JP2022024657A - Device, method and program for conference support - Google Patents

Abstract

To provide a conference support technique enabling efficient progress of a conference with improved conference quality.SOLUTION: In a conference support device 100 which supports a conference in a virtual space by a plurality of participants, a remark extraction unit 110 extracts a remark content from the voice communication or the text input of each participant received from a participant terminal through a network, so as to store into a remark history storage unit 140 as a remark history. A contribution determination unit 120 reads out the participant remark history from the remark history storage unit 140 to evaluate the degree of contribution of the remark contents to the proceeding and the purpose of the conference, according to the relation determination criterion stored in a determination criterion storage unit 150 and determine a score, so as to store into the remark history storage unit 140 the score in an associative manner with the remark content. A display unit 130 reads out the remark history of each participant from the remark history storage unit 140 to display the score in an associative manner with the remark content.SELECTED DRAWING: Figure 1

Description

The present invention relates to conference assistive technology.

In recent years, work style reforms have required improvements in work efficiency and productivity. In particular, since a lot of time is spent on meetings and meetings in business, there is a demand for a lean and highly effective method for managing meetings. In general, a facilitator is often assigned and the facilitator steers the progress of the meeting, and the quality of the meeting depends largely on the ability of the facilitator.

Patent Document 1 is a conference management device that determines the progress of a meeting based on keywords and key phrases in the statements of the conference participants, informs the conference participants of the judgment result of the progress, and supports the smooth progress of the conference. Is disclosed.

Japanese Unexamined Patent Publication No. 2011-666794

In the conference management device of Patent Document 1, since the information notified to the conference participants is the judgment result of the progress status based on the aggregation result, it is specific which statement of which participant himself / herself was evaluated and how. I can't know. The quality of the meeting will be further improved if each participant can recognize the content of his / her own remarks and the remarks of others, and the influence and contribution of these remarks on the progress of the meeting during the meeting.

The present invention has been made in view of these circumstances, and an object of the present invention is to provide a conference assistive technology capable of streamlining the progress of a conference and improving the quality of the conference.

In order to solve the above problems, the conference support device of a certain aspect of the present invention is a conference support device that supports a conference by a plurality of participants in a virtual space, and is a speech extraction unit that extracts the speech contents of the participants. , A determination unit for determining a score indicating the degree of contribution of the content of remarks to the meeting, and a display unit for displaying the score in association with the content of remarks of each participant.

Another aspect of the present invention is a conference support method. This method is a conference support method that supports a conference by multiple participants in a virtual space, and includes a step of extracting the content of the participants' remarks, a step of determining a score indicating the degree of contribution of the content of the remarks to the conference, and a step of determining the score. It includes a step of associating the score with the content of each participant's remark and displaying it.

It should be noted that any combination of the above components and the conversion of the expression of the present invention between methods, devices, systems, recording media, computer programs and the like are also effective as aspects of the present invention.

According to the present invention, the progress of the conference can be streamlined and the quality of the conference can be improved.

It is a block diagram of the conference support apparatus of this embodiment. It is a figure which shows the VR chat space exemplary. It is a figure explaining the remark history displayed on the sub-display. It is a figure which shows the example of the display mode of a participant's avatar in a VR chat space. It is a figure which shows another example of the display mode of a participant's avatar in a VR chat space. It is a figure explaining another display example of the speech history displayed on the sub-display. It is a figure explaining another display example of the speech history displayed on the sub-display. It is a flowchart which shows the meeting support procedure by a meeting support device.

FIG. 1 is a configuration diagram of the conference support device 100 of the present embodiment. The conference support device 100 includes a speech extraction unit 110, a contribution determination unit 120, a display unit 130, a speech history storage unit 140, a determination standard storage unit 150, a conference material storage unit 160, and an evaluation input unit for others. It is equipped with 170.

Participants in the conference participate in VR chat by logging in from a terminal for logging in to a chat (hereinafter referred to as "VR chat") in a VR (Virtual Reality) space such as a personal computer or a smartphone. Each terminal of the conference participant connects to the conference support device 100 via the network.

Participants hold a meeting in the VR chat space. Each participant wears an audio input / output device such as a headset and participates in VR chat using a text input device such as a keyboard. Participants can speak by voice and text input. At the time of voice input, it is converted into text data by a general voice recognition technology on the participant's terminal. The voice of each participant is transmitted to each terminal via the network and can be heard by all the participants.

FIG. 2 is a diagram illustrating an example of a VR chat space. In this example, six participants' avatars A to F are displayed around the conference desk 10 in the VR chat space. In the VR chat space, the main display 20 for displaying materials used in the proceedings such as the agenda of the meeting, presentation materials, review materials, and the sub that displays the remark history of the participants converted into character data by voice recognition. The display 30 and the display 30 are arranged.

Each configuration of the conference support device 100 will be described again with reference to FIG. The speech extraction unit 110 extracts the speech content from the participant's voice call or text input received from the participant's terminal via the network, and stores it in the speech history storage unit 140 as a speech history.

The contribution degree determination unit 120 reads out the statement history of the participants from the statement history storage unit 140, and the degree to which the statement content contributes to the proceedings / purpose of the meeting according to the relevance determination criteria stored in the determination criterion storage unit 150. Is evaluated to determine the score, and the score is associated with the content of the statement and stored in the statement history storage unit 140.

The contribution determination unit 120 generates a learning model using artificial intelligence (Artificial Intelligence) that learns the correlation between the agenda / purpose of the meeting and the content of the statement, and determines the degree of contribution of the content of the statement to the conference using AI. May be good. In this case, the relevance determination criterion stored in the determination criterion storage unit 150 is a trained model. Alternatively, the facilitator of the meeting may set the relevance determination criteria, and the contribution determination unit 120 may determine the contribution of the content of the statement to the conference in light of the relevance determination criteria set by the facilitator. The contribution determination unit 120 can immediately switch the relevance determination criteria according to the agenda and purpose of the meeting.

The display unit 130 displays the participants in the virtual space with an avatar. The virtual space is displayed on each participant's terminal. Whether or not the participant is speaking is determined by the voice / volume detection function that is common in the participant's terminal. The display unit 130 changes the display mode of the participant's avatar in the virtual space so that the participant can know that he / she is speaking during voice utterance or text input. For example, if a participant is speaking, the participant's avatar may move his body or mouth or change his facial expression so that other participants can see what he or she is saying. ..

The display unit 130 reads out the material used in the conference from the conference material storage unit 160 and displays it on the main display 20. Further, the display unit 130 reads the remark history of the participant from the remark history storage unit 140, associates the score with the remark content, and displays it on the sub-display 30.

The display unit 130 displays the purpose of the conference and the relevance judgment criteria on another display in the virtual space so that the participants of the conference can always know what kind of remarks can contribute to the progress of the conference. good.

The other person evaluation input unit 170 accepts the evaluation of another participant with respect to the content of the participant's remark, associates the evaluation of the other with the content of the remark, and stores it in the remark history storage unit 140. For example, when evaluating the content of a participant's remark as being good by another speaker, the display unit 130 enables the other participant to input an evaluation such as good for the content of the remark. The total number of evaluations is displayed as the good evaluation number in association with the content of the statement.

FIG. 3 is a diagram illustrating a speech history displayed on the sub-display 30. The remark history of the participants is displayed in a list on the sub-display 30. The speech history may be displayed on the timeline in chronological order, or may be displayed in descending order of evaluation.

The remark history includes an avatar icon 40, a remark 42, a tag 44, another person's evaluation 46, and a contribution degree 48.

The avatar icon 40 is an icon of the participant who made a statement. The remark 42 includes the speaker identification information, the remark time, and the remark content. The speaker identification information is a user name or the like, and is indicated by A to F here. The speaking time is the time when the speaker made a statement. The remark content is a text data display of the remark content of the speaker.

The tag 44 is identification information indicating the type of the content of the statement, such as "# Q & A", "# Response", and "# Suggestion". For example, if the meeting is brainstorming, it can be used to identify ideas from the speech history.

The other person's evaluation 46 includes the evaluation and the number of evaluations by other participants for the content of the speaker's remarks. For example, the number of positive / negative evaluations such as good / bad and the number of evaluations such as good / bad. Negative evaluations such as bad are optional and may not be included if they atrophy the participants' remarks.

Contribution level 48 is a score indicating the degree to which the participants' remarks contributed to the agenda / purpose of the meeting, and is evaluated by AI or the facilitator. Here, it is shown as a numerical value in 10 stages (the larger the numerical value, the higher the degree of contribution). If it is difficult to judge by the one-dimensional evaluation value, some items may be judged and the evaluation value may be displayed in multiple dimensions. The display of multidimensional evaluation values includes display by various graphs such as radar charts.

The display unit 130 can make it easier to see the highly evaluated remarks by changing the characters and colors of the remark history according to the contribution score and the good evaluation number. In addition, the display unit 130 can display the remarks having a high score on another display to raise the awareness of the participants.

When there is a technical term in the remark, the display unit 130 can automatically extract the technical term from the remark history and display the technical term as an auxiliary in another window, or display the explanation of the technical term. ..

FIG. 4 is a diagram showing an example of a display mode of a participant's avatar in a VR chat space. Participant F is currently speaking, and a balloon 50 including a text of the content of the statement is displayed in the avatar of participant F.

The display unit 130 visualizes the participant who is about to speak by changing the display mode of the participant's avatar based on the participant's breathing. For example, suppose Participant A is about to speak and his breathing is changing. In this case, a balloon 52 containing no text is displayed in the avatar of the participant A so that other participants can know that the participant A is about to speak. When a participant changes his or her breathing when trying to speak, or when he or she says something, a balloon is displayed on the participant's avatar, or the participant's avatar's body or mouth is moved or his facial expression is changed. This highlights the participants who are willing to speak.

In the real space, you can feel the breathing of the participants, so you can see who is trying to speak, but in VR chat, you cannot even know the breathing of the participants. By expressing the change in the breathing of the participants as the change in the display mode of the avatar in this way, it becomes possible to distinguish the participant who is trying to speak even in the VR chat.

FIG. 5 is a diagram showing another example of the display mode of the participant's avatar in the VR chat space. It is assumed that the scores indicating the degree of contribution of the statements made by the two participants C and D are low. The display unit 130 changes the display mode of the participant's avatar according to the degree of contribution of the participant's remarks. For example, by thinning the avatars of participants C and D or displaying them with a dotted line, the avatars of the participants with a low contribution of speech are faded out. On the contrary, the avatars of the participants who contributed a lot to the remarks are displayed darkly. Alternatively, the display unit 130 may change the display mode of the participant's avatar according to the number of remarks. If the number of remarks is large, the avatar is thickened, and if the number of remarks is small, the avatar is thin.

FIG. 6 is a diagram illustrating another display example of the speech history displayed on the sub-display 30. The display mode of the contribution score of the content of the statement by the display unit 130 may differ depending on the participant. For example, the display unit 130 displays the scores for all the participants' remarks to the facilitator of the conference, but displays only the scores for their own remarks to the other participants, and other participants. The score for the content of the statement is hidden. FIG. 6 is a remark history displayed on the terminal of the participant B, and the score for the remarks of the participant B itself is displayed as “7”, but the contribution degree to the remarks of the other participants is not displayed. It has become.

FIG. 7 is a diagram illustrating another display example of the speech history displayed on the sub-display 30. A link is attached to the content of the statement based on the relationship between the participants' statements, and the content of the statement is displayed in a cluster. In this example, the participant F responds to the remark 50 of the participant A with the remark 51, the participant B responds with the remark 52, and the participant E responds with the remark 55. Further, the participant C responds to the remark 52 of the participant B with the remark 53, and the participant E responds to the remark 53 of the participant C with the remark 54. Further, the participant B responds to the remark 55 of the participant E with the remark 56, and the participant D responds with the remark 57.

By displaying the statement history in the cluster in this way, the relationship between the statements can be visualized, and it becomes clear at a glance who the statement was the trigger for the proceedings. In this example, since a large number of remarks are continued triggered by the remark 50 of the participant A, it can be seen that the contribution of the remark 50 of the participant A to the meeting is large. The contribution degree determination unit 120 may calculate the contribution degree score according to the number of links of remarks.

FIG. 8 is a flowchart showing a conference support procedure by the conference support device 100.

Participants log in to the VR chat space from the terminal and start a conference using voice call / text input and screen sharing in the VR chat space (S10).

The remark extraction unit 110 extracts the remark contents of the participants during the meeting at any time and registers them in the remark history storage unit 140 (S20).

The contribution degree determination unit 120 calculates the contribution degree score according to the relevance determination standard for the content of the participant's remarks and outputs the score to the remark history storage unit 140 (S30). The relevance judgment may be performed by AI or by the facilitator. However, since the relevance judgment may involve the flow of proceedings before and after the statement, the score is not calculated immediately after the statement is made, and the display timing of the score is different each time. In addition, the score calculation standard differs depending on the content of the proceedings.

In addition, the contribution determination unit 120 can also quantify the contribution to the conference by calculating the average evaluation score of each participant and the total number of good evaluations. Further, the contribution determination unit 120 can quantify the quality of the meeting by calculating the average value of the scores of all the participants.

The display unit 130 reads the remark history from the remark history storage unit 140, and displays the score together with the remark contents of the participants on the sub-display 30 (S40). Finally, the remark history becomes the minutes of the meeting.

An example of conference support by the conference support device 100 is shown below. As a precondition of the embodiment, it is assumed that a meeting for reviewing the specifications of a certain function using the meeting support device 100 is held by five people (denoted as A, B, C, D, and E). In this example, it is assumed that A is the facilitator and B is the explainer. The evaluation score is 1 to 10 (10 is the best).

(1) A creates a VR chat room for this system and invites the rest of the members.
Suppose that E has contacted you that your login will be delayed.

(2) Participants individually log in to VR chat using a terminal.

(3) When everyone except E logs in, facilitator A displays the materials used in the meeting on the main display and explains the contents of the proceedings.
The facilitator A or the explainer B sets the agenda content and the relevance judgment criteria in the judgment criteria storage unit 150 so that the AI can use them, and starts the meeting.

(4) First, the explainer B explained the specifications.
The explanation of the explainer B is stored in the remark history storage unit 140, and is displayed on the sub-display 30 by the display unit 130.
Since the remark of the explainer B was the explanation of the specification in the facilitator A, "explanation" was set in the tag indicating the type of the remark content.
Since the content of this statement by the explainer B is "explanation", the contribution determination unit 120 may or may not evaluate the content of this statement by the explainer B.

(5) Next, C asked B a question as a question. Suppose that the question content of C was an important question.
Facilitator A set "Question" as the tag of the content of the statement because C's statement was a question.
B and D set a good evaluation for C's remark, and the good evaluation number of C's remark is "2".
The evaluation score for the content of C's remark by AI was "8" after evaluating the importance of the remark.

(6) It is assumed that B can accurately answer the question of C.
Facilitator A set "response" to the tag because B's remark is a reply.
C sets a good evaluation for B's reply, and the good evaluation number of B's reply is "1".
The evaluation score for B's response by AI was "8" after evaluating the response content.

(7) Next, D made a statement and made a proposal. However, it is assumed that the proposal has little relation to the agenda.
Facilitator A has set the tag "suggestion" for D's remarks.
Since there was no good evaluation for D's remark, the good evaluation number of D's remark is "0".
The evaluation score for D's proposal by AI was "2" in consideration of the low relationship.
At this point, D can recognize the error in his proposal.

(8) Here, it is assumed that E is late and logs in.
E was able to see what was said before E logged in by looking at the statement history.
E performed a good evaluation for the questions and answers of (5) and (6), and the good evaluation was incremented by +1 for each.

(9) Next, E asked a question related to the question of C in (6). Suppose this question wasn't very important.
Facilitator A set "Question" in the tag because E's remark was a question.
Since there was no good evaluation for E's remark, the good evaluation number of E's remark is "0".
The evaluation score for the question and answer of E by AI was "5" according to the importance.

(10) It is assumed that B cannot answer the question of E accurately.
Facilitator A set "response" to the tag because B's remark is a reply.
Since there was no good evaluation for B's reply, the good evaluation number for B's reply is "0".
The evaluation score for B's response by AI was "2" because an accurate response could not be made.
B realized that the response was not accurate at this point.

Below, the meeting will continue, and the content of each participant's remarks will be evaluated. In addition to displaying the score for one's remark to each participant in real time, the AI or facilitator may propose advice to each participant for the content of the remark.

As described above, according to the conference support device 100, various information can be shared among the participants by holding the conference in the VR chat space. By daring to use VR chat, it will be equipped with an individual voice input / output device, it will be easier to distinguish the speaker and digitize the speech, and the content of the speech will be visualized and displayed in the VR space to participate. Can be shared by everyone. Since facilitators and participants obtain shared information in real time, it is possible to eliminate the waste of overhearing and re-listening by referring to, for example, the remark history. Furthermore, by making it possible to receive evaluations by AI and evaluations from other participants for the remarks of individual participants, it is possible to suppress derailment of the agenda, which leads to efficiency and quality improvement of the meeting. be able to.

The present invention has been described above based on the embodiments. It is understood by those skilled in the art that the embodiments are exemplary and that various modifications are possible for each of these components and combinations of processing processes, and that such modifications are also within the scope of the present invention. ..

100 Conference support device, 110 Speech extraction unit, 120 Contribution judgment unit, 130 Display unit, 140 Speech history storage unit, 150 Judgment standard storage unit, 160 Conference material storage unit, 170 Others evaluation input unit.

Claims (6)

A conference support device that supports conferences by multiple participants in virtual space.
A remark extraction unit that extracts the remarks of participants, and a remark extraction unit
A contribution judgment unit that judges the score indicating the contribution of the content of the statement to the meeting, and
A conference support device including a display unit that displays the score in association with the content of each participant's remark. It also includes an other person's evaluation input unit that inputs the evaluation of other participants to the content of the participant's remarks.
The conference support device according to claim 1, wherein the display unit displays the content of each participant's remarks in association with the evaluations of other participants. The display unit displays the score for the content of the remarks of all the participants among the plurality of participants in the facilitator of the conference, but the participants other than the facilitator display the score for the content of their remarks. The conference support device according to claim 1 or 2, wherein the score is displayed while the score is hidden with respect to the content of statements made by other participants. Any of claims 1 to 3, wherein the display unit visualizes a participant who is trying to speak by changing the display mode of the participant's avatar in the virtual space based on the participant's breathing. Conference support device described in Crab. The conference support device according to any one of claims 1 to 4, wherein the display unit changes the display mode of the participant's avatar in the virtual space according to the score for the participant's remark content. A conference support program that supports conferences by multiple participants in virtual space.
Steps to extract the content of participants' remarks and
Steps to determine the score that indicates the contribution of the content of the statement to the meeting,
A conference support program characterized by having a computer perform a step of associating and displaying the score with the content of each participant's remark.

Images