JP6027476B2 - Dialog program, server, and method for inserting dynamic dialog node in dialog scenario - Google Patents

Description

  The present invention relates to a technology of a dialogue system that interacts with a user using a dialogue scenario.

  A dialogue system that realizes natural dialogue with human beings is becoming more popular, especially on smartphones and tablet terminals. Among them, there is a technique for acquiring necessary information from a user by alternately proceeding a dialogue process with the user according to a dialogue scenario (see, for example, Patent Document 1). According to this technology, a dialogue process can be transitioned between dialogue systems having a common dialogue scenario, and the dialogue process currently in progress can be continued in the dialogue system after the transition. It is executed according to a fixed dialogue scenario set in advance.

  There is also a technique that can appropriately respond to any input other than the prediction range in the dialogue scenario (see, for example, Patent Document 2). According to this technology, a natural dialogue can be promoted by preparing a large number of dialogue nodes and making a transition to a dialogue node having the maximum transition probability with respect to the current dialogue node. The transition probability is updated in consideration of past discrimination results. As a result, the system administrator has less time to create a fixed conversation scenario.

JP 2004-212802 A JP 2009-198614 A

"Calculation of similarity of documents by vector space method", [online], [Search February 24, 2013], Internet <URL: http://www.gifu-nct.ac.jp/elec/deguchi/ sotsuron / hayashi / node20.html>

  However, according to the technique described in Patent Document 1, since it is a fixed dialogue scenario, there is a problem that the user is easily bored when used for a long time. To deal with such problems, it is necessary for the operator of the dialogue system to increase the number of dialogue scenario branches and dialogue variations, which requires manpower. Further, according to the technique described in Patent Document 2, there is also a problem that partial correction of the dialogue scenario is difficult because all of the transition of the dialogue scenario depends on the probability model.

  On the other hand, the inventors also considered inserting a dynamic dialog node having the maximum transition probability based on Patent Document 2 during the progress of the fixed dialog node based on Patent Document 1. As a dynamic dialog node, in order to acquire user information, for example, a conversation for collecting profile information such as a user's hobbies and preferences and a conversation for providing information that the user is likely to be interested in are inserted. As a result, the fixed dialogue scenario is partially and automatically corrected, and the effort for increasing the variation of the dialogue is reduced.

  However, even if the scenario can be advanced naturally from a fixed first dialog node to a dynamic dialog node having a high transition probability, from the dynamic dialog node, When returning to the dialog node, the natural nature of the dialog suddenly drops. That is, there is no or little conversational relevance between the dynamic dialog node and the fixed second dialog node as the return destination.

  Therefore, an object of the present invention is to provide a dialog program, a server, and a method capable of inserting a dynamic dialog node between fixed dialog nodes in a dialog as natural as possible.

According to the present invention, there is provided an interactive program for causing a computer to function to advance a scenario interactively with a user,
A dialogue scenario accumulating means for accumulating a plurality of fixed dialogue nodes associated with a sentence to be clearly shown to a user and connecting them as a scenario from the first fixed dialogue node to the second fixed dialogue node;
Dynamic dialog node storage means for storing a plurality of dynamic dialog nodes that can be inserted between dialog nodes;
From the dynamic dialogue node storage means, the transition probability P (A, Si) of the first fixed dialogue node from the sentence and the return probability P (Si, B) of the return to the second fixed dialogue node A dynamic dialog node selecting means for selecting a dynamic dialog node having a maximum combination under a predetermined condition;
The selected dynamic dialog node is inserted from the first fixed dialog node, and the computer is caused to function as a dialog scenario advancing means for progressing the scenario to the second fixed dialog node.

According to another embodiment of the dialogue program of the present invention,
The dynamic dialog node selection means is:
Dynamic dialog node search means for searching a predetermined number of dynamic dialog nodes having a high transition probability P (A, Si) from the sentence of the first fixed dialog node from the dynamic dialog node storage means;
Return transition probability calculating means for calculating a return transition probability P (Si, B) from the sentence of each searched dynamic dialog node to the second fixed dialog node;
Dynamic dialog node determination in which a combination of a transition probability of going from the first fixed dialog node and a return transition probability from the second fixed dialog node is maximized under a predetermined condition It is also preferable to make the computer function to include the means.

According to another embodiment of the dialogue program of the present invention,
It is also preferable to cause the computer to function so that the transition probability in the dynamic dialog node search means and the return transition probability calculation means is based on the cosine distance between the sentence of the fixed dialog node and the sentence of the dynamic dialog node. .

According to another embodiment of the dialogue program of the present invention,
Using a key word dictionary that stores key words (nouns) that frequently appear in the dialogue corpus, the computer functions so that the cosine distance is calculated according to the key words extracted from each sentence and their categories. It is also preferable that

According to another embodiment of the dialogue program of the present invention,
Each first fixed dialogue node is given a question suitability level indicating whether or not to insert a question based on a dynamic dialogue node after a sentence included in the node,
An insertion determining means for determining that a dynamic dialog node is inserted next to the first fixed dialog node only when the question fitness of the first fixed dialog node is equal to or greater than a predetermined threshold;
It is also preferable to cause the computer to function so that the dynamic dialog node search means functions only when it is determined to be true by the insertion determination means.

According to another embodiment of the dialogue program of the present invention,
A positive / negative determining means for determining whether the content of the user's answer to the sentence of the first fixed dialog node is positive or negative;
The insertion determination means multiplies the question fitness by a coefficient assigned to each positive / negative, and inserts a dynamic dialog node next to the first fixed dialog node only when the multiplication value is equal to or greater than a predetermined threshold value. It is also preferable to make the computer function so as to make a determination.

According to the present invention, there is provided a dialogue server that interactively advances a scenario with a terminal based on a user operation,
A dialogue scenario accumulating means for accumulating a plurality of fixed dialogue nodes associated with a sentence to be transmitted to a terminal, and associating the scenario from the first fixed dialogue node to the second fixed dialogue node;
Dynamic dialog node storage means for storing a plurality of dynamic dialog nodes that can be inserted between dialog nodes;
From the dynamic dialogue node storage means, the transition probability P (A, Si) of the first fixed dialogue node from the sentence and the return probability P (Si, B) of the return to the second fixed dialogue node A dynamic dialog node selecting means for selecting a dynamic dialog node having a maximum combination under a predetermined condition;
A dialog scenario advancing means for inserting the selected dynamic dialog node from the first fixed dialog node and progressing the scenario to the second fixed dialog node is provided.

According to the present invention, there is provided an interactive method for interactively progressing a scenario with a user using an apparatus,
A dialogue scenario storage unit that accumulates a plurality of fixed dialogue nodes associated with sentences to be clearly shown to a user, and links them as a scenario from the first fixed dialogue node to the second fixed dialogue node;
A dynamic dialog node storage unit that stores a plurality of dynamic dialog nodes that can be inserted between dialog nodes;
From the dynamic dialogue node storage unit, the transition probability P (A, Si) of the first fixed dialogue node from the sentence and the return probability P (Si, B) of the return to the second fixed dialogue node A first step of selecting a dynamic interaction node whose combination is maximal under predetermined conditions;
A second step of inserting the selected dynamic dialog node from the first fixed dialog node and proceeding with the scenario to the second fixed dialog node.

  According to the dialogue program, server, and method of the present invention, by considering the transition probability when returning from the dynamic dialogue node to the fixed dialogue node, the dialogue between the stationary dialogue nodes is as natural as possible. Dynamic dialog nodes can be inserted.

It is a block diagram of the dialogue system in this invention. It is explanatory drawing showing the some fixed dialogue node memorize | stored in the dialogue scenario storage part. It is explanatory drawing which searches the dynamic dialog node which should be inserted between fixed dialog nodes. It is explanatory drawing which calculated the return transition probability about each searched dynamic dialog node. It is explanatory drawing which calculated the transition probability with respect to a dynamic dialog node brute force. FIG. 3 is an explanatory diagram illustrating a final dialog scenario in which a dynamic dialog node is inserted between fixed dialog nodes in FIG. 2.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a configuration diagram of a dialogue system according to the present invention.

  According to FIG. 1, the dialogue server module program 1 and the terminal module program 2 are configured. For example, the dialogue server module program 1 is a program that advances a scenario interactively with a user, and is executed by a server, for example. On the other hand, the terminal module program 2 is executed by a smartphone or a tablet terminal, and communicates with a server via a network. The dialogue server module program 1 interacts with the user through the terminal 2. Note that both the module programs 1 and 2 may be executed integrally by a single computer of a single device. According to the present invention, the target device on which the server module program and the terminal module program are to be executed is not specified.

  The terminal module program 2 has an input device and an output device for the user. When the input device is a microphone, the input voice processing unit converts the voice signal acquired by the microphone into an answer sentence, and transmits the answer sentence to the dialogue server module program 1. When the output device is a speaker, the answer sentence received from the dialog server module program 1 is converted into an audio signal, and the audio signal is output from the speaker.

  According to FIG. 1, the dialog server module program 1 includes a dialog scenario storage unit 10, a dynamic dialog node storage unit 11, a dynamic dialog node selection unit 12, a dialog scenario progression unit 13, and a question sentence transmission unit 14. An answer sentence receiving unit 15, a positive / negative determining unit 16, and an insertion determining unit 17. The question message transmitting unit 14 transmits a question message to the user based on the dialogue scenario advancing unit 13 to the terminal module program 2. In addition, the answer message receiving unit 15 receives an answer message from the user from the terminal module program 2.

[Dialogue scenario storage unit 10]
The dialogue scenario storage unit 10 accumulates a plurality of fixed dialogue nodes associated with sentences to be clearly shown to the user, and links them as a scenario from the first fixed dialogue node to the second fixed dialogue node. The “dialog scenario” is a dialog node composed of questions and answers configured in a tree shape.

  FIG. 2 is an explanatory diagram showing a plurality of fixed dialog nodes stored in the dialog scenario storage unit.

According to FIG. 2, the scenario progresses as follows by the connection of each fixed dialog node.
(System) “Please schedule”
(User) "Play golf in Karuizawa at 8:00 on Saturday"
(System) "Do you want to set an alarm 2 hours before the start of play?"
(User) "Yes"
For example, if the user does not respond to the “please schedule” question in the system, the scenario ends. Such a dialog scenario is stored in advance in the dialog scenario storage unit 10 by connecting fixed dialog nodes.

[Dynamic dialog node storage unit 11]
The dynamic dialog node storage unit 11 stores a plurality of dynamic dialog nodes that can be inserted between dialog nodes. The dynamic dialog node has the same structure as that of the fixed dialog node, but the dialog nodes are not linked in advance.

  Note that each of the fixed dialog node and the dynamic dialog node constitutes a short dialog scenario by a plurality of answers to the question to the user. Accordingly, among one node, questions and answers are configured as a dialogue scenario in a tree shape.

  Here, as an option, a positive / negative determination unit 16 and an insertion determination unit 17 may be further provided. The positive / negative determining unit 16 and the insertion determining unit 17 determine whether or not a dynamic dialog node may be inserted between fixed dialog nodes. That is, there is a case where a dynamic dialog node should not be inserted depending on the reaction of the user's dialog. For example, if the user reacts negatively, it is not desirable to ask the user further questions by inserting a dynamic interaction node. Hereinafter, two determination methods will be described.

<Only insertion determination unit 17>
For each first fixed dialog node, a question suitability level indicating whether or not to insert a question based on a dynamic dialog node is given after a sentence included in the node. For example, the question suitability is given as follows.
(Fixed conversation node) (Question applicability)
"Please schedule": 1
“Is it bad?”: 0.5
“Do you like golf?”: 2
The question application level becomes a higher value when it is appropriate to ask the user further questions after the conversation of the fixed dialog node. For example, when a fixed dialogue node such as “Do you like golf” is executed, the user has a relatively large margin and the question application level is a high value. On the other hand, the lower the value, the more appropriate it is not to ask further questions after the dialog of the fixed dialog node. For example, when a fixed dialogue node such as “Is it bad?” Is executed, the user has relatively little room and the question application level is a low value.

  Here, the insertion determination unit 17 determines to insert the dynamic dialog node next to the first fixed dialog node only when the question suitability of the first fixed dialog node is equal to or greater than a predetermined threshold. The predetermined threshold may be “0.8”, for example. In that case, no dynamic dialog node is inserted after the question “Is it bad”?

  Finally, the dynamic dialog node selection unit 12 is executed only when the insertion determination unit 17 determines true.

<Positive negative determination unit 16 + insertion determination unit 17>
The positive / negative determination unit 16 determines whether the content of the user's answer to the sentence of the first fixed dialog node is positive or negative. The positive / negative determination unit 16 has an evaluation polarity dictionary in which positive (positive) / negative (negative) semantic attributes are associated with evaluation words. For example:
(Evaluation word) (Semantic attribute)
Bullying Negative Marriage Positive Poor Negative Negative Pleasure Positive In addition, the following coefficients are associated with semantic attributes.
Positive = 2
Negative (negative) = 0.5

Next, according to the insertion determination unit 17, as described above, for each first fixed dialog node, the suitability of inserting a question based on the dynamic dialog node after the sentence included in the node is expressed. Question suitability is given. Then, the insertion determination unit 17 multiplies the question suitability by a coefficient assigned to each positive / negative.
(Dialogue example 1)
"Please enter a schedule": Question application level = 1
“We have a wedding on Sunday”: Positive = 2
Question applicability = 1 × 2 = 2
(Dialogue example 2)
"Isn't it bad?": Question application level = 0.5
"I was bullied": Negative = 0.5
Question applicability = 0.5 × 0.5 = 0.25
(Dialogue example 3)
"Do you like golf?"
“I cannot do it because it is poor”: Negative = 0.5
Question applicability = 2 × 0.5 = 1
Here, the insertion determination unit 17 determines to insert the dynamic dialog node next to the first fixed dialog node only when the question suitability of the first fixed dialog node is equal to or greater than a predetermined threshold. The predetermined threshold may be “0.8”, for example. In that case, no dynamic dialog node is inserted after the “bullyed” answer. On the other hand, a dynamic dialog node is inserted after “There is a wedding on Sunday” and “I cannot do it because it is poor”.

  Finally, the dynamic dialog node selection unit 12 is executed only when the insertion determination unit 17 determines true.

[Dynamic dialog node selection unit 12]
The dynamic dialog node selection unit 12 includes a dynamic dialog node search unit 121, a return transition probability calculation unit 122, and a dynamic dialog node determination unit 123.

<Dynamic dialog node search unit 121>
The dynamic dialog node search unit 121 searches the dynamic dialog node storage unit 11 for a predetermined number of dynamic dialog nodes having a high transition probability P (A, Si) from the sentence of the first fixed dialog node. .

  FIG. 3 is an explanatory diagram for searching for a dynamic dialog node to be inserted between fixed dialog nodes.

According to FIG. 3, it is assumed that seven dynamic dialog nodes are stored in the dynamic dialog node storage unit 11. [] Represents a category belonging to a noun.
(S1) Is golf [sports] a hobby?
(S2) Do you work at Karuizawa [Place]?
(S3) Do you often play sports in Karuizawa [place]?
(S4) Is *** [person name] a friend?
(S5) Do you set alarm [time] every day when you get on [time] train [things]?
(S6) Is something wrong?
(S7) Do you always play golf [sports] on the wake-up [alarm] on holiday [time]?

After each sentence is classified into morphemes, an important word is detected using an important word dictionary in which important words (nouns) having a high appearance frequency in the dialogue corpus are accumulated. Of course, all morpheme nouns may be used as elements for calculating the cosine distance without being detected as important words. The cosine distance is calculated according to an important word extracted from each sentence and its noun type. For example, it is calculated by the following conceptual formula.
Feature vector of key words of fixed dialogue node sentence: D
Feature vector of key words of dynamic dialogue node sentence: E
Similarity between two sentences: sim (D, E)
sim (D, E) = cos θ = (D · E) / (| D || E |)
For the cosine distance, the similarity weight is set to 1 when the words are the same, and the similarity weight is set to 1 when the words are the same category. Of course, the similarity weight of the same word category can be set to a value between 0 and 1.

  Here, the dynamic dialog node search unit 121 calculates a cosine distance between the answer text of the fixed dialog node “go golf in Karuizawa at 8:00 on Saturday” and the question text of each dynamic dialog node. Since holidays [time] and Saturday [time] are in the same category, the similarity weight is 1.

For example, the following important words are detected for “please schedule” and “go golfing at Karuizawa at 8:00 on Saturday”.
“Saturday”: [Time]
“Karuizawa”: [Place]
"Golf": [Sports]
"Schedule": [Schedule]

Here, the cosine distance of each dynamic dialogue node with respect to “please schedule” and “play golf in Karuizawa at 8:00 on Saturday” is calculated.
(S1) Is golf [sports] a hobby?
= "Golf": Cosine distance = 0.50
(S2) Do you work at Karuizawa [Place]?
= "Karuizawa": Cosine distance = 0.50
(S3) Do you often play sports in Karuizawa?
= "Karuizawa""Sports": Cosine distance = 0.71
(S4) Is *** [person name] a friend?
= ***: Cosine distance = 0.0
(S5) Do you set alarm [time] every day when you get on [time] train [vehicle]?
= "Everyday""Train""Alarm": Cosine distance = 0.29
(S6) Is something wrong?
= Cosine distance = 0.0
(S7) Do you always play golf [sports] on the wake-up [alarm] on holiday [time]?
= "Holiday""Wake up early""Golf": Cosine distance = 0.58

  According to FIG. 3, for example, it is assumed that the threshold value “0.5” of the cosine distance to be searched is set. In this case, S1, S2, S3, and S7 are searched as dynamic dialog node candidates. As a method for calculating a cosine distance between sentences, there is a method for calculating a similarity using a vector space method (see, for example, Non-Patent Document 1).

<Return Transition Probability Calculation Unit 122>
The return transition probability calculation unit 122 calculates a return transition probability P (Si, B) from the retrieved sentence of each dynamic dialog node to the second fixed dialog node. Similarly to the dynamic dialog node search unit 121, the return transition probability calculation unit 122 is based on the cosine distance between the sentence of the fixed dialog node and the sentence of the dynamic dialog node.

  FIG. 4 is an explanatory diagram in which a return transition probability is calculated for each searched dynamic dialog node.

Here, the return transition probability calculation unit 122 includes a question sentence of each dynamic dialogue node and a question sentence of the return (second) fixed dialogue node “Do you want to set an alarm 2 hours before the start of play?”. Calculate the cosine distance between them. For example, it is calculated as follows.
“Do you want to set an alarm 2 hours before the start of play?” = “Play” “Time” “Alarm”
(S1) Is golf [sports] a hobby?
= "Golf": Cosine distance = 0.58
(S2) Do you work at Karuizawa [Place]?
= "Karuizawa": Cosine distance = 0.0
(S3) Do you often play sports in Karuizawa [place]?
= "Karuizawa""Sports": Cosine distance = 0.41
(S7) Do you always play golf [sports] on the wake-up [alarm] on holiday [time]?
= "Holiday""Wake up early""Golf": Cosine distance = 1.0

  Here, “golf” and “play” belong to the same category of [sports]. Further, “2 hours” and “holiday” belong to the same category of [time]. “Wake up early” and “alarm” belong to the same category of “alarm”. The cosine distance between words in the same category is shortened. As a result, the cosine distance between the sentences “Do you always play golf on holidays?” And “Do you want to set an alarm 2 hours before the start of play?” Is the shortest (highest value). )

<Dynamic dialog node determination unit 123>
The dynamic dialog node determination unit 123 combines the transition probability P (A, Si) going from the first fixed dialog node and the return transition probability P (Si, B) from the second fixed dialog node. Selects the dynamic interaction node that is maximal under the given conditions. For example, with a predetermined condition as a multiplication value, the dynamic interaction node that selects the maximum value of the multiplication value of the going transition probability P (A, Si) and the returning transition probability P (Si, B) is selected. Also good.
P (A, Si) x P (Si, B)
(S1) Is golf a hobby? = 0.50 x 0.58 = 0.29
(S2) Do you work at Karuizawa? = 0.50 × 0 = 0
(S3) Do you often play sports in Karuizawa? = 0.71 x 0.41 = 0.29
★ (S7) Do you always get up early and play golf on holidays? = 0.58 x 1.00 = 0.58
Finally, it is determined as the dynamic dialog node of S7.

  Alternatively, the value may be a product of the return transition probability P (Si, B) by weighting the outgoing transition probability P (A, Si). Further, it may be an added value of the outgoing transition probability P (A, Si) and the return transition probability P (Si, B). Here, “maximum under a predetermined condition” can employ various calculation methods.

<Higher-level conceptual calculation method of the dynamic dialog node selection unit 12>
The above-described dynamic dialog node selection unit 12 has been described as executing three steps. However, from the beginning, the dynamic dialog node is determined by combining the outgoing transition probability and the return transition probability from the beginning. It may be a choice. That is, the dynamic dialogue node selection unit 12 sends the transition probability P (A, Si) from the sentence of the first fixed dialogue node to the second fixed dialogue node from the dynamic dialogue node storage unit 11. A dynamic dialog node whose combination with the return transition probability P (Si, B) is maximized under a predetermined condition is selected.

  FIG. 5 is an explanatory diagram in which the transition probabilities for the dynamic dialogue nodes are calculated brute force.

  According to FIG. 5, the dynamic dialogue node selection unit 12 calculates the cosine distance between the answer sentence of the fixed dialogue node “Golf golf in Karuizawa at 8:00 on Saturday” and the question sentences of all the dynamic dialogue nodes. To do. Also, the dynamic dialogue node selection unit 12 asks the questions of all the dynamic dialogue nodes and the question of the return (second) fixed dialogue node “Do you want to set an alarm 2 hours before the start of play?”. The cosine distance between is calculated. Then, for each dynamic dialogue node, for example, a multiplication value (addition value or predetermined condition) of two cosine distances is calculated. Then, for example, the dynamic dialog node having the maximum multiplication value is selected.

[Dialogue scenario progression unit 13]
The dialogue scenario progression unit 13 inserts the selected dynamic dialogue node from the first fixed dialogue node, and advances the scenario to the second fixed dialogue node.

  FIG. 6 is an explanatory diagram showing a final dialog scenario in which a dynamic dialog node is inserted between the fixed dialog nodes in FIG.

According to FIG. 6, the dialogue scenario proceeds as follows.
(System) “Please schedule” (fixed dialog node)
(User) "Play golf in Karuizawa at 8:00 on Saturday"
(System) “Do you always get up early on the weekend to play golf?” (Dynamic Dialogue Node)
(User) "It's always a weekday"
(System) “Do you want to set an alarm 2 hours before the start of play?” (Fixed dialogue node)
(User) "Yes"
Here, the dialogue system can collect information (profile information such as user's hobbies and preferences) that the user is “going golf on weekdays” by the dynamic dialogue node.

  As described above in detail, according to the dialogue program, server, and method of the present invention, the transition probability when returning from the dynamic dialogue node to the fixed dialogue node is taken into consideration in the natural dialogue as much as possible. , Dynamic dialog nodes can be inserted between fixed dialog nodes.

  The present invention is suitable for a voice dialogue system in which a “character agent” is displayed on a display of a terminal based on a user operation, and a dialogue between the user and the agent advances by voice. Of course, the present invention is not limited to a voice dialogue, but may be a text dialogue.

  Various changes, modifications, and omissions of the above-described various embodiments of the present invention can be easily made by those skilled in the art. The above description is merely an example, and is not intended to be restrictive. The invention is limited only as defined in the following claims and the equivalents thereto.

DESCRIPTION OF SYMBOLS 1 Dialog server module program 10 Dialog scenario storage part 11 Dynamic dialog node storage part 12 Dynamic dialog node selection part 121 Dynamic dialog node search part 122 Return transition probability calculation part 123 Dynamic dialog node determination part 13 Dialog scenario progress part 14 Question sentence sending part 15 Answer sentence receiving part 16 Positive / negative judgment part 17 Insertion judgment part 2 Terminal module program

Claims (8)

An interactive program that causes a computer to function to proceed with a scenario interactively with a user,
A dialogue scenario accumulating means for accumulating a plurality of fixed dialogue nodes associated with a sentence to be clearly shown to a user and connecting them as a scenario from the first fixed dialogue node to the second fixed dialogue node;
Dynamic dialog node storage means for storing a plurality of dynamic dialog nodes that can be inserted between the dialog nodes;
From the dynamic dialog node storage means, the transition probability P (A, Si) of the return from the sentence of the first fixed dialog node, and the transition probability P (Si, B) of the return to the second fixed dialog node, A dynamic dialog node selection means for selecting a dynamic dialog node that has a maximum combination under a predetermined condition;
An interactive program that inserts the selected dynamic interactive node from a first fixed interactive node and causes the computer to function as an interactive scenario proceeding unit that advances the scenario to the second fixed interactive node. The dynamic dialog node selection means includes:
Dynamic dialog node search means for searching a predetermined number of dynamic dialog nodes having a high transition probability P (A, Si) from the sentence of the first fixed dialog node from the dynamic dialog node storage means;
Return transition probability calculating means for calculating a return transition probability P (Si, B) from the sentence of each searched dynamic dialog node to the second fixed dialog node;
Dynamic dialog node determination in which a combination of a transition probability of going from the first fixed dialog node and a return transition probability from the second fixed dialog node is maximized under a predetermined condition The interactive program according to claim 1, wherein the computer functions so as to include a means.   Causing the computer to function such that transition probabilities in the dynamic dialog node search means and the return transition probability calculation means are based on a cosine distance between a sentence of a fixed dialog node and a sentence of the dynamic dialog node. The dialogue program according to claim 2, wherein:   Using a key word dictionary storing key words (nouns) having a high appearance frequency in the dialogue corpus, the cosine distance is calculated according to the key words extracted from each sentence and their categories. 4. The interactive program according to claim 3, wherein the interactive program is made to function. Each first fixed dialogue node is given a question suitability level indicating whether or not to insert a question based on a dynamic dialogue node after a sentence included in the node,
An insertion determining unit that determines to insert a dynamic dialog node next to the first fixed dialog node only when the question suitability of the first fixed dialog node is equal to or greater than a predetermined threshold;
5. The interactive program according to claim 2, wherein the dynamic interactive node search unit causes a computer to function only when it is determined to be true by the insertion determining unit. 6. . A positive / negative determining means for determining whether the content of the user's answer to the sentence of the first fixed dialog node is positive or negative;
The insertion determination means multiplies the question fitness by a coefficient assigned to each positive / negative, and only when the multiplied value is equal to or greater than a predetermined threshold value, the dynamic dialog node follows the first fixed dialog node. The interactive program according to claim 5, wherein the computer is caused to function so as to determine that the character is inserted.
A dialogue server that interactively advances a scenario with a terminal based on a user operation,
A dialogue scenario accumulating means for accumulating a plurality of fixed dialogue nodes associated with a sentence to be transmitted to the terminal, and associating them as a scenario from the first fixed dialogue node to the second fixed dialogue node;
Dynamic dialog node storage means for storing a plurality of dynamic dialog nodes that can be inserted between the dialog nodes;
From the dynamic dialog node storage means, the transition probability P (A, Si) of the return from the sentence of the first fixed dialog node, and the transition probability P (Si, B) of the return to the second fixed dialog node, A dynamic dialog node selection means for selecting a dynamic dialog node that has a maximum combination under a predetermined condition;
A dialog server, comprising: a dialog scenario advancing means for inserting the selected dynamic dialog node from a first fixed dialog node and progressing a scenario to a second fixed dialog node. A dialogue method in which a scenario is progressed interactively with a user using a device,
A dialogue scenario storage unit that accumulates a plurality of fixed dialogue nodes associated with sentences to be clearly shown to a user, and links them as a scenario from the first fixed dialogue node to the second fixed dialogue node;
A dynamic dialog node storage unit that stores a plurality of dynamic dialog nodes that can be inserted between the dialog nodes;
From the dynamic dialog node accumulating unit, the transition probability P (A, Si) of the return from the sentence of the first fixed dialog node, and the transition probability P (Si, B) of the return to the second fixed dialog node, A first step of selecting a dynamic interaction node whose combination is maximal under predetermined conditions;
A dialog method comprising: a second step of inserting the selected dynamic dialog node from a first fixed dialog node and advancing a scenario to a second fixed dialog node.

Images