JP2006018691A - Method, device and program for structuring comments - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To easily grasp a content and a flow of discussion and an influence of a comment on the discussion. <P>SOLUTION: An attribute information assignment part 16 provides an interface for inputting attribute information to a user to which the attribute information of an input comment (object comment) is assigned. The user inputs a reference comment to the input comment (object comment) by selecting a comment displayed in a comment display area, and inputs a comment kind by selecting a comment kind button. An attribute information narrowing part 20 performs the narrowing of attribute information and reflects the narrowing result to the attribute information assignment part 16. A comment storage part 12 stores input comments and attribute information assigned thereto to a (comment kind mark and reference comment mark) comment database 13. A connection unit formation part 17 forms, based on the information accumulated in the comment database 13, a connection unit of comments based on the flow of questions and answers of the comments. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an electronic conference system, and more particularly to an electronic conference support method.

  In recent years, with the development of technologies such as computers and networks, discussions and meetings using video conferencing systems, chat systems, electronic bulletin boards, mailing lists, and the like have been actively conducted.

  In this case, in order to provide a means for the non-participants of the discussion to easily grasp the contents of the discussion, the flow of the discussion, and the impact of the comments on the discussion of the participants, It is conceivable to devise a display method. It is possible to display the history of utterances in chronological order, but it is difficult to understand the relationship between utterances and utterances. It is also difficult to grasp the status of activities.

  Patent Document 1 proposes a method for displaying only a utterance that includes a specific keyword.

  Patent Document 2 and Patent Document 3 propose a method of structuring and displaying a relationship between utterances while paying attention to the utterance reference relationship (parent-child relationship).

Patent Document 2, Patent Document 4, and Patent Document 5 propose a method of structuring and displaying the relationship between statements using information on the types of statements.
JP 2003-131550 A Japanese Patent No. 2990919 JP 2002-207682 A JP-A-9-231040 JP-A-6-131396

  The above-described conventional technology has the following problems in grasping the activity status of individual and individual discussion participants with respect to the discussion.

  In the remark display method according to Patent Document 1, only the remarks of part and part are displayed, so the relationship between remarks and remarks is not displayed, and the influence of the remarks on the subject has been grasped. Can not do it.

  In the statement structuring method according to Patent Document 2, a large tree structure may be displayed, and it is difficult to grasp the activity status of individuals and individual discussion participants.

  In the method according to Patent Document 3, a large tree structure is once constructed and divided into fine tree structures. However, because it uses only information about the reference relationship (parent-child relationship) of remarks, it is important to understand the activity status of individual and individual discussion participants when dividing into fine tree structures. May be deleted.

  In the methods according to Patent Document 2 and Patent Document 4, it has been proposed to change the display color and display order depending on the type of speech. However, since only the reference relationship of the statements is used when structuring the statements, it is important to understand the activities of individuals and individual discussion participants, It may contain irrelevant statements.

  The method according to Patent Document 5 uses two kinds of utterances such as a question and an answer, manages the utterances in pairs, and enables display in pairs. However, even if the display is performed in units of pairs, it is impossible to grasp whether the questioner has agreed or not agreed through the question and answer.

  An object of the present invention is to provide a speech structuring method that allows participants or non-participants of a discussion to easily grasp the content of the discussion, the flow of the discussion, and the influence of each comment on the discussion participant's discussion. , Apparatus, and program.

  The present invention uses the information about the types of statements and the information about the reference relationship between the types of statements, and structures the statements of each speaker by referring to the question-and-answer model that models the comments about the questions and answers. Is.

  By structuring the speech of each speaker, it is possible to easily grasp the content and flow of the discussion and the impact of the speech on the discussion.

  Next, embodiments of the present invention will be described with reference to the drawings.

  As shown in FIG. 1, an electronic conference system according to an embodiment of the present invention includes a speech input unit 11, a speech storage unit 12, a speech database 13, a speech retrieval unit 14, a display device 15, an attribute information adding unit 16, and a connection unit creation. A unit 17, a concatenation unit storage unit 18, a concatenation unit tree creation unit 19, and an attribute information narrowing unit 20 are included.

  The speech input unit 11 provides a conference participant with an interface for inputting speech. FIG. 2 shows an example of a speech input interface. A conference participant inputs a speech by writing a speech in the speech input area.

  The attribute information assigning unit 16 provides an interface for inputting attribute information to a user (attribute granter) who assigns attribute information (speech type, reference speech code) of the input speech (target speech). The attribute grantor selects a comment displayed in the comment display area to input a reference message for the input message (target message), and selects a message type button to input a message type. In the present embodiment, the utterances of each speaker are classified into four utterance types: “question utterance”, “answer utterance”, “consent utterance”, and “non-agreement utterance”. Further, it is assumed that a reference / referenced relationship according to the “question and answer model” as shown in FIG. 3 exists between the messages of these four message types. In the present embodiment, by performing such classification and assumption, it becomes possible to create a connection unit in which the utterances of each speaker are structured and connected.

  The attribute information narrowing unit 20 narrows down the attribute information and reflects the narrowing result to the attribute information adding unit 16 to provide an input interface that makes it easy for the attribute giver to input the attribute information. Flowcharts of attribute information narrowing processing performed by the attribute information narrowing unit 20 are shown in FIGS. 4, 5, and 6.

  When a conference participant inputs a speech (target speech) in the speech input unit 11 (step 101) and inputs a speech type of the target speech in the attribute information adding unit 16 (in the case of A in FIG. 3) (step 101, 102) It is possible to display a display screen that facilitates the input of the reference utterance of the target utterance to the conference participants. In addition, when a conference participant inputs a utterance (target utterance) in the utterance input unit 11 and the system estimates the utterance type of the target utterance, a display screen that makes it easy to input a reference utterance of the target utterance may be displayed. It becomes possible.

  FIG. 7 shows an example of a display screen that enables selection of only a utterance that can be a reference utterance of the target utterance, for example, when the target utterance is assumed to be an answer utterance through the above processing. In FIG. 7, the conference participant inputs a speech (target speech) as “not a university?” To the speech input area, and at the same time, the conference participant becomes an attribute information granter, and “answer” is selected from the speech type button. The button is selected. Based on the processing shown in FIG. 5 (steps 111 to 115), the attribute information narrowing unit 20 narrows down that the reference utterance of the target utterance is a question utterance. Based on the result narrowed down by the attribute information narrowing unit 20, the attribute information adding unit 16 can select only (question) speech that can be a reference speech from among the speech displayed in the speech display area. .

  When a conference participant inputs a utterance (target utterance) in the utterance input unit 11 (step 101) and inputs a reference utterance of the target utterance in the attribute information adding unit 16 (steps 102 and 103) (in FIG. 3B ), It is possible to display a display screen that facilitates the input of the message type of the target message to the conference participants. In addition, when a conference participant inputs a utterance (target utterance) in the utterance input unit 11 and the system estimates a reference utterance of the target utterance, a display screen that facilitates input of the utterance type of the target utterance may be displayed. It becomes possible.

  Through the above processing, for example, when it is determined that the reference utterance of the target utterance is a question utterance, the target utterance is automatically determined as an answer utterance. An example of this screen display is shown in FIG. In FIG. 8, the conference participant inputs the speech (target speech) as “XX university?” To the speech input area, and at the same time, the conference participant becomes the attribute information granter, and “K "Where are you from?" Is selected as the reference utterance for the target utterance. Based on the processing shown in FIG. 6 (steps 121 to 124), the attribute information narrowing unit 20 narrows down that the target speech is an answer speech. The attribute information giving unit 16 automatically selects the “answer” message type button based on the result narrowed down by the attribute information narrowing unit 20.

  The utterance storage unit 12 utters the utterance input by the utterance input unit 11, the utterer, and the attribute information (utterance identification code, utterance type code, and reference utterance code) given by the utterance time and the attribute information giving unit 16. Store in database 13. An example of accumulation of information stored in the message storage unit 12 is shown in Table 1.

  The message extraction unit 14 has a function of extracting information stored in the message database 13. Specifically, information stored in the message database 13 is extracted based on the message identification code and the message time.

  The connection unit creation unit 17 creates a speech connection unit based on the flow of questions and answers. In order to create a connection unit for a message, a message type code and a reference message code assigned to each message are used. FIG. 9 shows a procedure for creating a connection unit of messages. Using the accumulation example of the database 13 shown in Table 1, the flow of processing for creating a connection unit will be described based on the flowchart of FIG.

1. The statement identification code “1” (“How is K?”) Of the question statement is substituted into the statement X (step 201).
2. Since there is a referenced utterance of the utterance X (utterance identification code “1”), the routine proceeds to step 203 (step 202).
3. The statement identification code “3” (“suppressing pitcher jan”), which is the referenced statement of the statement X (the statement identification code “1”), is substituted into the statement Y (step 203).
4). Since there is no referred utterance of the utterance Y (utterance identification code “3”), the process proceeds to step 211 (step 204).
5. A linking unit (1, 3,) is created from the utterance X (utterance identification number “1”) and utterance Y (utterance identification number “3”), and the linking unit (1, 3,) is stored in the database 13 (step 211). ).
6). In order to repeat the loop of Y, the process proceeds to step 203 (step 208).
7). The statement identification code “4” (“Two Armies”), which is the referenced statement of the statement X (the statement identification code “1”), is substituted into the statement Y (step 203).
8). Since there is a referred utterance of the utterance Y (utterance identification code “4”), the routine proceeds to step 205 (step 204).
9. The statement identification code “5” (“I think so, two troops”), which is the referenced statement of the statement Y (the statement identification code “4”), is substituted into the statement Z (step 205).
10. Concatenation unit (1, 4, 5) is created from utterance X (utterance identification code “1”, utterance Y (utterance identification code “4”, utterance Z (utterance identification code “5”)), and connected unit (1, 4, 5). 5) is stored in the database 13 (step 206).
11. In order to repeat the loop of Z, the process proceeds to step 205 (step 207).
12 Since there is no referred utterance of the utterance Y (utterance identification code “4”), the routine proceeds to step 208 (step 205).
13. In order to repeat the loop of Y, the process proceeds to step 203 (step 208).
14 Since there is no referred utterance of the utterance X (utterance identification code “1”), the process proceeds to step 209 (step 203).
15. In order to repeat the loop of X, the process proceeds to step 201 (step 209).

16. The statement identification code “2” (“Where is K from?”) Of the question statement is substituted into the statement X (step 201).
17. Since there is no referred utterance of the utterance X (utterance identification code “2”), the routine proceeds to step 210 (step 202).
18. A concatenation unit (2,) is created from the remark X (utterance identification code “2”), and the concatenation unit (2,) is stored in the database 13 (step 210).
19. In order to repeat the loop of X, the process proceeds to step 201 (step 209).
20. Since there is no question utterance, proceed to “END”.

  The connection unit storage unit 18 stores the connection unit of the message created by the connection unit creation unit 17 in the database 13. An example of accumulation of information stored in the concatenation unit storage unit 18 is shown in Table 2.

  The connection unit tree creation unit 19 uses a speech as a node based on the flow of questions and answers of a speech, expresses the connection unit of the speech in a tree structure, and displays it on the display device 15. FIG. 10 shows a flowchart for creating a tree structure from the connected units of messages. Here, the X coordinate l for displaying the question utterance, the X coordinate m for displaying the reply utterance, the X coordinate n for displaying the consent utterance or the non-agreement utterance, and the vertical interval d between the utterance and the utterance are 100, 200, 300, respectively. , 50.

1. First, the initial value of the Y coordinate y is set to 0 (step 301).
2. The utterance is taken out in a unit of connection, and the utterance identification code “1” (“How is K?”) Is substituted for the utterance A, and the utterance identification code “3” (“Justice Pitcher Jan”) is substituted for the utterance B. (Step 302).
3. Since the utterance A is not already displayed, the process proceeds to step 304 (step 303).
4). The utterance A (utterance identification code “1”) is displayed at the coordinates (100, 0) (step 304).
5. Since the utterance B is not empty, the process proceeds to step 306 (step 305).
6). The utterance B (utterance identification code “3”) is displayed at the coordinates (200, 0) (step 306).
7). The displayed message A and the displayed message B are connected by a line (step 307).
8). Since the statement C is empty, the process proceeds to step 311 (step 308).
9. 50 is added to y, and y becomes 50 (step 311).
10. Proceed to step 302 (step 312).
11. The utterances are taken out in a connected unit, the utterance identification code “1” (“What about K?”) For utterance A, the utterance identification code “4” (“Two Armies”) for utterance B, and the utterance identification code for utterance C “5” (“I think so, 2 troops”) is substituted (step 302).
12 Since the utterance A is already displayed, the process proceeds to step 305 (step 303).
13. Since the utterance B is not empty, the process proceeds to step 306 (step 305).
14 The utterance B (utterance identification code “4”) is displayed at the coordinates (200, 50) (step 306).
15. The displayed message A and the displayed message B are connected by a line (step 307).
16. Since the statement C is not empty, the process proceeds to Step 309 (Step 308).
17. The utterance C (utterance identification code “5”) is displayed at the coordinates (300, 50) (step 309).
18. The displayed message B and the displayed message C are connected by a line (step 310).
19. 50 is added to y, and y becomes 100 (step 311).
20. The utterance is taken out in the unit of connection, and the utterance identification code “2” (“Where is K from?”) Is substituted for utterance A, and utterance B and utterance C are emptied.
21. Since the utterance A is not already displayed, the process proceeds to step 304 (step 303).
22. The utterance A (utterance identification code “1”) is displayed at the coordinates (100, 100) (step 304).
23. Since the utterance B is empty, the process proceeds to step 311 (step 305).
24. 50 is added to y, and y becomes 150 (step 311).
25. Since all connected units have been extracted, proceed to the end.

  An example of the created tree structure is shown in FIG.

  In addition to what is implemented by dedicated hardware, the present invention records a program for realizing the function on a computer-readable recording medium, and the program recorded on the recording medium is stored in a computer system. It may be read and executed. The computer-readable recording medium refers to a recording medium such as a flexible disk, a magneto-optical disk, and a CD-ROM, and a storage device such as a hard disk device built in a computer system. Furthermore, a computer-readable recording medium is a server that dynamically holds a program (transmission medium or transmission wave) for a short period of time, as in the case of transmitting a program via the Internet, and a server in that case. Some of them hold programs for a certain period of time, such as volatile memory inside computer systems.

It is a block diagram of the electronic conference system of one Embodiment of this invention. It is a figure which shows the example of a speech input interface. It is a figure which shows the question and answer model of an utterance. It is a flowchart of the narrowing-down process of attribute information. It is a flowchart of the narrowing-down process of a reference message. It is a flowchart of the narrowing-down process of the message type. It is a figure which shows the example of the display screen which makes easy the input of a reference message. It is a figure which shows the example of the display screen which makes the input of utterance kind easy. It is a flowchart of the process which produces the connection unit of an utterance. It is a flowchart of the process which produces the tree structure of a connection unit. It is a figure which shows the example of a tree structure display.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Comment input part 12 Message storage part 13 Message database 14 Message extraction part 15 Display apparatus 16 Attribute information provision part 17 Connection unit creation part 18 Connection unit storage part 19 Connection unit tree preparation part 20 Attribute information narrowing-down part 101-103,111 115, 121-124, 201-209, 301-312 steps

Claims (11)

  A speech structuring method for structuring a speech by referring to a question-and-answer model that models speech related to a question and answer, using information related to the type of speech and information related to a reference relationship between the types of speech.   The method according to claim 1, wherein, when information on the type of input utterance is determined, the type of utterance in which the utterance has a reference relationship is narrowed down or determined based on the question and answer model.   The method according to claim 1, wherein, when information related to a reference relationship of speech is determined, information regarding the type of speech is narrowed down or determined based on the question and answer model.   4. The message structuring method according to claim 1, wherein the structured result is displayed in a tree structure with each message as a node.   The method according to any one of claims 1 to 4, wherein there are four types of the utterances: question utterance, answer utterance, consent utterance, and non-agreement utterance. A speech input unit for a speaker to input a speech,
An attribute information giving unit that gives attribute information including an identification code, a code indicating the type, and a reference message code that is an identification code of the message referred to by the message, for the input message;
A message structuring apparatus comprising: a linking unit creating unit that creates a linking unit of utterances based on a question and answer of utterances using a utterance type code and a reference utterance code assigned to each utterance.   When the information about the type of the input utterance is determined, the type of the utterance in which the utterance is in the reference relationship is narrowed down based on the question-and-answer model that models the utterance using the information about the reference relationship between the utterance types. The apparatus according to claim 6, further comprising an attribute information refinement unit for determining.   Attribute information that narrows down or determines information about the type of speech based on a question-and-answer model that models the speech, using information about the reference relationship between each type of speech when information about the reference relationship of the speech is determined The method according to claim 6, further comprising a narrowing portion.   The message structuring apparatus according to claim 5, further comprising a connection unit tree creating unit that expresses a connection unit of the message as a node in a tree structure and displays the connection unit on the display device.   The apparatus according to any one of claims 6 to 9, wherein the types of the utterances are four of a question utterance, an answer utterance, a consent utterance, and a non-agreement utterance.   A program for executing the message structuring method according to any one of claims 1 to 5 on a computer.

Images