and key points of an online meeting when you join late, means for collecting voice data and converting said voice data into text; a means for analyzing the text data converted from the voice data by natural language processing and extracting the main points and important points of the discussion; means for collecting chat data, analyzing the chat data, and extracting important comments; means for analyzing the voice data and the chat data by sentiment analysis to identify the atmosphere of the conference; and means for providing the summarized discussion points, key points, and atmosphere of the meeting to participants.

2. 2. The system according to claim 1, wherein the audio data is collected in real time at the start of an online conference.

3. The system according to claim 1 , wherein the text data analyzed by the natural language processing is integrated with a summary of the chat data.

Images