Conference support system, conference support method, and program

Abstract

According to the present invention, a meeting support system is provided that provides support for meetings in accordance with the progress of the meeting. [Solution] The meeting support system in this embodiment includes a scenario storage unit that stores a scenario that defines the progress of the meeting, an audio data acquisition unit that acquires audio data relating to the meeting in accordance with the scenario stored in the scenario storage unit, a stage identification unit that identifies the stage of the scenario corresponding to the audio data based on the acquired audio data, an emotion identification unit that identifies the emotions of the meeting participants at the stage of the scenario identified by the stage identification unit, an utterance state identification unit that identifies the utterance state of the meeting participants based on the acquired audio data, and a comment generation unit that generates comments based on the emotions identified by the emotion identification unit and the utterance state identified by the utterance state identification unit, wherein the utterance state identification unit identifies at least the turn-take, the amount of utterance, and the utterance in the meeting.

Description

Technical Field

[0001] The present invention relates to a meeting support system, a meeting support method, and a program.

Background Art

[0002] For example, Patent Document 1 discloses a meeting support system including: a meeting information acquisition means for acquiring information related to a meeting; a speech acquisition means for acquiring speech contents in the meeting; an evaluation input means for inputting information related to a subjective evaluation of the meeting by participants in the meeting; a score calculation means for calculating an evaluation score of the meeting from the information related to the meeting, the speech contents in the meeting, and the information related to the subjective evaluation of the meeting; and a notification means for notifying the evaluation score of the meeting.

[0003] Further, Patent Document 2 discloses an information processing system that constitutes a virtual space or an online meeting and provides information to users participating in the virtual space or the online meeting. The information processing system includes a server device and a terminal device connected to each other via a network. The server device includes: a login user identification unit for identifying a user logged in to the virtual space or the online meeting; an action grasping unit for grasping the actions of the user identified by the login user identification unit; and a point management unit for giving points to the user based on the actions of the user grasped by the action grasping unit. The terminal device includes: a login instruction unit for sending instructions for logging in and out of the virtual space or the online meeting to the server device; an action detection unit for detecting the actions of the user; an action transmission unit for transmitting information based on the actions of the user detected by the action detection unit to the server device via the network; and a point reception unit for receiving information related to the points corresponding to the user managed by the point management unit of the server device from the server device via the network.

Prior Art Documents

Patent Documents

[0004] [Patent Document 1] Japanese Patent Publication No. 2021-163405 [Patent Document 2] Japanese Patent Publication No. 2024-67914 [Overview of the project] [Problems that the invention aims to solve]

[0005] The present invention aims to provide a meeting support system that provides support for meetings in accordance with their progress. [Means for solving the problem]

[0006] The conference support system according to the present invention includes a scenario storage unit that stores a scenario that defines the progress of a conference; an audio data acquisition unit that acquires audio data relating to a conference in accordance with the scenario stored in the scenario storage unit; a stage identification unit that identifies the stage of the scenario corresponding to the audio data acquired by the audio data acquisition unit; an emotion identification unit that identifies the emotions of the conference participants at the stage of the scenario identified by the stage identification unit; an utterance state identification unit that identifies the utterance state of the conference participants based on the audio data acquired by the aforementioned unit; and a comment generation unit that generates comments based on the emotions identified by the emotion identification unit and the utterance state identified by the utterance state identification unit, wherein the utterance state identification unit identifies at least the turntakes, the amount of utterances, and the utterances in the conference.

[0007] The conference support system according to the present invention includes: a scenario storage unit that stores a scenario that defines the progress of a conference; an audio data acquisition unit that acquires audio data relating to a conference in accordance with the scenario stored in the scenario storage unit; a stage identification unit that identifies the stage of the scenario corresponding to the audio data based on the audio data acquired by the audio data acquisition unit; an emotion identification unit that identifies the emotions of the conference participants at the stage of the scenario identified by the stage identification unit; an utterance state identification unit that identifies the utterance state of the conference participants based on the audio data acquired by the audio data acquisition unit; a display unit that displays the emotions identified by the emotion identification unit and the utterance state identified by the utterance state identification unit on a conference status display terminal; and a comment receiving unit that accepts comments created by the user based on the information displayed by the display unit, wherein the utterance state identification unit identifies at least the turn-takes, the amount of speech, and the utterances in the conference.

[0008] Preferably, the emotion identification unit further includes an emotion transition analysis unit that identifies the participant's emotions at predetermined intervals based on the audio data acquired by the audio data acquisition unit, and associates the transition of the scenario stages identified by the stage identification unit with the transition of the participant's emotions analyzed by the emotion identification unit, and the comment generation unit generates a comment based on the transition of the scenario stages associated by the emotion transition analysis unit and the transition of the participant's emotions.

[0009] Preferably, the emotion identification unit identifies the emotions of the meeting participants corresponding to the stage of the scenario identified by the stage identification unit, based on the voice information including nonverbal information from the voice data acquired by the voice data acquisition unit.

[0010] Preferably, the speech state identification unit further includes a speech transition analysis unit that identifies speech states at predetermined intervals based on the voice data acquired by the voice data acquisition unit, and associates the transitions between scenario stages identified by the stage identification unit with the transitions in speech states identified by the speech state identification unit, and the comment generation unit generates comments based on the transitions between scenario stages and the transitions in speech states associated by the speech transition analysis unit.

[0011] Preferably, the system further includes: a participant identification unit that identifies participants based on audio data acquired by the audio data acquisition unit; a role identification unit that identifies the role of the participant identified by the participant identification unit in the meeting based on the utterances in the audio data acquired by the audio data acquisition unit; and a participant evaluation unit that evaluates the participant identified by the participant identification unit based on the role identified by the role identification unit, the transitions between scenario stages and the transitions in the utterance state associated by the utterance transition analysis unit, the transitions between scenario stages associated by the emotion transition analysis unit, and the transitions in the participant's emotions.

[0012] Preferably, the system further includes an evaluation receiving unit that receives evaluations from participants of the meeting for comments generated by the comment generation unit, and a comment evaluation storage unit that stores the evaluations received by the evaluation receiving unit, the comments that are the subject of the evaluation, the transitions between stages of the associated scenario and the transitions in the speech state as determined by the speech transition analysis unit, and the transitions between stages of the associated scenario and the transitions in the emotions of the participants as determined by the emotion transition analysis unit.

[0013] Preferably, the system includes a meeting evaluation storage unit that stores a meeting classification, an index indicating the optimal progression of speech states in the classified meeting, and an index indicating the optimal progression of emotions in the classified meeting, in association with each other; and a meeting evaluation unit that classifies the meeting based on the transitions of scenario stages and the progression of speech states associated by the speech progression analysis unit, and the transitions of scenario stages and the emotions of participants associated by the emotion progression analysis unit, wherein the meeting evaluation unit generates advice based on the meeting classification to bring the progression of speech states and the emotions of participants in the meeting closer to the optimal index stored in the meeting evaluation storage unit.

[0014] Preferably, the system further includes: an emotion transition analysis unit that associates the transitions of scenario stages identified by the stage identification unit with the transitions of participants' emotions analyzed by the emotion identification unit; an utterance transition analysis unit that associates the transitions of scenario stages identified by the stage identification unit with the transitions of utterance states identified by the utterance state identification unit; a meeting evaluation storage unit that stores a meeting classification, an index indicating the optimal transition of utterance states in the classified meeting, and an index indicating the optimal transition of emotions in the meeting, in association with each other; and a meeting evaluation unit that classifies the meeting based on the transitions of scenario stages and transitions of utterance states associated by the speech transition analysis unit and the transitions of scenario stages and transitions of participants' emotions associated by the emotion transition analysis unit, wherein the meeting evaluation unit accepts advice to bring the transitions of utterance states and transitions of participants' emotions in the meeting closer to the optimal index stored in the meeting evaluation storage unit based on the meeting classification, and the advice is created by the user.

[0015] The conference support method according to the present invention includes: a scenario storage step in which a computer stores a scenario that defines the progress of a conference; an audio data acquisition step in which the computer acquires audio data relating to a conference in accordance with the scenario stored in the scenario storage step; a stage identification step in which the computer identifies a stage of the scenario corresponding to the audio data based on the audio data acquired in the audio data acquisition step; an emotion identification step in which the computer identifies the emotions of the conference participants at the stage of the scenario identified in the stage identification step; an utterance state identification step in which the computer identifies the utterance state of the conference participants based on the audio data acquired in the audio data acquisition step; and a comment generation step in which the computer generates a comment based on the emotions identified in the emotion identification step and the utterance state identified in the utterance state identification step, wherein in the utterance state identification step, at least the turntakes, the amount of utterances, and the utterances in the conference are identified.

[0016] The program according to the present invention causes a computer to execute the following steps: a scenario storage step for storing a scenario that defines the progress of a meeting; an audio data acquisition step for acquiring audio data relating to a meeting in accordance with the scenario stored in the scenario storage step; a stage identification step for identifying the stage of the scenario corresponding to the audio data based on the audio data acquired in the audio data acquisition step; an emotion identification step for identifying the emotions of the meeting participants at the stage of the scenario identified in the stage identification step; an utterance state identification step for identifying the utterance state of the meeting participants based on the audio data acquired in the audio data acquisition step; and a comment generation step for generating comments based on the emotions identified in the emotion identification step and the utterance state identified in the utterance state identification step, wherein in the utterance state identification step, at least the turntakes, the amount of speech, and the utterances in the meeting are identified.

[0017] The conference support method according to the present invention comprises: a scenario storage step in which a computer stores a scenario that defines the progress of a conference; an audio data acquisition step in which the computer acquires audio data relating to a conference in accordance with the scenario stored in the scenario storage step; a stage identification step in which the computer identifies a stage of the scenario corresponding to the audio data based on the audio data acquired in the audio data acquisition step; an emotion identification step in which the computer identifies the emotions of the conference participants at the stage of the scenario identified in the stage identification step; an utterance state identification step in which the computer identifies the utterance state of the conference participants based on the audio data acquired in the audio data acquisition step; a display step in which the computer displays the emotions identified in the emotion identification step and the utterance state identified in the utterance state identification step; and a comment reception step in which the computer receives comments created by a user based on the information displayed in the display step, wherein in the utterance state identification step, at least the turntakes, the amount of utterances, and the utterances in the conference are identified.

[0018] The program according to the present invention includes a scenario storage step for storing a scenario that defines the progress of a meeting, an audio data acquisition step for acquiring audio data related to the meeting along the scenario stored in the scenario storage step, a stage identification step for identifying the stage of the scenario corresponding to the audio data based on the audio data acquired in the audio data acquisition step, an emotion identification step for identifying the emotions of the participants in the meeting at the stage of the scenario identified in the stage identification step, a speech state identification step for identifying the speech state of the participants in the meeting based on the audio data acquired in the audio data acquisition step, a display step for displaying the emotions identified in the emotion identification step and the speech state identified in the speech state identification step, and a comment reception step for receiving comments created by a user based on the information displayed in the display step, and causes a computer to execute them. In the speech state identification step, at least the turn-taking, the amount of speech, and the speech in the meeting are identified.

Effects of the Invention

[0019] According to the present invention, it is possible to provide support for a meeting according to the progress of the meeting.

Brief Description of the Drawings

[0020] [Figure 1] It is a diagram illustrating the overall configuration of the meeting support system 1. [Figure 2] It is a diagram illustrating the hardware configuration of the meeting support server 5. [Figure 3] It is a diagram illustrating the functional configuration of the meeting support server 5. [Figure 4] (a) is a diagram illustrating the profile of a meeting, and (b) is a diagram illustrating the profile of a questioner. [Figure 5] It is a diagram illustrating the scenario information table 60. [Figure 6] It is a diagram illustrating the meeting evaluation table 61. [Figure 7]It is a diagram illustrating the role evaluation table 62. [Figure 8] It is a diagram illustrating the comment evaluation table 63. [Figure 9] It is a flowchart for explaining the meeting evaluation process (S10) by the meeting support system 1. [Figure 10] It is a flowchart for explaining the participant evaluation process (S20) by the meeting support system 1. [Figure 11] It is a flowchart for explaining the comment evaluation process (S30) by the meeting support system 1. [Figure 12] It is a diagram illustrating the overall configuration of the meeting support system 1 in Embodiment 2. [Figure 13] It is a diagram illustrating the functional configuration of the meeting support server 5 in Embodiment 2. [Figure 14] It is a diagram illustrating the meeting state display screen displayed on the meeting state display terminal 9 in Embodiment 2.

Mode for Carrying Out the Invention

[0021] First, the background of the present invention will be described. For meetings, there are, for example, stages of meeting progress such as issue sharing, current situation grasping, discussion, and conclusion. During the flow of the meeting, the speeches and emotions of the participants change. In order to realize a highly productive meeting, it is important to draw out the opinions and questions of the participants after grasping the changes in the meeting state and the emotions of the participants accompanying the progress of the meeting. Therefore, the present invention provides support for activating the meeting and leading to problem solving by presenting comments in real time based on the changes in the meeting state and the emotions of the participants accompanying the progress of the meeting.

[0022] Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the scope of the present invention is not limited to the illustrated examples. Furthermore, in this embodiment, the meeting is described as a face-to-face meeting format in which participants gather in the same place (such as a conference room or office). However, the meeting in this invention is not limited to face-to-face meetings, but also includes online meetings. In this embodiment, the case of a face-to-face meeting will be described.

[0023] (Embodiment 1) Figure 1 is a diagram illustrating the overall configuration of the conference support system 1. As illustrated in Figure 1, the conference support system 1 comprises a sound collection device 3, a conference support server 5, and a participant terminal 7, which are connected to each other via a network NW. The sound collection device 3 includes a microphone array containing multiple sound collection units (microphones) arranged in different orientations. The sound collection device 3 collects sound from 360° around it and transmits the collected sound as audio data to the conference support server 5. The sound collection device 3 also functions as a speaker that converts comments generated by the conference support server 5 into audio data and outputs it. The conference support server 5 is a server device that analyzes the conference status as the conference progresses based on audio data acquired from the sound collection device 3. The conference support server 5 generates comments on the conference based on the analysis results and performs evaluations of the conference and the conference participants. The conference support server 5 displays the evaluation of the conference on the participant terminal 7. The meeting will proceed according to a pre-prepared scenario, and participants will speak according to the script based on that scenario.

[0024] Participant terminal 7 is a computer terminal such as a smartphone, tablet, or personal computer, and is a terminal that meeting participants operate and view. Participant terminal 7 displays the user interface provided by the meeting support system 1 and displays information that participants should share, as well as evaluation results from the meeting support server 5. Information that participants should share includes, for example, the meeting agenda and decisions made during the meeting. There are 3 to 8 participants in the meeting, and all participants may view one participant terminal 7, or each participant may have their own individual participant terminal 7. The display or hiding of evaluation results on participant terminal 7 is optional.

[0025] Figure 2 is a diagram illustrating the hardware configuration of the conference support server 5. As illustrated in Figure 2, the conference support server 5 includes a CPU 550, memory 552, HDD 554, network interface 556 (network IF 556), display device 558, and input device 560, and these components are connected to each other via a bus 562. The CPU550 is, for example, a central processing unit. Memory 552 is, for example, volatile memory and functions as main memory. The HDD554 is, for example, a hard disk drive, which stores computer programs and other data files as a non-volatile recording device. Network IF556 is an interface for wired or wireless communication, and for example, it enables communication with the sound collection device 3 and the participant terminal 7. The display device 558 is, for example, a liquid crystal display. The input device 560 is, for example, a keyboard and a mouse. The conference support server 5 may also be a cloud server.

[0026] Figure 3 is a diagram illustrating the functional configuration of the conference support server 5. As illustrated in Figure 3, the conference support server 5 in this example has the conference support program 50 installed, and is configured with a scenario database 600 (scenario DB600), a conference evaluation database 610 (conference evaluation DB610), a role evaluation database 620 (role evaluation DB620), and a comment evaluation database 630 (comment evaluation DB630). The meeting support program 50 includes an operation detection unit 500, a condition reception unit 502, a field reception unit 504, an audio data acquisition unit 506, a stage identification unit 508, an emotion identification unit 510, an utterance state identification unit 512, an emotion transition analysis unit 514, an utterance transition analysis unit 516, a timing determination unit 518, a comment generation unit 520, a participant identification unit 522, a role identification unit 524, an evaluation value calculation unit 526, a participant evaluation unit 528, an evaluation reception unit 530, and a meeting evaluation unit 532. Furthermore, part or all of the conference support program 50 may be implemented using hardware such as an ASIC. The conference support program 50 is stored on a recording medium such as a CD-ROM and installed on the conference support server 5 via this recording medium.

[0027] In the meeting support program 50, the operation detection unit 500 detects operations on the user interface of the meeting support system 1 displayed on the participant terminal 7. For example, the operation detection unit 500 detects the pressing of the meeting start button, the meeting end button, and the button indicating the evaluation of comments. The condition reception unit 502 accepts the selection of meeting profile candidates, which are meeting information, displayed on the participant terminal 7. As illustrated in Figure 4(a), a meeting profile is meeting information that includes industry, area, job title, theme, and number of employees. More specifically, the meeting information received by the condition reception unit 502 becomes input information for the learning model used by the comment generation unit 520 to generate comments. The field reception unit 504 accepts the selection of candidate questioner profiles, which are questioner information, displayed on the participant terminal 7. As illustrated in Figure 4(b), the questioner profile is information about the questioner that includes industry, area, occupation, position, job duties, age, and number of employees. More specifically, the questioner information received by the field reception unit 504 becomes input information for the learning model used by the comment generation unit 520 to generate comments. Here, the questioner is an avatar, a character in a virtual space that emits comments generated by the conference support server 5. Furthermore, the field reception unit 504 may accept a selection of the expression method for the generated comments, in addition to the questioner information. For example, the field reception unit 504 may accept a selection of the expression method, such as "mild" to indicate that the comments are gentle or mild, or "harsh" to indicate that the comments are harsh or strong. The comment generation unit 520 generates the comments based on the accepted expression method.

[0028] The audio data acquisition unit 506 acquires audio data of the meeting audio collected by the sound collection device 3 from the sound collection device 3. Specifically, the audio data acquisition unit 506 acquires audio data related to the meeting in accordance with the scenarios stored in the scenario DB 600. The audio information obtained from the audio data includes non-verbal information and linguistic information. The stage identification unit 508 identifies the stage (stage) of the scenario corresponding to the acquired audio data, based on the audio data acquired by the audio data acquisition unit 506. Specifically, the stage identification unit 508 identifies which stage of the scenario the acquired audio data corresponds to, based on the participants' statements in the audio data, the elapsed time since the start of the meeting, or the participants' statements and the elapsed time since the start of the meeting. The stage identification unit 508 is an example of a stage identification unit according to the present invention.

[0029] The emotion identification unit 510 identifies the emotions of meeting participants at the stage of the scenario identified by the stage identification unit 508, based on the audio data acquired by the audio data acquisition unit 506. Specifically, the emotion identification unit 510 identifies the emotions of participants at predetermined intervals based on the audio data acquired by the audio data acquisition unit 506. For example, the emotion identification unit 510 identifies the emotions of participants every 5 minutes based on the audio data. More specifically, the emotion identification unit 510 identifies the emotions of meeting participants corresponding to the stage of the scenario based on audio information, including nonverbal information, from the audio data acquired by the audio data acquisition unit 506. Nonverbal information includes the tone of voice, volume, speaking speed, and length of silence of the participants' speech. The linguistic information included in the audio information is meaningful words or phrases whose content can be understood. The emotions identified are at least positive and negative emotions, but other emotions may also be identified.

[0030] The speech state identification unit 512 identifies the speech state of meeting participants from the audio data acquired by the audio data acquisition unit 506. The speech state identification unit 512 identifies at least the turn-taking of the meeting, the amount of speech by the participants, and the participants' speech as speech states. The speech state identification unit 512 identifies the speech state at predetermined intervals. For example, the speech state identification unit 512 identifies the speech state every 5 seconds based on the audio data. Here, the speech identified by the speech state identification unit 512 refers to the linguistic information uttered by the participants. The emotion transition analysis unit 514 correlates the transitions between scenario stages identified by the stage identification unit 508 with the emotional transitions of participants identified by the emotion identification unit 510. Here, the correlated information is referred to as emotion transition information. The emotion transition analysis unit 514 also visualizes the emotion transition information and displays it on the participant terminal 7. The emotion transition analysis unit 514 displays the emotion transition information on the participant display terminal 7 from the start of the meeting up to the stage division after a certain period of time has elapsed. In addition, the emotion transition analysis unit 514 displays the emotion transition information on the participant display terminal 7 from the start of the meeting up to the point determined by the meeting facilitator, based on the facilitator's judgment.

[0031] The speech transition analysis unit 516 associates the transitions between scenario stages identified by the stage identification unit 508 with the transitions between speech states identified by the speech state identification unit 512. Here, the associated information is called speech state transition information. The speech transition analysis unit 516 also visualizes the speech state transition information and displays it on the participant terminal 7. The speech transition analysis unit 516 displays the speech state transition information on the participant display terminal 7 from the start of the meeting up to the stage division after a certain period of time has elapsed. In addition, the speech transition analysis unit 516 displays the speech state transition information on the participant display terminal 7 from the start of the meeting up to the point determined by the meeting facilitator, based on the facilitator's judgment. The emotion transition analysis unit 514 and the speech transition analysis unit 516 do not constantly display emotion transition information and speech state transition information on the participant terminal 7, but only at specific timings. This reduces the psychological stress on participants compared to when emotion transition information and speech state transition information are constantly displayed. Furthermore, temporarily displaying emotion transition information and speech state transition information up to the stage division during the meeting helps motivate participants to move on to the next stage. The timing determination unit 518 determines the timing for generating a comment. Specifically, the timing determination unit 518 determines that it is time to generate a comment if it determines that a participant is emotional based on the emotions of the participants identified by the emotion identification unit 510, or if it determines that there is a bias in the participants who are speaking or that there are long periods of silence, based on the speech state identification unit 512.

[0032] The comment generation unit 520 generates comments based on the emotions identified by the emotion identification unit 510 and the speech state identified by the speech state identification unit 512. Specifically, the comment generation unit 520 generates comments based on the emotion transition information analyzed by the emotion transition analysis unit 514 and the speech state transition information analyzed by the speech state transition analysis unit 516. More specifically, the comment generation unit 520 generates comments based on the meeting information received by the condition reception unit 502, the questioner information and expression method received by the field reception unit 504, the emotion transition information, and the speech state transition information. The comment generation unit 520 is a generative AI and generates comments based on a pre-trained learning model for comment generation. The comment generation unit 520 generates comments when the timing determination unit 518 determines that it is time to generate comments. Furthermore, the comment generation unit 520 converts the generated comments into audio data from the sound collection device 3 and outputs it. The comment generation unit 520 may also display the generated comments on the participant terminal 7.

[0033] The participant identification unit 522 identifies participants based on the audio data acquired by the audio data acquisition unit 506. Specifically, the participant identification unit 522 identifies meeting participants based on the characteristics of the voice (e.g., pitch, range) and the location from which the voice is emitted. The role identification unit 524 identifies the role of the participant identified by the participant identification unit 522 in the meeting based on the utterances in the audio data acquired by the audio data acquisition unit 506. Specifically, the role identification unit 524 matches the lines associated with the scenarios stored in the scenario DB 600 with the statements identified by the utterance state identification unit 512 to identify the role of the participant.

[0034] The evaluation value calculation unit 526 calculates an evaluation value of the meeting state based on the progression of speech states from the start to the end of the meeting and the progression of emotions of the meeting participants from the start to the end of the meeting. Specifically, the evaluation value calculation unit 526 calculates an evaluation value based on the speech variance from the start to the end of the meeting, the turn-take variance, and the emotion progression information. Furthermore, the evaluation value calculation unit 526 calculates evaluation values ​​for each participant's speech state progression and emotional progression. Specifically, the evaluation value calculation unit 526 calculates evaluation values ​​for each participant's total speaking time in the meeting, the number of turns taken during the meeting, and emotional progression information. The participant evaluation unit 528 evaluates the participants identified by the participant identification unit 522 based on the roles identified by the role identification unit 524, the progression of speech states identified by the speech progression analysis unit 516, and the progression of emotions analyzed by the emotion progression analysis unit 514. Specifically, the participant evaluation unit 528 evaluates the participant's role based on the participant's speaking time during the meeting, the number of turns taken during the meeting, the evaluation value of the emotion data, and the optimal value of the role in the participant evaluation table 62.

[0035] The evaluation reception unit 530 receives evaluations from meeting participants of comments generated by the comment generation unit 520. Specifically, the evaluation reception unit 530 detects when a button indicating an evaluation is pressed by a participant on the user interface displayed on the participant terminal 7. The evaluation reception unit 530 also associates the received evaluation with the comment being evaluated, the speech state progression identified by the speech progression analysis unit 516 until the comment was generated, and the emotion progression information identified by the emotion progression analysis unit 514, and stores this information in the comment evaluation DB 630. The meeting evaluation unit 532 classifies meetings based on the emotional transitions identified by the emotion transition analysis unit 514 and the speech transitions identified by the speech transition analysis unit 516. The meeting evaluation unit 532 is also a generative AI and, based on the meeting classification, generates advice to bring the evaluation value of the meeting state calculated by the evaluation value calculation unit 526 closer to the optimal index corresponding to the meeting classification, which is managed by the meeting evaluation table 61. The meeting evaluation unit 532 displays the meeting classification and the generated advice on the participant terminal 7.

[0036] Figure 5 is an example of a scenario information table 60 stored in the scenario DB 600. Scenario Information Table 60 manages the association between Scenario ID, Stage ID, Stage Name, Time Allocation, Role, and Dialogue. The Scenario ID is identification information that identifies the meeting scenario. The Stage ID is information that identifies the stage of the scenario. The Stage Name is the name of each stage in the scenario. The Time Allocation is an estimate of the time to be spent on each stage of the scenario. The Role is the role of the character appearing in the scenario. The Dialogue defines the lines spoken by each role in each stage of the scenario. The dialogue defined in Scenario Information Table 60 is at least dialogue that can identify the stage, and explanations of meeting topics and issues are improvised by the participants during the meeting. Participants are provided with a script for each scenario the meeting follows, and they speak lines according to their role. Any lines other than those assigned are improvised by the participant in that role. The stage identification unit 508 identifies the stage by comparing the spoken words in the acquired audio data with the dialogue in the scenario stored in the scenario information table 60.

[0037] Figure 6 is an example of a meeting evaluation table 61 stored in the meeting evaluation DB 610. As illustrated in Figure 6, the meeting evaluation table 61 has evaluation values ​​that indicate the optimal state of a meeting for each type of meeting. Specifically, the meeting evaluation table 61 represents the state of a meeting using the variance of utterances among participants, the variance of the number of turntakes among participants, and the emotion transition data value. The emotion transition data value is a numerical value that quantifies emotion transition information. The emotion transition data value includes at least the values ​​of positive emotion transition data and negative emotion transition data, but may also include other emotion transition data. The meeting evaluation unit 532 compares the evaluation values ​​of the progression of speech states and emotional states in the conducted meeting with the optimal value of the meeting state managed by the meeting evaluation table 61, and identifies the type of meeting that has the optimal value that approximates the conducted meeting. In addition, the meeting evaluation unit 532 classifies the type of meeting by the variance of speech among participants, the variance of the number of turntakes among participants, and the emotional state data value, but the type of meeting may also be classified by including other indicators.

[0038] Figure 7 is an example of a role evaluation table 62 stored in the role evaluation DB 620. As illustrated in Figure 7, the role evaluation table 62 contains evaluation values ​​that indicate the optimal meeting state for each role. Specifically, the role evaluation table 62 contains a scenario ID, a role, and the optimal value for the meeting state. The scenario ID is information that identifies the scenario that determines the progress of the meeting. The role is the role that the participants are assigned in the scenario. The optimal value for the meeting state represents the evaluation values ​​for the optimal speaking time during the meeting, the number of turns during the meeting, and the emotion transition data value for each role. The participant evaluation unit 528 evaluates the participants assigned to the roles identified by the role identification unit 524 based on the evaluation values ​​in the role evaluation table 62.

[0039] Figure 8 is an example of a comment evaluation table 63 stored in the comment evaluation DB 630. As illustrated in Figure 8, the comment evaluation table 63 manages comment evaluation information by associating the meeting ID, scenario ID, comment ID, comment, evaluation, emotional progression, and speech state progression. The meeting ID is information that identifies the meeting. The scenario ID is information that identifies the scenario used in the meeting. The comment ID is information that identifies the comment generated in the meeting. The comment is the generated comment. The rating is the participant's rating of the generated comment. The emotional transition and the speech state transition represent the state of the meeting that led to the generation of the comment. Specifically, the emotional transition shows the scenario stage that led to the generation of the comment and the emotional transition at each stage (emotional transition data), and the speech state shows the scenario stage that led to the generation of the comment and the transition of the participants' speech state at each stage (speech transition data) in the meeting.

[0040] The evaluation reception unit 530 updates the comment evaluation table 63 by associating the received evaluation with the meeting ID, scenario ID, comment ID, comment, emotional progression, and speech state progression. The comment generation unit 520 learns a model for generating comments using comment evaluation information managed by the comment evaluation table 63 as input information. As comment evaluation information is accumulated, a more accurate comment generation learning model can be constructed.

[0041] Figure 9 is a flowchart illustrating the meeting evaluation process (S10). Prior to the meeting, it will be decided that the meeting will follow the scenario with scenario ID S1. The script for scenario S1 will be distributed to the meeting participants, and the participants will conduct the meeting and speak according to the script. As illustrated in Figure 9, in step 100 (S100), the condition receiving unit 502 of the conference support server 5 accepts the selection of a candidate conference profile representing conference information displayed by the participant terminal 7. In step 105 (S105), the field reception unit 504 of the conference support server 5 accepts the selection of candidate questioner profiles and the selection of the method of representation for the questioner information displayed by the participant terminal 7. In step 110 (S110), the operation detection unit 500 of the conference support server 5 detects that the participant terminal 7 has pressed the "Start Conference" button. The sound collection device 3 begins collecting audio of the conference. The sound collection device 3 transmits the collected audio data to the conference support server 5. The audio data acquisition unit 506 of the conference support server 5 acquires the audio data of the conference collected by the sound collection device 3 from the sound collection device 3. Audio collection by the sound collection device 3 and acquisition of audio data by the audio data acquisition unit 506 continue until the conference ends.

[0042] In step 115 (S115), the stage identification unit 508 compares the participant's statements with the lines corresponding to scenario S1 in the scenario information table 60, based on the audio data acquired by the meeting data acquisition unit 500, and identifies which stage ID the participant's statements correspond to. In step 120 (S120), the participant identification unit 522 identifies the participant based on the characteristics of the voice data acquired by the voice data acquisition unit 506 and the location of the sound source. In step 125 (S125), the emotion identification unit 510 identifies the emotions of each participant in the meeting, identified by the participant identification unit 522, at 5-minute intervals, based on the audio data acquired by the audio data acquisition unit 506, for each stage ID identified by the stage identification unit 508.

[0043] In step 130 (S130), the speech state identification unit 512 identifies the turn-taking between meeting participants identified by the participant identification unit 522, the amount of speech from each participant, and the speech itself, every 5 seconds, based on the voice data acquired by the voice data acquisition unit 506. In step 135 (S135), the emotion transition analysis unit 514 correlates the transition of scenario stages identified by the stage identification unit 508 with the transition of participants' emotions identified by the emotion identification unit 510. The speech transition analysis unit 516 correlates the transition of scenario stages identified by the stage identification unit 508 with the transition of speech states identified by the speech state identification unit 512.

[0044] In step 140 (S140), the timing determination unit 518 determines, based on the emotions of the participants identified by the emotion identification unit 510, that a participant is emotional, or that there is a bias in the participants who are speaking, or that there is a long period of silence (S140: Yes), that it is necessary to generate a comment and proceed to S145. Otherwise (S140: No), it proceeds to S150. In step 145 (S145), the comment generation unit 520 generates a comment based on the meeting information received by the condition reception unit 502, the questioner information and expression method received by the field reception unit 504, the emotion transition information analyzed by the emotion transition analysis unit 514, and the speech state transition information identified by the speech transition analysis unit 516. The comment generation unit 520 converts the generated comment into audio data from the sound collection device 3 and outputs it.

[0045] In step 150 (S150), if the operation detection unit 500 detects that the end-of-meeting button has been pressed (S150: Yes), the process proceeds to S155. If the operation detection unit 500 does not detect that the end-of-meeting button has been pressed, the meeting evaluation process (S10) proceeds to S110 and continues to acquire audio data. In step 155 (S155), the evaluation value calculation unit 526 calculates evaluation values ​​for the speech variance, the turn-take variance, and the emotion transition data value based on the progression of speech in the meeting and the progression of emotions of the meeting participants. The meeting evaluation unit 532 identifies the classification of the meeting based on the evaluation values ​​calculated by the evaluation value calculation unit 526 and the evaluation values ​​in the meeting evaluation table 61. In step 160 (S160), the meeting evaluation unit 532 generates advice to bring the meeting closer to the optimal indicator managed by the meeting evaluation table 61, based on the classification of the identified meeting. The meeting evaluation unit 532 displays the meeting classification and the generated advice on the participant terminal 7.

[0046] This allows for comments to be made that are in line with the flow of the meeting, resulting in the generation of natural comments that follow the conversation. Furthermore, instead of forcing productive meetings into a single pattern, it provides advice tailored to each type of meeting, enabling the generation of optimal advice that is appropriate for the challenges and participants.

[0047] Figure 10 is a flowchart illustrating the participant evaluation process (S20). The participant evaluation process (S20) assumes that the meeting follows scenario ID S1, that the script for the predetermined scenario S1 is distributed to the meeting participants, and that the participants conduct the meeting and speak according to the script. It is also assumed that the meeting has ended and that the analysis of emotional transitions and speech transitions has been completed. The person to be evaluated is Participant A. As illustrated in Figure 9, in step 200 (S200), the role identification unit 524 identifies the role of participant A, who has been identified by the participant identification unit 522. Specifically, the role identification unit 524 compares the lines in the scenario information table 60 where the scenario ID is scenario S1 with the statements identified from the audio data by the speech state identification unit 512, and identifies the role of participant A who speaks the matching lines.

[0048] In step 205 (S205), the evaluation value calculation unit 526 calculates evaluation values ​​for participant A's emotional changes and speech state changes. Specifically, the evaluation value calculation unit 526 calculates participant A's speaking time during the meeting, the number of turns taken during the meeting, and the evaluation value of the emotional data. The participant evaluation unit 528 evaluates participant A, who is assigned a role identified by the role identification unit 524, based on the evaluation values ​​in the role evaluation table 62 and the evaluation values ​​calculated by the evaluation value calculation unit 526. The participant evaluation unit 528 displays the evaluation results on participant A's participant terminal 7. Since individual participant evaluations are obtained, it is possible to understand each participant's strengths and areas for improvement, which can be used to improve future meetings. Furthermore, since evaluations of roles are obtained, the requirements for each role become clearer.

[0049] Figure 11 is a flowchart illustrating the comment evaluation process (S30). The meeting will proceed according to scenario S1, and participants will speak according to the script of scenario S1. Furthermore, the comment evaluation process (S30) will occur when the processing from S100 to S145 in the meeting evaluation process (S10) has been completed, and the timing determination unit 518 has determined that a comment should be generated. The comment evaluation process when a comment is generated will be described below. As illustrated in Figure 10, in step 300 (S300), the operation detection unit 500 detects the pressing of a button indicating an evaluation by the participant on the user interface displayed on the participant terminal 7. In step 305 (S305), the evaluation reception unit 530 associates the evaluation detected by the operation detection unit 500, the comment that is the subject of the evaluation, the emotion transition information up to the generation of the comment analyzed by the emotion transition analysis unit 514, and the speech state transition information up to the generation of the comment identified by the speech transition analysis unit 516, and stores them in the comment evaluation DB 630. This allows the information managed in the comment evaluation table 63 to be used as an indicator for generating comments that make meetings more effective. Furthermore, the comment evaluation table 63 serves as input data for the learning model used by the comment generation unit 520, which helps in building the learning model.

[0050] As explained above, the meeting support system 1 in this embodiment generates comments based on the progression of speech during the meeting and the progression of participants' emotions, enabling the output of questions and comments that take the flow of the meeting into consideration. Furthermore, the output of comments, evaluation of the meeting, and evaluation of participants are based not only on the progression of speech during the meeting and the progression of participants' emotions, but also on meeting information and questioner information, making it possible to provide more accurate feedback. Moreover, since the progression of speech during the meeting and the progression of participants' emotions are obtained from the collected audio data of the meeting, there is no need to install cameras to film participants, so the meeting can proceed without causing participants the psychological stress of "being filmed," making it easier for participants to express their opinions in a relaxed state. In addition, since the meeting proceeds according to a scenario, it is clear what participants should say and how to proceed, reducing the psychological burden on participants and making the meeting proceed smoothly.

[0051] (Embodiment 2) In the above embodiment 1, we described a case where the comment is generated by a comment generation unit 520, which is a generation AI. However, the invention is not limited to this, and for example, the comment may be created by a human questioner. Elements having substantially the same function and configuration as those in Embodiment 1 described above are denoted by the same reference numerals, thus omitting redundant explanations. Figure 12 is a diagram illustrating the overall configuration of the conference support system 1 according to Embodiment 2. As illustrated in Figure 12, the conference support system 1 in Embodiment 2 includes a conference status display terminal 9 in addition to the overall configuration shown in Figure 1. The meeting status display terminal 9 is a computer terminal operated by the questioner who creates comments and advice for the meeting. For example, the meeting status display terminal 9 is a mobile device such as a smartphone or tablet, or a computer terminal such as a personal computer. The meeting status display terminal 9 displays the progression of speech states indicating the meeting status and the progression of participants' emotions. The questioner creates comments based on the meeting status displayed on the meeting status display terminal 9 and the audio of the meeting. Comments created by the human questioner are output as being spoken by an avatar, which is a character in a virtual space. This reduces the tension and anxiety that a human questioner might feel when making comments with their face visible, and also reduces the emotional impact on meeting participants who receive the comments.

[0052] Figure 13 is a diagram illustrating the functional configuration of the conference support server 5 according to Embodiment 2. As illustrated in Figure 13, the conference support server 5 of Embodiment 2 has a conference support program 55 installed, and is configured with a scenario database 600 (scenario DB600), a conference evaluation database 610 (conference evaluation DB610), a role evaluation database 620 (role evaluation DB620), and a comment evaluation database 630 (comment evaluation DB630). The meeting support program 55 includes an operation detection unit 500, a condition reception unit 502, a field reception unit 504, an audio data acquisition unit 506, a stage identification unit 508, an emotion identification unit 510, an utterance state identification unit 512, an emotion transition analysis unit 514, an utterance transition analysis unit 516, a timing determination unit 518, a participant identification unit 522, a role identification unit 524, an evaluation value calculation unit 526, a participant evaluation unit 528, and an evaluation reception unit 530, in addition to a display unit 534, a comment reception unit 536, and a meeting evaluation unit 538.

[0053] The display unit 534 displays the meeting status on the meeting display terminal 9 operated by the questioner. Specifically, the display unit 534 displays the emotions identified by the emotion identification unit 510 and the speech states identified by the speech state identification unit 512 on the meeting status display terminal 9. More specifically, the display unit 534 displays the emotion transition information associated by the emotion transition analysis unit 514 and the speech state transition information associated by the speech transition analysis unit 516 on the meeting status display terminal 9. The comment receiving unit 536 receives comments created by the questioner based on the information displayed by the display unit 534. In Embodiment 2, the questioner is an example of a user according to the present invention. Specifically, the questioner creates a comment based on the meeting information received by the condition receiving unit 502, the meeting status displayed on the meeting status display terminal 9, and the audio collected by the sound collection device 3. The comment receiving unit 536 receives the input of the comment created by the questioner to the meeting status display terminal 9.

[0054] The meeting evaluation unit 538 classifies the meeting based on the emotional transitions identified by the emotion transition analysis unit 514 and the speech transitions identified by the speech transition analysis unit 516. Based on the meeting classification, the questioner creates advice to bring the meeting state evaluation value calculated by the evaluation value calculation unit 526 closer to the optimal index corresponding to the meeting classification, which is managed by the meeting evaluation table 61. The meeting evaluation unit 538 accepts the advice created by the questioner. The meeting evaluation unit 538 displays the meeting classification and the received advice on the participant terminal 7.

[0055] Figure 14 is an example of the meeting status display screen displayed on the meeting status display terminal 9 in Embodiment 2. As illustrated in Figure 14, the meeting status display screen shown on the meeting status display terminal 9 displays a turn-taking diagram, a speech state transition diagram, and an emotion transition diagram, and is provided with a comment input field and a comment insertion button. A turn-taking diagram is a visual representation of how each participant's statements are exchanged. A speech state transition diagram is a graph showing how much each participant speaks at each stage of the scenario. An emotion transition diagram is a graph showing what emotion each participant is feeling at each stage of the scenario. In Figure 14, emotional states are represented as positive and negative, but other emotional states may also be displayed. Additionally, the meeting status display screen may show the total speaking time for the entire meeting. The comment input field is where the questioner enters their comments. The comment insertion button outputs the comment entered in the comment input field from the sound collection device 3. The meeting status display screen is updated in real time during the meeting, allowing questioners to create comments based on the meeting status. Having a human questioner allows them to grasp the atmosphere of the meeting based on what is being said and the reactions of the participants, and to make comments appropriate to the situation.

[0056] The above embodiment describes a case where comments and advice for a meeting are generated by AI or a human, but it is not limited to this, and comments and advice for a meeting may be generated by both AI and a human. For example, the output may be a modified version of the comments and advice generated by the AI, or the comments and advice generated by each individual may be output. [Explanation of Symbols]

[0057] 1…Meeting support system 3…Sound collection device 5…Conference support server 7…Participant terminals 9…Conference status display terminal 50, 55… Conference support programs 500... Operation detection unit 502...Conditions Reception Department 504... Department Reception 506...Audio data acquisition unit 508... Stage Specific Section 510…Emotion identification part 512... Speech state identification unit 514…Emotional transition analysis department 516... Speech Transition Analysis Department 518... Timing determination unit 520... Comment generation unit 522...Participant Identification Unit 524... Role Identification Department 526...Evaluation value calculation unit 528...Participant Evaluation Department 530... Evaluation Reception Department 532, 538… Meeting Evaluation Department 534...Display section 536... Comment Reception Department 600... Scenario Database 610...Meeting Evaluation Database 620...Role Evaluation Database 630... Comment Rating Database

Claims

1. A scenario storage unit that stores the scenario that defines the progress of the meeting, A voice data acquisition unit acquires voice data related to a meeting in accordance with the scenario stored by the scenario storage unit, A stage identification unit identifies the stage of the scenario corresponding to the audio data acquired by the audio data acquisition unit, An emotion identification unit that identifies the emotions of the participants in the meeting at the stage of the scenario identified by the stage identification unit, Based on the audio data acquired by the aforementioned audio data acquisition unit, the speech state identification unit identifies the speech state of the participants in the meeting, A comment generation unit generates a comment based on the emotion identified by the emotion identification unit and the speech state identified by the speech state identification unit. It has, The speech state identification unit identifies at least the turn-taking, the amount of speech, and the speech during the meeting. Meeting support system.

2. A scenario storage unit that stores the scenario that defines the progress of the meeting, A voice data acquisition unit acquires voice data related to a meeting in accordance with the scenario stored by the scenario storage unit, A stage identification unit identifies the stage of the scenario corresponding to the audio data acquired by the audio data acquisition unit, An emotion identification unit that identifies the emotions of the participants in the meeting at the stage of the scenario identified by the stage identification unit, Based on the audio data acquired by the aforementioned audio data acquisition unit, the speech state identification unit identifies the speech state of the participants in the meeting, A display unit that displays the emotion identified by the emotion identification unit and the speech state identified by the speech state identification unit on a conference state display terminal, A comment receiving unit accepts comments created by the user based on the information displayed by the aforementioned display unit. It has, The speech state identification unit identifies at least the turn-taking, the amount of speech, and the speech during the meeting. Meeting support system.

3. The emotion identification unit identifies the participant's emotions at predetermined intervals based on the audio data acquired by the audio data acquisition unit. An emotion transition analysis unit that correlates the transitions between scenario stages identified by the stage identification unit with the changes in participants' emotions analyzed by the emotion identification unit. It further possesses, The comment generation unit generates comments based on the transitions between scenario stages associated by the emotion transition analysis unit and the changes in the participants' emotions. The conference support system according to claim 1.

4. The emotion identification unit identifies the emotions of the meeting participants corresponding to the stage of the scenario identified by the stage identification unit, based on the voice information, including nonverbal information, from the voice data acquired by the voice data acquisition unit. A conference support system according to claim 1 or claim 2.

5. The speech state identification unit identifies the speech state at predetermined intervals based on the voice data acquired by the voice data acquisition unit. Speech transition analysis unit that correlates the transition of the scenario stages identified by the stage identification unit with the transition of the speech state identified by the speech state identification unit. It further possesses, The comment generation unit generates comments based on the transitions between scenario stages and the transitions in speech states associated by the speech transition analysis unit. The conference support system according to claim 3.

6. A participant identification unit identifies participants based on the audio data acquired by the aforementioned audio data acquisition unit, Based on the speech in the voice data acquired by the voice data acquisition unit, a role identification unit identifies the role of the participant identified by the participant identification unit in the meeting, A participant evaluation unit evaluates the participant identified by the participant identification unit based on the role identified by the role identification unit, the transitions between scenario stages and speech states associated with the speech transition analysis unit, the transitions between scenario stages associated with the emotion transition analysis unit, and the participant's emotion transition. It further possesses The conference support system according to claim 5.

7. An evaluation receiving unit that receives evaluations from participants of the meeting regarding the comments generated by the comment generation unit, A comment evaluation storage unit stores the evaluation received by the evaluation reception unit, the comment that is the subject of the evaluation, the transition of the scenario stages and the transition of the speech state associated by the speech transition analysis unit, and the transition of the scenario stages and the transition of the participant's emotions associated by the emotion transition analysis unit, in association with each other. It further possesses The conference support system according to claim 5.

8. A meeting evaluation storage unit stores the classification of meetings, an index indicating the transition of the optimal speech state in the classified meetings, and an index indicating the transition of the optimal emotion, in association with each other. Based on the transitions between scenario stages and speech states associated by the speech transition analysis unit, and the transitions between scenario stages and participants' emotions associated by the emotion transition analysis unit, the meeting evaluation unit classifies the meeting. It further possesses, The meeting evaluation unit generates advice based on the meeting classification to bring the progression of speech patterns and the emotional changes of the meeting participants closer to the optimal indicators stored by the meeting evaluation storage unit. The conference support system according to claim 5.

9. An emotion transition analysis unit that correlates the transitions between scenario stages identified by the stage identification unit with the changes in the emotions of participants analyzed by the emotion identification unit, A speech transition analysis unit that correlates the transition of the scenario stages identified by the stage identification unit with the transition of the speech state identified by the speech state identification unit, A meeting evaluation storage unit stores the classification of meetings, an index indicating the transition of the optimal speech state in the classified meetings, and an index indicating the transition of the optimal emotion, in association with each other. Based on the transitions between scenario stages and speech states associated by the speech transition analysis unit, and the transitions between scenario stages and participants' emotions associated by the emotion transition analysis unit, the meeting evaluation unit classifies the meeting. It further possesses, The aforementioned meeting evaluation unit receives advice based on the classification of the meeting, to bring the progression of the speech state in the meeting and the progression of the emotions of the meeting participants closer to the optimal indicators stored by the aforementioned meeting evaluation storage unit. The aforementioned advice is created by the user. The conference support system according to claim 2.

10. A scenario storage step in which the computer stores a scenario that defines the progress of the meeting, A computer acquires audio data in a step of acquiring audio data relating to a meeting in accordance with the scenario stored in the scenario storage step, A step in which a computer identifies the stage of the scenario corresponding to the audio data based on the audio data acquired in the audio data acquisition step, A computer performs an emotion identification step to identify the emotions of the participants in the meeting at the stage of the scenario identified in the stage identification step, A speech state identification step in which a computer identifies the speech state of a participant in the meeting based on the speech data acquired in the speech data acquisition step, A computer generates a comment based on the emotion identified in the emotion identification step and the speech state identified in the speech state identification step. It has, In the step of identifying the speech state, at least the turn-taking, the amount of speech, and the speech itself in the meeting are identified. Meeting support methods.

11. A scenario storage step that stores the scenario that defines the progress of the meeting, The scenario storage step includes an audio data acquisition step that acquires audio data relating to a meeting in accordance with the scenario stored in the scenario storage step, A step to identify the stage of the scenario corresponding to the audio data, based on the audio data acquired in the aforementioned audio data acquisition step, An emotion identification step to identify the emotions of the participants in the meeting at the stage of the scenario identified in the stage identification step, Based on the audio data acquired in the aforementioned audio data acquisition step, a speech state identification step is performed to identify the speech state of the participants in the meeting, A comment generation step that generates a comment based on the emotion identified in the emotion identification step and the speech state identified in the speech state identification step. Have the computer run it, In the step of identifying the speech state, at least the turn-taking, the amount of speech, and the speech itself in the meeting are identified. program.

12. A scenario storage step in which the computer stores a scenario that defines the progress of the meeting, A computer acquires audio data in a step of acquiring audio data relating to a meeting in accordance with the scenario stored in the scenario storage step, A step in which a computer identifies the stage of the scenario corresponding to the audio data based on the audio data acquired in the audio data acquisition step, A computer performs an emotion identification step to identify the emotions of the participants in the meeting at the stage of the scenario identified in the stage identification step, A speech state identification step in which a computer identifies the speech state of a participant in the meeting based on the speech data acquired in the speech data acquisition step, A computer displays the emotion identified in the emotion identification step and the speech state identified in the speech state identification step, A computer receives comments created by a user based on the information displayed in the display step. It has, In the step of identifying the speech state, at least the turn-taking, the amount of speech, and the speech itself in the meeting are identified. Meeting support methods.

13. A scenario storage step that stores the scenario that defines the progress of the meeting, The scenario storage step includes an audio data acquisition step that acquires audio data relating to a meeting in accordance with the scenario stored in the scenario storage step, A step to identify the stage of the scenario corresponding to the audio data, based on the audio data acquired in the aforementioned audio data acquisition step, An emotion identification step to identify the emotions of the participants in the meeting at the stage of the scenario identified in the stage identification step, Based on the audio data acquired in the aforementioned audio data acquisition step, a speech state identification step is performed to identify the speech state of the participants in the meeting, A display step that displays the emotion identified in the emotion identification step and the speech state identified in the speech state identification step, A comment receiving step that accepts comments created by users based on the information displayed in the aforementioned display step. Have the computer run it, In the step of identifying the speech state, at least the turn-taking, the amount of speech, and the speech itself in the meeting are identified. program.

Images