JP3751898B2 - Dialogue simulation apparatus and dialogue support method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a dialogue simulation apparatus.
[0002]
[Prior art]
For example, in preparation for a regular report meeting of a meeting or a meeting in a foreign language, it is expected to prepare remarks to anticipate and respond to comments and questions from multiple participants who are scheduled to participate Things are often done. In meetings and the like, the purpose of the meeting is clear and discussions are being developed, so it is easy to predict the contents of comments from participants in advance and to develop a scenario for discussion development.
[0003]
However, when chatting with various people at parties, social gatherings, dinners, and other occasions, especially when trying to talk with people or foreigners who are meeting for the first time, In reality, it is often difficult to find a topic, such as "Isn't it interested?" Although there are a lot of topics about work and my field of specialization, there should be many people who have had experiences of feeling awkward because conversation was interrupted immediately because there were few topics when chatting. It is.
[0004]
In such a case, it is encouraging if you can rehearse in advance while selecting a suitable topic in accordance with the situation, assuming a situation where you interact with others.
[0005]
[Problems to be solved by the invention]
Conventionally, there are many large-scale simulators such as simulators for pilot training of aircraft and simulators for the purpose of training astronauts for landing on the moon. For example, Japanese Patent Laid-Open No. 2000-122520 relates to a virtual reality simulator and a simulation method thereof, and particularly to cope with an emergency situation such as an accident occurring inside a large-scale and complicated three-dimensional structure such as a chemical plant or a thermal power plant. Virtual simulation training can be performed on a daily basis.
[0006]
However, there is no dialogue simulator that can be used at the individual level for the purpose of dialogue training, assuming the situation of dialogue with other people.
[0007]
In view of the above, an object of the present invention is to provide a dialog simulation apparatus and a dialog support method that are used at an individual level for the purpose of dialog training, assuming a scene of dialog with another person.
[0008]
[Means for Solving the Problems]
The simulation apparatus of the present invention sets an input means for inputting a desired scene and a condition for setting a desired conversation partner, and sets the scene and the conversation partner based on the condition input by the input means. Setting means, acquisition means for acquiring latest information for each field based on the conditions, display means for displaying the latest information of the field selected by the user among the latest information collected for each field, and the conditions and topics The presenting means includes a presenting means for presenting advice to the user so that a dialog suitable for the selected field can be performed, and a generating means for generating the dialogue partner's dialogue in response to the content of the speech from the user, When a topic word is extracted from the content of a statement from the user, the dialogue partner's dialogue corresponding to the topic word is generated.
[0009]
According to the present invention, it is possible to easily train a dialogue with various people in various scenes encountered at an individual level.
[0010]
Preferably, the presenting unit inserts a desired topic word included in the latest information displayed on the display unit as the advice so that the user can perform a conversation about the latest information displayed on the display unit. By doing so, we will present a dialogue template that completes the dialogue.
[0011]
Further, the presenting means presents a speech that the user preferably utters as the advice.
[0012]
The dialogue support method of the present invention sets the scene and the dialogue partner based on the input desired scene and the condition for setting the desired dialogue partner, and acquires the latest information for each field based on the condition. In the latest information collected by field, the latest information of the field selected by the user is displayed, and advice is presented to the user so that a dialogue suitable for the condition and the field selected as the topic can be performed. When there is a utterance from the user, the dialogue partner corresponding to the content of the utterance is generated, and when a topic word is extracted from the utterance content from the user, the conversation partner corresponding to the topic word is generated. It is characterized by generating the dialogue.
[0013]
According to the present invention, it is possible to easily train a dialogue with various people in various scenes encountered at an individual level.
[0014]
Preferably, a dialogue template that completes the user's dialogue is presented as the advice by inserting a desired topic word included in the displayed latest information so that a conversation about the latest information displayed can be performed. .
[0015]
In addition, as the advice, a speech that is desired to be uttered by the user is presented.
[0016]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
[0017]
FIG. 1 shows a schematic configuration example of a dialogue simulation apparatus according to the present embodiment.
[0018]
The input unit 2 is used to input a scene for interacting with a desired conversation partner and conditions necessary for setting the desired conversation partner, and for the user to input various instructions to the system.
[0019]
The conditions for setting the scene include, for example, when the desired scene is a party venue, the party type, the date and time the party is held, the location (country name, place name, etc.), and the purpose of the party For example, data that can be known from the invitation of the party, formal or informal, and data that can be input within the range that the user can understand, such as the age group and expertise of the attendees. . Preferably, as detailed conditions as possible are input. In addition, examples of conditions for setting a conversation partner include sex, age, expertise, hobbies, occupations, relationships with users, and the like. Again, preferably, it is desirable to input as detailed conditions as possible.
[0020]
The scene setting unit 3 combines CG (computer graphics) components stored in advance in the component storage unit 4 based on the scene conditions input to the input unit 2, for example, a meal at a party venue or restaurant, etc. Are created three-dimensionally.
[0021]
The dialogue partner setting unit 5 creates a dialogue partner desired by the user by combining CG (computer graphics) components stored in advance in the component storage unit 4 based on the scene conditions input to the input unit 2.
[0022]
Based on various conditions for setting a scene and a conversation partner input from the input unit 2, the information acquisition unit 6 collects the latest information that meets these conditions, for example, using the Internet. .
[0023]
The information storage unit 7 stores the latest information acquired by the information acquisition unit 6. In addition, common-sense information that varies from country to country and from scene to scene may be stored in advance.
[0024]
The information selection unit 8 selects appropriate information from the information storage unit 7 based on the scene input from the input unit 2, the conditions of the conversation partner, and other instructions from various users.
[0025]
The scenario storage unit 9 stores a plurality of scenarios for dialogues of all kinds (formal, informal, female, male, etc.) in all fields that are expected to be encountered in various scenes. . The simplest scenario is, for example, a dialog (dialog) between a user and a conversation partner.
[0026]
For example, in a party venue held in the United States, when a Japanese user interacts with an American first-time male, start with a first-time greeting, create a self-introduction, and start a conversation. If there is something that should be called a series of dialogue strategies to develop, a dialog for greetings with men who meet Americans for the first time (this is often composed of clichés), followed by For example, a dialog for dialogue on a desired topic can be given.
[0027]
For example, in a party venue held in the United States, a scenario for a Japanese user to interact with an American man in the political field, for example, about “election” is shown in FIG. Also good.
[0028]
In the dialog shown in FIG. 2, the user or the system side inserts a desired word into the “x1” and “x2” portions. Here, the words included in the displayed information inserted in the portions “x1” and “x2” are called topic words.
[0029]
In this embodiment, the user inserts a desired word (topic word) in the latest information, for example, into the “x1” or “x2” portion, and the user utters the dialogue to use the latest topic. The training of the dialogue that was possible can be done now. At the same time, the topic, manners and common sense of the scene can be acquired.
[0030]
In FIG. 2, a Japanese dialog is shown, but this may be an English dialog. In this case, if the user's native language is not English, it is also practiced for a foreign language.
[0031]
The scenario stored in the scenario storage unit 9 may have various variations according to the user's utterance content based on the scenario as shown in FIG.
[0032]
The scenario selection unit 10 selects an optimal scenario from the scenario storage unit 9 based on the scene input from the input unit 2, the conditions of the conversation partner, other instructions from various users, the user's remarks, and the like.
[0033]
The dialogue control unit 11 generates advice to be presented to the user or emits it on the system side so that the dialogue can proceed between the system side (conversation partner) and the user according to the scenario selected by the scenario selection unit 10. This is for selecting a power line.
[0034]
The voice input unit 12 is for inputting a user's speech (uttered speech). The speech of the user input to the speech input unit 12 is recognized by the speech recognition unit 13 and further analyzed by the analysis unit 14 from the content of the speech into, for example, a dialogue template (template) on the user side of the scenario shown in FIG. For example, a topic word inserted in “x1” is extracted. When the topic word is extracted, the dialogue control unit 11 requests the dialogue generation unit 15 to generate the next dialogue partner dialogue in order to generate the dialogue partner dialogue corresponding to the topic word. At that time, if necessary, the scenario selection unit 10 reselects the scenario, the information selection unit 8 selects necessary information, and the information acquisition unit 6 updates the latest topic word corresponding to the extracted topic word. Information may be acquired.
[0035]
The dialogue generation unit 15 is for generating dialogue dialogue, for example, to respond to the user's utterance content (if the user's utterance content includes a topic word). For example, a topic is generated by inserting a topic word extracted from the content of a user's speech into “x1” included in the dialogue template on the system side of the scenario shown in FIG.
[0036]
The dialogue selected by the dialogue control unit 11 and the dialogue generated by the dialogue generation unit 15 are synthesized by the speech synthesis output unit 16 in correspondence with the dialogue partner set by the dialogue partner setting unit 5, and As a remark, it is output from a speaker or the like.
[0037]
The display control unit 17 performs control for displaying on the display 20 the scene set by the scene setting unit 3 and the dialog partner set by the dialog partner setting unit 5.
[0038]
The conversation partner control unit 18 controls the expression, movement, and manner of speaking of the conversation partner based on the analysis result of the speech content input by the user in the analysis unit 14. In other words, the analysis unit 14 determines whether or not the content of remarks input from the user includes remarks that are taboo, contrary to manners, content that the conversation partner is pleased with, or content that makes you feel bad. You may make it do. Based on the determination result, the conversation partner control unit 18 can convey the fact by a facial expression when, for example, the user hears a violation of manners.
[0039]
The control part 1 is for controlling each said parts 2-18.
[0040]
Next, processing operations of the dialogue simulation apparatus in FIG. 1 will be described with reference to flowcharts shown in FIGS.
[0041]
First, with reference to FIG. 4, the processing operation in the preparation stage until the dialogue is started will be described. The user can input a desired scene condition (for example, type = party, purpose = for deepening social gatherings after the end of the conference, date / time = ○ month × day 12:00 pm, informal, etc.) The user inputs the conditions of the desired conversation partner (for example, gender = male, nationality = America, age = 40 years, specialty = information engineering, hobby = golf, occupation = professor at Y University, relationship with user = first meeting) (Step S1, Step S2).
[0042]
The scene setting unit 3 creates a scene based on the inputted scene conditions, and the dialogue partner setting unit 5 creates a dialogue partner based on the inputted dialogue partner conditions (steps S3 and S4).
[0043]
The information acquisition unit 6 collects the latest information from the Internet, for example, using the input condition as a keyword (step S5). In this case, for example, it is desirable that the latest information in all fields such as current affairs, incidents, entertainment, food and sports in the United States can be collected. However, the latest information corresponding to a predetermined field may be collected.
[0044]
The display control unit 17 displays the CG scene created by the scene setting unit 3, the CG dialogue partner created by the dialogue partner setting unit 5 on the display 20, and the latest information collected by the information collection unit 6. A button for selecting the field of information is displayed (step S6, step S7).
[0045]
FIG. 3 shows a screen display example displayed on the display 20 of the dialogue simulation apparatus.
[0046]
As shown in FIG. 3, this display screen includes a display area W1 in which CG scenes and dialogue partners are displayed, an area W2 in which a selection button for selecting a field of information is provided, and the latest information. The display area W3 displays information and the display area W4 displays advice to the user.
[0047]
In step S7 in FIG. 4, the area W1 and the area W2 are displayed among the areas shown in FIG.
[0048]
In this state, when the user selects a button corresponding to a desired field among a plurality of selection buttons displayed in the area W2 in FIG. 3 using a pointing device such as a mouse (step S11 in FIG. 5), the scenario The selection unit 10 selects a scenario suitable for the selected field and the conditions input in Step S1 and Step S2 from the scenarios stored in the scenario storage unit 4 (Step S12).
[0049]
Further, the information selection unit 8 reads out information (latest information, general common sense, etc.) belonging to the relevant field selected by the user from the information storage unit 7, and the information is stored in the area W3 in FIG. It is displayed (step S13, step S14).
[0050]
Note that the fields displayed in the area W2 in FIG. 3 are major classification fields, and when any of these fields is selected, the minor classification is further displayed, and a desired field is selected from among them. You may do it.
[0051]
For example, as the above conditions, the scene conditions are type = party, purpose = to deepen social gathering after the end of the conference, and the desired conversation partner conditions are gender = male, nationality = America, age = 40 years old If the field selected by the user is the sub-category “election” belonging to “politics”, the scenario selected in step S12, that is, the dialog may be, for example, as shown in FIG. . In addition, although the scenario shown in FIG. 2 is Japanese, it is actually English. However, here, for simplicity of explanation, the case of Japanese is taken as an example.
[0052]
Now, when the scenario as shown in FIG. 2 is selected, when it is predetermined that the user starts the dialogue (for example, the user may set the dialogue), the dialogue control unit 11 2, the role “A” in FIG. 2 is determined as the user, and the role “B” is determined as the system, that is, the conversation partner. Therefore, the dialog control unit 11 first asks the user “A to perform a dialog according to the scenario, that is, a dialog suitable for the condition input by the user and the field selected as the topic by the user. Is displayed in the area W4 of FIG. 3 (step S15). For example, as an advice, a desired topic word included in the latest information displayed in the area W4 as a speech that the user preferably utters so that the user can perform a conversation about the latest information displayed in the area W3. A dialogue template that completes the dialogue by inserting “”, that is, for example, “Is interested in x1 (what election)?” Corresponding to the first dialogue of “A” in this case, is displayed in the area W4. You may do it.
[0053]
Since the information selected for “election” (in this case, the latest information) is displayed in the area W3 of the screen, the user looks at this and the area W4, for example, “To the“ American presidential election ” Are you interested? " In the above notation, the portion surrounded by “” is a topic word inserted by the user.
[0054]
The user's speech is input from the voice input unit 12, is recognized by the voice recognition unit 13, and is further analyzed by the analysis unit 14 (steps S21 to S23).
[0055]
As a result of analyzing the content of the user's speech by the analysis unit 14, it is assumed that a topic word is extracted from the content of the speech (step S25). For example, in this case, “American presidential election” is extracted. For example, the inserted topic word can be extracted by obtaining a difference between the content of the user's statement and the dialogue template displayed as advice.
[0056]
The dialogue control unit 11 first checks whether information corresponding to the extracted topic word has already been acquired (stored in the information storage unit 7) (step S28). At this time, when the information is not stored in the information storage unit 7, the information acquisition unit 6 may be requested to acquire the latest information corresponding to the topic word. In this case, the information acquisition unit 6 may acquire information from the Internet, for example, using the passed topic word as a keyword (step S29).
[0057]
On the other hand, when the information corresponding to the extracted topic word has already been acquired (stored in the information storage unit 7), the dialogue control unit 11 prepares the dialogue on the system side (the dialogue partner). . That is, a dialogue-corresponding dialogue corresponding to the extracted topic word is generated (step S30). Here, the dialogue that the dialogue partner should next utter is, for example, the first dialogue in FIG. 2B in this case. In this dialogue “Of course. Who will win x1 (what election)?”, The topic word “American presidential election” extracted from the user's remarks needs to be included in “x1”. Therefore, the dialogue generation unit 15 inserts “American presidential election” into the “x1” portion of the dialogue, and generates the dialogue “Of course. Who thinks the American presidential election will win?” To do.
[0058]
Furthermore, when the analysis unit 14 determines in step S23 that the utterance content input from the user includes, for example, a taboo utterance contrary to manners, the determination result corresponds to the determination result. Determine the facial expression and movement of the conversation partner and how to speak (step S25). For example, the expression of the conversation partner may be strengthened, or the tone of the voice may be changed so as to be angry.
[0059]
In step S24, when a topic word is not extracted from the content of a statement from the user, that is, for example, when a phrase of greeting is spoken, or a dialogue template for inserting a topic word as advice is displayed. If it is not, it is not expected that the topic word is extracted from the content of the user's remarks. Therefore, the dialogue control unit 11 simply selects the next system side (conversation partner) dialogue from the scenario. Good.
[0060]
In step S26, the speech on the system side (the conversation partner) generated or selected in step S30 is synthesized by speech and output. After that, in response to the comment of the dialogue partner that has been output, the dialogue that is desired to be spoken by the user or the template of the dialogue is displayed as an advice on the screen area W4 so that the dialogue can be performed in accordance with the previously selected scenario. (Step S27).
[0061]
For example, in the area W4, “I don't know. For now, x2 (the name of the first candidate) is considered dominant” is displayed. Thereafter, in the same manner as in the description of step S15 in FIG. 5, the user selects “for example, John W” as a topic word from the information displayed in the area W3, for example. However, he says, “I think John W is dominant.” The subsequent processing operations are the same as described above, and the processing shown in FIG. 6 is repeated until the scenario ends.
[0062]
As described above, according to the above-described embodiment, the user can perform a conversation while developing a story about the desired conversation partner and the desired topic by inputting a desired scene and a condition of the desired conversation partner. Appropriate advice is displayed so that prior dialogue training can be performed easily. In particular, as the above-mentioned advice, the dialogue that is desirable to be uttered by the user is displayed, which is useful for mastering phrases such as greetings. Furthermore, as the above-mentioned advice, the latest information can be obtained by presenting a dialogue template that completes the dialogue by inserting a desired topic word included in the latest information that is already displayed as a dialogue that the user preferably utters. Since conversational training can be performed on the topic, realistic conversational training can be performed in accordance with actual conversational situations. Also, by conducting the above dialogue in a foreign language, dialogue training in a foreign language becomes possible. In addition, since the latest information and general common sense, etc., which are collected and selected based on the desired scene and the conditions of the desired conversation partner, and which are most suitable for the desired conversation partner, are presented, preliminary knowledge useful for actual conversations Can be obtained.
[0063]
In the above-described embodiment, the scenario is described in which the system side automatically selects in step S12 of FIG. 5. However, the present invention is not limited to this, and the system side selects the field selected by the user in step S12. A plurality of optimal scenario candidates may be selected from the input conditions, and the user may select a desired scenario.
[0064]
The method of the present invention described in the embodiment of the present invention (in particular, the method of the present invention corresponding to the processing operation shown in FIGS. 4 to 6) is a magnetic disk as a program that can be executed by a computer. It can also be stored and distributed in a recording medium such as a floppy disk or hard disk, an optical disk (CD-ROM, DVD, etc.), or a semiconductor memory.
[0065]
In addition, this invention is not limited to the said embodiment, In the implementation stage, it can change variously in the range which does not deviate from the summary. Further, the above embodiments include inventions at various stages, and various inventions can be extracted by appropriate combinations of a plurality of disclosed constituent requirements. For example, even if some constituent elements are deleted from all the constituent elements shown in the embodiment, the problem (at least one of them) described in the column of the problem to be solved by the invention can be solved, and the column of the effect of the invention If at least one of the effects described in (1) is obtained, a configuration from which this configuration requirement is deleted can be extracted as an invention.
[0066]
【The invention's effect】
As described above, according to the present invention, it is possible to easily train dialogues with various people in various scenes encountered at the individual level.
[Brief description of the drawings]
FIG. 1 is a diagram showing a schematic configuration example of a dialogue simulation apparatus according to an embodiment of the present invention.
FIG. 2 is a diagram showing an example of a dialogue scenario.
FIG. 3 is a diagram showing a screen display example of the dialogue simulation apparatus in FIG. 1;
4 is a flowchart for explaining a processing operation of the dialogue simulation apparatus of FIG. 1;
FIG. 5 is a flowchart for explaining a processing operation of the dialogue simulation apparatus in FIG. 1;
6 is a flowchart for explaining the processing operation of the dialogue simulation apparatus in FIG. 1;
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Control part 2 ... Input part 3 ... Scene setting part 4 ... Parts storage part 5 ... Dialog partner setting part 6 ... Information acquisition part 7 ... Information storage part 8 ... Information selection part 9 ... Scenario storage part 10 ... Scenario selection part 11 ... Dialog control unit 12 ... Voice input unit 13 ... Speech recognition unit 14 ... Analysis unit 15 ... Dialogue generation unit 16 ... Speech synthesis unit 17 ... Display control unit 18 ... Dialog partner control unit 20 ... Display

Claims (4)

Desired scene type, date / time, place, purpose, and conditions for setting the scene including at least one of the formal and informal scenes, and the gender, age, and expertise of the desired conversation partner An input means for inputting conditions for setting the conversation partner including at least one of hobbies, occupations, and relationships with users ;
A scenario storage means for storing a plurality of scenarios for dialogue including dialogue patterns of dialogue between the user and the conversation partner, in which a topic word insertion position is predetermined;
Information collecting means for collecting the topic words via a network using, as keywords, the scene input by the input means and the conditions for setting the conversation partner for each of a plurality of predetermined fields; ,
Information storage means for storing topic words collected by the information collection means;
Scenario selection means for selecting a scenario belonging to the field selected by the user from among the plurality of fields among the scenarios stored in the scenario storage means in a scene that satisfies the condition input by the input means. When,
Information selection means for selecting a topic word belonging to the field selected by the user from among the plurality of fields among the topic words stored in the information storage means;
Display means for displaying at least the topic word selected by the information selection means and the user's dialogue template in the scenario selected by the scenario selection means;
Voice input means for inputting a user's speech;
Extraction means for extracting a topic word in the user's speech by obtaining a difference between the user's speech input by the voice input means and a dialogue template displayed by the display means;
Generating means for generating the dialogue partner dialogue by inserting the topic word extracted by the extraction means at a predetermined insertion position of the dialogue partner dialogue template in the selected scenario;
Voice output means for synthesizing and outputting the speech generated by the generation means;
An interactive simulation apparatus characterized by comprising: CG of the scene and conversation partner ( Computer graphics ) Component storage means for storing CG components for creating an image;
A scene setting means for creating a CG image of the scene using the CG parts stored in the parts storage means based on the conditions for setting the scene input by the input means;
Based on the conditions for setting the dialogue partner input by the input means, the dialogue partner setting means for creating a CG image of the dialogue partner using the CG component stored in the component storage means;
  Further comprising
    The dialogue simulation apparatus according to claim 1, wherein the display unit further displays the scene and a CG image of the dialogue partner. Desired scene type, date / time, place, purpose, and conditions for setting the scene including at least one of the formal and informal scenes, and the gender, age, and expertise of the desired conversation partner An input means for inputting conditions for setting the conversation partner including at least one of hobbies, occupations, and relationships with users ;
A scenario storage means for storing a plurality of scenarios for dialogue including dialogue patterns of dialogue between the user and the conversation partner, in which a topic word insertion position is predetermined;
Information collecting means for collecting the topic words via a network;
Information storage means for storing topic words collected by the information collection means;
Display means;
Voice input means for inputting a user's speech;
Audio output means;
Control means;
A dialogue support method in a dialogue simulation apparatus comprising:
A first step in which the input means inputs a condition for setting the scene and a condition for setting the conversation partner;
For each of a plurality of predetermined fields, the information collection unit uses the conditions input for the scene and the conversation partner input by the input unit as keywords, and the topic word is obtained via a network. A second step of collecting;
The control means belongs to a field selected by the user from among the plurality of fields, from among the scenarios stored in the scenario storage unit, in a scene satisfying the condition input in the first step. A third step of selecting a scenario;
A fourth step in which the control means selects a topic word belonging to the field selected by the user from the plurality of fields from the topic words stored in the information storage means;
A fifth step in which the display means displays at least the topic word selected in the fourth step and the user's dialogue template in the scenario selected in the third step;
A sixth step in which the voice input means inputs a user's statement;
The control means obtains the difference between the user's utterance input in the sixth step and the dialogue template displayed by the display means in the fifth step, so that the topic word in the user's utterance A seventh step of extracting
The control means inserts the topic word extracted in the seventh step into a predetermined insertion position of the dialogue model of the dialogue partner in the scenario selected in the third step, An eighth step of generating dialogue for the dialogue partner;
A ninth step in which the voice output means synthesizes and outputs the speech generated in the eighth step;
Dialog support method including   The dialogue simulation apparatus comprises:
  CG of the scene and conversation partner ( Computer graphics ) Component storage means for storing CG components for creating an image;
  Further comprising
  The control means creating a CG image of the scene using the CG parts stored in the parts storage means based on the conditions for setting the scene input in the first step; ,
  The control unit creates a CG image of the conversation partner using the CG component stored in the component storage unit based on the condition for setting the conversation partner input in the first step. Steps,
  Further including
  4. The dialogue support method according to claim 3, wherein in the fifth step, the display means further displays the scene and a CG image of the dialogue partner.

Images