JP2003280687A - Scenario creation supporting device and program for the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a scenario creation supporting device which reduces the man-hours required for scenario creation and easily creates a scenario for realizing interactive processing. <P>SOLUTION: Considering that the description of a scenario created in an interactive system part at least includes a users' input word, a corresponding speech sentence at the side of the system and the transition position of the scenario, the automatic creation of the scenario is performed by using an interactive scenario specification table 21. The interactive scenario specification table 21 expresses the correspondence relation of each information including the input word, a scenario name to be added, a speech sentence and a transition scenario name. Thereby, an input of required information is simplified by expressing the scenario which becomes complicated by a hierarchical structure and a transition description in the structure in order to realize natural interactive processing by the interactive scenario specification table 21 for each unit of a dialogue. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique for creating a scenario used for interactive processing.

[0002]

2. Description of the Related Art Conventionally, there is known a technique of recognizing a user's utterance by voice and controlling an apparatus or the like. In such a case, there is an idea of more naturally realizing a dialogue with a user, creating a flow of the dialogue, and controlling the device by a series of dialogue contents. In order to realize this, it is necessary to change the answer and the next question on the system side in response to the user's utterance.

As a method therefor, it is conceivable to use a scenario in which output information and transition information on the system side in response to a user's utterance are described in advance. By using such a scenario, the flow of the dialogue processing on the system side is realized by the scenario, and if this scenario is prepared, the dialogue can be easily realized in various systems and the user's utterance can be realized. Appropriate output to
You can make the appropriate transitions.

An outline of the scenario is illustrated in FIG. 13 (a). As shown in FIG. 13A, the scenario may be described in a hierarchical structure such as scenario → act → scene. Also, a label for implementing the transition process may be described in the scene. Performance is described in this scenario. As shown in FIG. 13 (b), the performance is as follows: user-side input information such as user-side utterance information, system-side output information corresponding to the user-side input information, system control information, and after the system-side output. The transition position of the scenario is included.

[0005]

In order to realize natural dialogue processing by a scenario, it is necessary to create a plurality of hierarchical structures in the scenario, and the scenario description must be performed in a predetermined description format. I have to.

Therefore, there is a problem that the number of man-hours for creating a scenario becomes large in the system design. Moreover, the scenario can be created only by those who have mastered the description format of the scenario. An object of the present invention is to provide a scenario creation support device for easily creating a scenario, which contributes to a reduction in scenario creation man-hours.

[0007]

Means for Solving the Problems and Effects of the Invention A scenario creation support apparatus of the present invention creates a scenario used in a dialogue system for executing dialogue processing. This scenario realizes the transition in the dialogue processing.

In this scenario, user-side input information input by the user is described, and when the user-side input information is input, the system-side output information that is output from the interactive system and the next The transition position of the scenario to be executed is described. Based on such a premise, in the present invention, the input means uses the user-side input information, the description position of the user-side input information, the system-side output information, and the transition position as a dialog in which the correspondence relationship between each information is expressed. Input in the format of insertion into the scenario specification table. It is conceivable that each information input by the input means is given by the user from a keyboard or the like. Then, the scenario creating means creates a scenario according to the created interactive scenario specification table.

That is, in the present invention, the description of the scenario is
Focusing on at least the user-side input information, the corresponding system-side output information, and the transition position, and by specifying the description position of the user-side input information, create a dialogue scenario specification table that is the basis of scenario creation. To do.
At this time, the dialogue scenario specification table was made to be able to express the correspondence relationship of each information, and the format of inserting information into the dialogue scenario specification table was used to realize easy-to-understand input for the user.

As a result, even if the description format of the scenario is not acquired, the scenario can be easily created, which contributes to the reduction of the scenario creation man-hour. It should be noted that a plurality of user-side input information may be inserted into one correspondence relationship in the dialogue scenario specification table (claim 2). This is for the output information on the system side,
It means that a plurality of user-side input information are associated. By doing so, the input on the user side for obtaining the output on a certain system side is not fixed.

Similarly, it is conceivable that a plurality of system-side output information can be inserted into one correspondence relationship in the dialogue scenario specification table (claim 3). Also, a plurality of transition positions may be inserted (claim 4). In the former configuration, the system can be flexibly operated in response to an input from a certain user, and in the latter configuration, the flow of dialogue can be changed depending on the situation.

In particular, assuming the latter configuration, it is possible to insert transition conditions corresponding to a plurality of transition positions, and the scenario creating means may adopt a configuration for creating a scenario in consideration of the transition conditions. It is considered (Claim 5). By doing so, the flow of the dialogue can be appropriately changed depending on the transition condition.

By the way, the scenario creating means may add a description to a pre-existing scenario so that the scenario can be created. That is, according to the present invention, since the description position of the input information on the user side is designated, it is possible to easily realize the extension of the existing scenario. This allows the scenario to be expanded by adding a description to the existing scenario.

The above-mentioned user-side input information may be user's utterance information (claim 7). In this case, system control by the user's voice becomes possible. The dialogue system may be adapted to judge the utterance information of the user by using the dialogue recognition dictionary. Therefore, the scenario creating means may create the dialog recognition dictionary according to the dialog scenario specification table (claim 8). In this case, not only the scenario but also the interactive recognition dictionary is automatically created, which is convenient for the user.

Further, it is conceivable that the above-mentioned output information on the system side is speech information of the dialogue system (claim 9). In this case, information output from the dialogue system is also performed using voice. The dialogue system may output the utterance information of the dialogue system according to the utterance list in which the utterance information is described. Therefore, it is preferable that the scenario creating means adopts a configuration for creating the utterance list according to the dialogue scenario specification table (claim 10). In this case, not only the scenario but also the utterance list is automatically created, which is convenient for the user.

By the way, depending on the dialogue system, there is also a dialogue system in which information about a spoken voice such as voice quality and intonation can be set. Therefore, also in the present invention, when the information about the uttered voice is set, the scenario creating means may reflect the information to enable the scenario to be created (claim 11). By this, the setting description of the interactive system in the scenario is automatically generated.

The output information on the system side may include operation information on the interactive system side (claim 12). It is considered that the operation information here is control information of the dialogue system or designation information of an application started from the dialogue system. By doing so, for example, in response to a voice input on the user side such as "I want to see the news", an operation of reproducing a news image, that is, device control via a dialogue system based on a scenario can be realized.

Incidentally, the scenario creating means may be capable of creating a plurality of scenarios according to the dialogue scenario specification table (claim 13). In the dialogue scenario group consisting of multiple scenarios, the root scenario is the top layer,
It consists of several sub-scenarios. Therefore, the hierarchical structure between such scenarios is created from one dialogue scenario specification table, like the hierarchical structure inside the scenario. This makes it easy to create a dialogue scenario group having a hierarchical structure.

It has already been described that the dialogue scenario specification table is created in the form of inserting each information. At this time,
The user may directly input information via a keyboard or the like, but the following configuration may be adopted from the viewpoint of simplifying the input.

That is, it is conceivable to adopt a configuration including a graphic editor for inputting graphic information that expresses at least the description position and the transition position of the input information on the user side. In this case, the input means can create the dialog scenario specification table based on the graphic information input by the graphic editor (claim 14). If the description position and transition position of the input information on the user side are expressed using the graphic information, it is easy for the user to understand the hierarchical structure between scenarios and the inside of the scenario, rather than directly inputting information to the dialogue scenario specification table. It is easy to describe the hierarchical structure. At this time, the user side input information and the system side output information may be input as part of the graphic information.

Since the user can easily understand the hierarchical structure of the scenario by using the graphic information in this way, the graphic information can be used to instruct the hierarchical structure of the entire dialogue processing, and the input means can specify the hierarchical structure. May be reflected to create the dialogue scenario specification table (claim 1
5). This technical idea focuses on the fact that the flow of dialogue processing can be comprehensively expressed by using graphic information, and reflects the hierarchical structure of the whole dialogue processing in the dialogue scenario specification table.

Further, the corresponding dialogue scenario specification table may be read and modified by using the figure information input by the figure editor (claim 16). This makes it easy to partially edit the dialogue scenario specification table. In addition to input by graphic information, the input means is
When the slot information describing a series of input items is input, the dialogue scenario specification table may be created (claim 17). The slot information is a description of a series of input items such as the name of the destination, the prefecture in which the destination is located, and the desired arrival time at the destination, which is prepared in the process of setting the destination in the navigation device. From such slot information, "where do you want to go?",
It is possible to determine the utterances on the system side such as "What prefecture is there?" And "By what time do you want to arrive?" Therefore, if it is possible to create the dialogue scenario specification table based on the slot information, it is convenient to be able to semi-automatically create a scenario for realizing a series of dialogue processing regarding input items.

Although the above is the configuration that focuses on the point that a scenario is created based on the dialog scenario specification table, it is also possible to have a configuration that further includes a dialog system. That is, an interactive system may be provided so that interactive processing can be executed based on the scenario created by the scenario creating means (claim 18). By doing so, the interactive processing based on the created scenario can be actually checked, and as a result, the scenario can be added or modified based on the check result.

For example, it is conceivable that a test file in which user-side input information is described is input to the interactive system so that interactive processing can be executed (claim 19). If the interactive process can be executed by using such a test file, it is convenient because the user does not need to perform a voice test.

At this time, it is preferable that the history of the execution of the interactive processing can be recorded (claim 20). This is because it is easy to confirm the check result. In addition, using the graphic information input in the graphic editor as described above,
The dialogue processing based on the corresponding scenario may be executed (claim 21). If partial checking is possible using the graphic information in this way, it is possible to easily check a relatively large-scale scenario.

The invention of the scenario creation support device has been described above, but the function of the scenario creation support device can be realized by a computer. In this sense, the present invention can be realized as an invention of a program that causes a computer to function (claim 22).

In the case of such a program, for example, F
It can be used by recording it on a computer-readable recording medium such as D, MO, DVD, CD-ROM, hard disk, etc., loading it to a computer and starting it as necessary. In addition, ROM and backup RA
The program may be recorded by using M as a computer-readable recording medium, and this ROM or backup RAM may be incorporated into a computer for use.

[0028]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a functional block diagram showing a schematic configuration of the scenario creation support device of the embodiment.

The scenario creation support device is a well-known computer system, and includes a user interface 10 and
An input unit 20, a scenario creation unit 30, a dialogue system unit 40, and a graphic editor 50 are provided. The user interface 10 is a configuration for exchanging information with the user and includes an input / output mechanism. This input / output mechanism enables not only means such as a keyboard and a display but also input / output of information by voice, and has a voice recognition function and a voice synthesis function for this purpose.

In the input section 20, the dialogue scenario specification table 21
Is created by the information input via the user interface 10. Further, the dialogue scenario specification table 21 can be created based on the figure information from the figure editor 50. The scenario creating unit 30 creates a scenario, a dialog recognition dictionary, and an utterance list based on the dialog scenario specification table 21 created by the input unit 20.

As shown in FIG. 2, the dialogue system unit 40 analyzes a scenario and executes a dialogue process.
The scenario interpreter is equipped with an interpreter, and executes a dialogue process based on a scenario (a dialogue scenario group), a dialogue recognition dictionary, and an utterance list.

Next, the dialogue scenario specification table 21 will be described. FIG. 3 shows an example of the dialogue scenario specification table 21. As can be seen from FIG. 3, the dialogue scenario specification table 21 is divided into “input” and “output”, and columns of “number”, “input word” and “scenario name to be added” are provided on the input side. On the output side, columns of “number”, “uttered sentence” and “transition scenario name” are provided.

The input word mentioned here is a phrase described in the above-mentioned dialogue recognition dictionary, and this corresponds to "user side input information". Also, the scenario name to be added indicates the description position for describing the processing by the input word,
This corresponds to the "description position of the input information on the user side". The utterance sentence is described in the utterance list, and is read and output by performing the performance based on the input word. This corresponds to "system side output information". Then, the scenario name of the transition corresponds to the "transition position".

Although the scenario name to be added and the scenario name to be transitioned are described here, an act name and a scene name may be described in order to represent the transition in the scenario.
“Question” shown in FIG. 3 is assumed to be a scene name. All numbers are automatically assigned to input words and utterance sentences, and are information for associating a dialogue recognition dictionary with a scenario and a utterance list with a scenario.

Next, the scenario creation processing executed by the scenario creation support apparatus of this embodiment will be described with reference to the flowchart of FIG. In the first step (hereinafter, step is simply indicated by symbol S) 100, input information is inserted. The input information is the user interface 10
It may be given from the user via. This process corresponds to the insertion of the input word in FIG. For example, inserting the word "eye".

At S110, the description position is inserted.
This process corresponds to the insertion of the scenario name to be added in FIG. The scene name "Q" is associated with the input word "eye".
"Ustion" is inserted. Next S
At 120, output information is inserted. This process corresponds to the insertion of the uttered sentence in FIG. The utterance sentence “There are two eyes.” Is inserted in correspondence with the input word “eye”.

At S130, the transition position is inserted.
This process corresponds to the insertion of the transition scenario name in FIG. The scene name "Q" is associated with the input word "eye".
"Ustion" is inserted. Then, in the next S140, it is determined whether or not the input is completed.
This determination process is performed based on the input from the user. When it is determined that the input of the dialogue scenario specification table 21 is completed (S140: YES), the process proceeds to S150. On the other hand, while the input is not completed (S140: N
O) The processing from S100 is repeated.

By repeating the processing of S100 to S140, the columns of "input word", "scenario to be added", "uttered sentence" and "transition scenario name" in the dialogue scenario specification table 21 shown in FIG. It will be inserted. Here, input word → scenario name to be added →
It is assumed that the utterance sentences are inserted in the order of the transition scenario name, but the input order is not particularly limited. Also,
A plurality of input words, spoken sentences, and transition scenario names may be described in one column. This is because more natural dialogue processing can be realized.

In S150, the creation information is input. For example, as shown in FIG. 5, it is conceivable to prepare selection items such as uttered voice, character, transition condition, graphic input, and configuration in scenario. The user can set the instructed item by instructing the item to be set with a pointing device such as a mouse.

For example, when the uttered voice, the character, and the transition condition are instructed, a screen display as shown in FIG. 6 may be displayed to prompt the user to set each item. Specifically, as shown in FIG. 6, for the uttered voice, the type of voice, voice quality, pitch, intonation, tone,
Prompt for settings such as speed and volume. For the type of voice, male voice or female voice is selected, and other information can be set in five levels from "1" to "5".

Regarding the character,
You can choose from older sister, priest, young lady, grandfather, grandmother, and robot. The transition condition designates a transition condition when there are a plurality of transition positions described in the dialogue scenario specification table 21, and the transition condition such as the number of times, language, route, etc. can be selected. By setting such transition conditions, it is possible to appropriately change the flow of the interactive processing.

When the figure input in FIG. 5 is selected, the figure processing described later is executed and the dialogue scenario specification table 21 is created based on the information from the figure editor 50.
Further, when the configuration in the scenario is selected, detailed item setting in the scenario can be performed.

When a series of information is input to the dialogue scenario specification table 21, the dialogue scenario specification table 21
A number is automatically assigned in the "Number" column of. In S160 of FIG. 4, based on the dialogue scenario specification table 21,
Create a scenario. The scenario created from the dialogue scenario specification table 21 shown in FIG. 3 is an explanatory diagram of FIG. 7.

In all the input words shown in FIG. 3, the scene name "Question" is the description position. Therefore, as shown in FIG. 7, the scene name “Questio
n ”is described by the label definition. In Figure 7,
"EYE", "HEAD", "LANGUAGE",
Label definitions of "YES" and "NO" are made.

Then, the input side number "101,1",
“101,2”, ...
It is described as an argument of the branch instruction "branch-label" (symbol a). Further, in each label, an output command “static-phrase” of the utterance sentence is described. For example, the label "EYE" contains "stat
Although the description "ic-phase (1)" is made (symbol b), "1" of this argument corresponds to the output side number.

Further, in each label, the transition command "g
"oto" is described. For example, in the label “EYE”, “goto-sce” is associated with the transition scene name.
The description is "ne (Question)" (symbol c). This indicates a transition to scene (Question).

Under the label "LANGUAGE", the labels "SPEAK" and "UNDERSTAN" are placed.
"D" is described as a hierarchical structure (symbols d, e, f).
This is because "speakable" and "understand" are described under the input word "word" in the dialogue scenario specification table 21 shown in FIG.

In S170 of FIG. 4, a dialogue recognition dictionary is created. The dialogue recognition dictionary is the dialogue scenario specification table 21.
The input side number, the input word, and the reading of the input word are associated with each other. Dialog scenario specification table 2 shown in FIG.
From 1, a dialogue recognition dictionary as shown in FIG. 8 is created.

In subsequent S180, an utterance list is created. The utterance list corresponds to the numbers on the output side of the dialogue scenario specification table 21 and the utterance sentences. An utterance list as shown in FIG. 9 is created from the dialogue scenario specification table 21 shown in FIG. Then, after the processing of S180 ends, the scenario creation processing ends.

Based on the scenario, the recognition dictionary for dialogue, and the utterance list created in this way, the dialogue system unit 40
Will execute the interactive process. This interactive process is executed by inputting a test file through the user interface 10. The test file is a text file corresponding to the user's utterance.

First, when the scenario shown in FIG. 7 is executed and the scene name "Question" is entered, the branch instruction "de"
f-branch-label "is executed. At this time, the interactive recognition dictionary is searched for a word or phrase in the test file acquired via the user interface 10, and a branch is made to a label in the scenario according to the search result.
For example, if the phrase is “eye”, the search result is the number “10
Since it is output as "1,1", the corresponding label "EY"
Branch to "E". Then, the utterance sentence output command “sta
tic-phrase (1) "is analyzed, the utterance sentence" there are two eyes "corresponding to the number" 1 "is read from the utterance list, and the user interface 10 synthesizes and outputs the voice. After that, the transition command "goto-sc
ene (Question) is executed, and the branch instruction "def-branch-label" is executed again. At this time, the history of the execution of the interactive processing is also recorded.

The scenario is divided into a root scenario and sub-scenarios 1, 2, 3, ..., N. The transition inside the scenario is shown here, but from the root scenario to sub-scenarios 1 to N, or sub-scenario 1
Transitions across scenarios between ~ N are similarly realized. Then, by repeating such interactive processing,
A natural dialogue is realized by the dialogue system unit 40.

Next, based on the flowchart of FIG.
Graphic processing via the graphic editor 50 will be described.
First, in step S200, graphic information is input via the graphic editor 50. It is conceivable that this graphic information indicates, for example, the description position or transition position of the scenario in the scenario hierarchical structure. As shown in FIG. Here, the root scenario and 15 sub-scenarios A, B, C, D, E, F, G, H, I, J, K,
L, M, N and O are shown. At this time, sub-scenario A ~
Sub-scenarios D to I are located under C, and sub-scenarios J to O are further located under it. Then, the scenario connected by the solid line indicates that the upper scenario transits to the lower scenario. For example, the sub-scenario D and the sub-scenario J are connected by a solid line (symbol a), but this indicates that the upper sub-scenario D makes a transition to the lower sub-scenario J. It should be noted that although the description position and transition position of the scenario are shown here as an example, information such as an input word or a spoken sentence may be described in the graphic information.

At S210, it is determined whether the input of the graphic information is completed. This determination process is performed by input from the user. When it is determined that the input of the graphic information is completed (S210: YES), the process proceeds to S220. On the other hand, if the input of the graphic information is not completed (S
210: NO), the process from S200 is repeated.

In S220, the dialogue scenario specification table 21 is created from the graphic information. In subsequent S230, it is determined whether or not the input is completed. When all the information is inserted in the dialogue scenario specification table 21, a positive determination is made here. When it is determined that the input is completed (S230: YES),
The process proceeds to S150 in FIG. In this case, a scenario, a dialogue recognition dictionary, and a speech list are created (S160 to S180 in FIG. 4). On the other hand, when the input is not completed (S230: NO), the process proceeds to S240.

In S240, it is determined whether or not there is a correction instruction. This correction instruction is made by clicking the graphic information via the graphic editor 50. If it is determined that the correction instruction has been given (S240: YES), the dialogue scenario specification table 21 is read in S250, and then the process proceeds to S100 in FIG. On the other hand, when there is no correction instruction (S240: NO), the process of S250 is not executed and the process of FIG.
The process proceeds to S100 inside. For example, in FIG. 11, when the sub-scenario J is clicked (symbol b), the dialogue scenario specification table corresponding to the sub-scenario J is read (S250).

By the way, in the present embodiment, it has already been described that the dialogue system section 40 automatically verifies the dialogue processing based on the test file. In this connection
In this embodiment, when a part of the graphic information is selected via the graphic editor 50, the interactive process corresponding to the selected part can be verified. For example, referring to FIG. 11, when a part of the scenario (symbol c) indicated by the broken line is selected, the interactive processing corresponding to the root scenario and the sub-scenarios B, C, F and G is executed.

Next, the effect of the scenario creation support device of this embodiment will be described. In the present embodiment, focusing on that the description of the scenario executed by the dialogue system unit 40 includes at least the input word on the user side, the utterance sentence on the system side corresponding thereto, and the transition position of the next scenario,
A scenario is automatically generated using the dialogue scenario specification table 21. The dialogue scenario specification table 21 includes an input word, a scenario name to be added, an uttered sentence, and a transition scenario name, and expresses a correspondence relationship of each information. That is, a scenario that becomes complicated by a hierarchical structure and a transition description in the structure for realizing a natural dialogue process is expressed in the dialogue scenario specification table 21 for each dialogue unit, thereby simplifying the input of necessary information. did.

As a result, even if the description format of the scenario is not acquired, the scenario can be easily created, which contributes to the reduction of the scenario creation man-hour. Further, in the present embodiment, as shown in FIG. 3, since the "scenario name to be added" is designated as the description position, it is possible to easily realize the extension of the existing scenario. Furthermore, the dialogue recognition dictionary and the utterance list required in the dialogue system unit 40 are also created from the dialogue scenario specification table 21 (S in FIG. 4).
170, S180). This is convenient because there is no need for the user to create the recognition dictionary for dialogue and the utterance list.

Furthermore, in this embodiment, scenario creation information can also be input (S150 in FIG. 4), and a scenario is created by reflecting this creation information (S160). By this, the setting description of the interactive system in the scenario is automatically generated. Further, in the present embodiment, the graphic editor 50 is provided, and the dialogue scenario specification table 21 can be created or modified based on the graphic information. As a result, the user can easily understand the hierarchical structure, and the partial correction of the dialogue scenario specification table 21 also becomes easy.

Furthermore, in this embodiment, by including the dialogue system section 40, the dialogue processing based on the created scenario can be executed and the check can be performed. At this time, a test file in place of the user's utterance can be input to the dialogue system unit 40 to execute the dialogue processing, which is convenient because the user does not need to bother performing the voice test. Also, since the history of the execution of dialogue processing can be recorded,
Check results can be easily confirmed. Furthermore, since it is possible to perform a partial check using graphic information, it is easy to check a relatively large-scale scenario.

In the present embodiment, the input section 20 corresponds to "input means", the scenario creating section 30 corresponds to "scenario creating section", the dialogue system section 40 corresponds to "dialogue system", and a graphic editor. 50 corresponds to a "graphic editor". As described above, the present invention is not limited to such embodiments and can be implemented in various forms without departing from the gist of the present invention.

(A) In the above embodiment, the utterance sentence is input to the dialogue scenario specification table 21. However, as shown in FIG. 12, operation information on the system side may be inserted. In this way, it is possible to realize an operation of reproducing a news image in response to a user's voice input of “news” or “want to see”, that is, device control via a dialogue system based on a scenario.

(B) In the above embodiment, the creation of the dialogue scenario specification table 21 using graphic information was mentioned, but using graphic information in this way makes it easier for the user to understand the hierarchical structure of the scenario. Therefore, the hierarchical structure of the entire dialogue processing may be designated by the graphic information, and the input unit 20 may create the dialogue scenario specification table 21 by reflecting the designated hierarchical structure.

(C) In addition to the input by the graphic information,
When the slot information in which a series of input items is described is input, it may be possible to create the dialogue scenario specification table 21 by the input unit 20. The slot information is a description of a series of input items such as the name of the destination, the prefecture in which the destination is located, and the desired arrival time at the destination, which is prepared in the process of setting the destination in the navigation device. From such slot information,
"Where do you want to go?", "How many prefectures are there?",
It is possible to determine the utterance on the system side such as "By what time do you want to arrive?" And the input order. If a dialogue scenario specification table can be created from slot information, it is convenient to be able to semi-automatically create a scenario for realizing a series of dialogue processing regarding input items.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing a schematic configuration of a scenario creation support device according to an embodiment.

FIG. 2 is an explanatory diagram showing a schematic configuration of a dialogue system unit.

FIG. 3 is an explanatory diagram illustrating a dialogue scenario specification table.

FIG. 4 is a flowchart showing a scenario creating process.

FIG. 5 is an explanatory diagram illustrating a selection item as creation information.

FIG. 6 is an explanatory diagram exemplifying setting items for selection items.

FIG. 7 is an explanatory diagram showing a description of a created scenario.

FIG. 8 is an explanatory diagram showing the contents of a created recognition dictionary for dialogue.

FIG. 9 is an explanatory diagram showing the contents of the created utterance list.

FIG. 10 is a flowchart showing graphic processing.

FIG. 11 is an explanatory diagram illustrating an example of graphic information created by a graphic editor.

FIG. 12 is an explanatory diagram showing insertion of operation information into a dialogue scenario specification table.

FIG. 13 is an explanatory diagram showing an outline of a scenario and contents of performance.

[Explanation of symbols] 10 ... User interface 20 ... Input section 21… Dialog scenario specification table 30 ... Scenario creation section 40 ... Dialog system section 50 ... Graphic editor

Continued front page    (72) Inventor Fumihiko Murase             1-1, Showa-cho, Kariya city, Aichi stock market             Inside the company DENSO F-term (reference) 5D015 KK02 LL10 LL11                 5D045 AB30

Claims (22)

[Claims] 1. A scenario creation support device for use in an interactive system for executing interactive processing, for creating a scenario for realizing a transition of interactive processing, wherein the scenario has a user-side input input by a user. Assuming that the information is described and the system-side output information that is output from the interactive system when the user-side input information is input and the transition position of the scenario to be executed next are described. User side input information, the description position of the user side input information, the system side output information, and the transition position,
Input means for inputting in the form of insertion into a dialogue scenario specification table expressing the correspondence of each information, and scenario creating means for creating the scenario according to the dialogue scenario specification table created via the input means. A scenario creation support device characterized by being provided. 2. The scenario creation support apparatus according to claim 1, wherein a plurality of user-side input information can be inserted into one correspondence relationship in the dialogue scenario specification table. apparatus. 3. The scenario creation support apparatus according to claim 1, wherein a plurality of system-side output information can be inserted into one correspondence relationship in the dialogue scenario specification table. Creation support device. 4. The scenario creation support device according to claim 1, wherein a plurality of transition positions can be inserted into one correspondence relationship in the dialogue scenario specification table. Scenario creation support device. 5. The scenario creation support device according to claim 4, wherein transition conditions corresponding to the plurality of transition positions can be inserted, and the scenario creation unit considers the transition conditions and sets the scenario A scenario creation support device characterized by creating. 6. The scenario creation support device according to claim 1, wherein the scenario creation unit can create the scenario by making an additional description to a preexisting scenario. Scenario creation support device. 7. The scenario creation support apparatus according to any one of claims 1 to 6, wherein the user-side input information is user utterance information. 8. The scenario creation support apparatus according to claim 7, wherein the dialog system is configured to judge the utterance information of the user by using a dialog recognition dictionary, and the scenario creating unit is configured to execute the dialog. A scenario creation support device which creates the dialog recognition dictionary according to a scenario specification table. 9. The scenario creation support device according to claim 1, wherein the system-side output information is utterance information of the dialogue system. 10. The scenario creation support apparatus according to claim 9, wherein the dialogue system stores utterance information of the dialogue system,
The scenario creation support device is configured to output according to the utterance list in which the utterance information is described, and the scenario creating means creates the utterance list according to the dialogue scenario specification table. 11. The scenario creation support device according to any one of claims 1 to 10, wherein when the information on the uttered voice is set, the scenario creation means can create the scenario by reflecting the information. A scenario creation support device characterized by: 12. The scenario creation support apparatus according to claim 1, wherein the system-side output information includes operation information on the interactive system side. apparatus. 13. The scenario creation support apparatus according to claim 1, wherein the scenario creation means is capable of creating a plurality of scenarios in accordance with the dialogue scenario specification table. Support device. 14. The scenario creation support apparatus according to claim 1, further comprising: a graphic editor for inputting graphic information that expresses at least the description position of the user-side input information and the transition position, The scenario creation support device, wherein the input means can create the dialogue scenario specification table based on the graphic information input by the graphic editor. 15. The scenario creation support device according to claim 14, wherein the graphic information is used to instruct a hierarchical structure of the entire dialogue process, and the input means reflects the instructed hierarchical structure. A scenario creation support device for creating the dialog scenario specification table. 16. The scenario creation support apparatus according to claim 14 or 15, wherein the corresponding dialogue scenario specification table can be read and modified by using the graphic information input by the graphic editor. Scenario creation support device. 17. The scenario creation support apparatus according to any one of claims 1 to 16, wherein when the slot information in which a series of input items is described is input, the input means is based on the slot information. A scenario creation support device capable of creating a dialog scenario specification table. 18. The scenario creation support apparatus according to claim 1, further comprising the dialog system, wherein the dialog processing can be executed based on the scenario created by the scenario creating means. Characteristic scenario creation support device. 19. The scenario creation support apparatus according to claim 18, wherein when a test file in which utterance information of the user is described based on the input information on the user side is input to the interactive system, interactive processing is performed based on the test file. A scenario creation support device characterized by executing. 20. The scenario creation support apparatus according to claim 18 or 19, wherein a history of execution of the dialog processing can be recorded. 21. The scenario creation support apparatus according to claim 18, further comprising: a graphic editor for inputting graphic information that expresses at least the description position of the user-side input information and the transition position. On the premise that the scenario creation support device is capable of executing interactive processing based on a corresponding scenario by using the graphic information designated by the graphic editor. 22. A program for causing a computer to function as each unit of the scenario creation support apparatus according to any one of claims 1 to 21.