JP2009134008A - Emotional answer generation device and emotional answer generation program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an emotional answer generation device and an emotional answer generation program, for generating an appropriate emotional answer, depending on whether user's emotion is negative. <P>SOLUTION: The emotional answer generation device includes an input part 11 for inputting a user's speech and the user's emotion corresponding to the speech; an emotional state estimation part 12, which estimates whether the state of the user's emotion is negative, based on the user's emotion input through the input part 11; an emotional answer generating part 13 which does not generate emotional answer, when the state of the user's emotion estimated by the estimation part 12 is negative, and generates an emotional answer conforming with the user's emotion input through the input part 11, when the state is not negative; and an output part 14 which outputs the emotional answer generated by the emotional answer generating part 13. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an emotion response generation device and an emotion response generation program that generate an appropriate emotion response corresponding to a user's emotion in consideration of the user's emotion state.

In listening, etc., an invention for generating a response to a user utterance based on user emotion has been proposed. As an apparatus that generates such a response, a dialogue process in which the system generates a response that sympathizes with the user's emotion by generating a positive response if the estimated user emotion is positive and a negative response if the estimated user emotion is negative An apparatus (see, for example, Patent Document 1) is known. In the dialog processing device of Patent Document 1, even if the user's emotion is negative, the system responds with negative emotion or responds to deeply digging the emotion content (see FIG. 6 of Patent Document 1 “So then? Etc.).
JP 2006-178063 A

  However, in the dialogue processing device of Patent Literature 1, when the user has a negative emotion, an appropriate emotion response corresponding to the user emotion cannot be generated. Therefore, there is a problem that when the user feeling is negative, it cannot be kept quiet.

  The present invention has been made to solve the above-mentioned problems, and an object thereof is to provide an emotion response device and an emotion response generation program that generate an appropriate emotion response depending on whether or not the user emotion is negative. And

  In order to achieve the above object, an emotion response generating apparatus according to claim 1 is provided with an input unit for inputting a user's utterance and a user's emotion corresponding to the utterance, and a user's emotion input by the input unit. Based on emotion state estimation means for estimating whether or not the emotional state of the user is negative, and negative when the emotional state of the user estimated by the emotional state estimation means is negative and negative In other cases, an emotion response generation unit that generates an emotion response synchronized with the user's emotion input by the input unit, and an output unit that outputs the emotion response generated by the emotion response generation unit are provided. .

  According to the first aspect of the present invention, when the user's emotion is negative, it is possible to keep the user quiet without responding.

  The emotion response generation apparatus according to claim 2, wherein an input unit that inputs a user's utterance and a user's emotion corresponding to the utterance, and a state of the user's emotion based on the user's emotion input by the input unit An emotional state estimating means for estimating whether or not is negative, and when the emotional state of the user estimated by the emotional state estimating means is negative, an emotional response is not generated; otherwise, the input means Emotion response generation means for generating an emotion response that digs deeply into the input user's emotion, and output means for outputting the emotion response generated by the emotion response generation means.

  According to the second aspect of the present invention, when the user's emotion is negative, it is possible to leave the user gently without responding to the emotion, and when the user's emotion is not negative, it is possible to make an emotion response that digs deeply into the user's emotion.

  The emotion response generation device according to claim 3 is the emotion response generation device according to claim 2, wherein the emotion response generation device is configured such that when the emotional state of the user estimated by the emotional state estimation means is other than negative, An emotional response is generated by querying any one of a reason, a time, and a place where the emotion of the user input by the input means is generated, or an emotional response by a fixed expression that further extracts information about the emotional state.

  According to the third aspect of the present invention, when the emotional state of the user is other than negative, the emotional response by a fixed expression that further extracts information on the emotional state, or the reason why the user's emotion has occurred, By generating a response that asks any of the places, it is possible to dig deeply into the user's emotions.

  According to a fourth aspect of the present invention, there is provided the emotion response generating apparatus according to the first aspect, wherein when the emotional state of the user estimated by the emotional state estimating means is negative, a conflict is generated.

  According to the fourth aspect of the present invention, when the user's emotion is negative, the user's emotion can be replied and the user's emotion can be sympathized.

  The emotional response generation device according to claim 5 generates an affinity when the emotional state of the user estimated by the emotional state estimation means is negative in the emotional response generation device according to claim 2 or claim 3. .

  According to the fifth aspect of the present invention, when the user's emotion is negative, it is possible to respond to the user's emotion by responding to each other, and when the user's emotion is not negative, an emotional response can be made to deepen the user's emotion.

  The emotional response generation program according to claim 6, wherein the computer is configured to estimate whether the state of the user's emotion is negative based on the emotion of the user input by the input unit, and the emotional state estimation Emotion response generation means for generating an emotion response that does not generate an emotion response when the state of the user's emotion estimated by the means is negative, and that is synchronized with the user's emotion input by the input means when the state is other than negative To function as.

  According to the sixth aspect of the present invention, when the user's emotion is negative, the user can be kept quiet without making an emotional response.

  8. The emotion response generation program according to claim 7, wherein the emotion state estimation means for estimating whether or not the state of the user's emotion is negative based on the emotion of the user input by the input means, and the emotion state estimation Emotion response generation that does not generate an emotion response when the state of the user's emotion estimated by the means is negative, and generates an emotion response that digs deeply into the user's emotion input by the input means otherwise Function as a means.

  According to the seventh aspect of the present invention, when the user's emotion is negative, it is possible to leave the user gently without responding to the emotion, and when the user's emotion is not negative, it is possible to make an emotional response that digs deeply into the user's emotion.

  As described above, according to the present invention, it is possible to generate an appropriate emotion response depending on whether or not the user's emotion is negative.

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

  FIG. 1 is a block diagram showing a configuration of an emotion response generation apparatus according to an embodiment of the present invention. As shown in the figure, the emotion response generating apparatus includes an input unit 11 for inputting a user's utterance and a user's emotion corresponding to the utterance, and an emotional state for estimating the state of the user's emotion input by the input unit 11. An estimation unit 12, an emotion response generation unit 13 that generates an emotion response based on the emotion state estimated by the emotion state estimation unit 12, an output unit 14 that outputs an emotion response generated by the emotion response generation unit 13, It includes an emotion state dictionary 15 that defines the correspondence between emotions and emotion states, and an emotion response dictionary 16 that defines emotions and emotion responses to them.

  The input unit 11 inputs a user's utterance and an emotion estimated when the user utters. The user's utterance is input by a microphone or the like. In addition, any known method may be used as a method for estimating the user's emotion, and any method may be used in the present embodiment, such as estimation from a face image, user utterance, and task achievement state.

  The output unit 14 outputs the generated emotion response by voice, using a speaker, or outputs it by an operation such as swinging by a robot capable of expressing the emotion response content.

  Next, the flow of operation of the emotion response generation apparatus according to each embodiment of the present invention will be described with reference to the flowcharts shown in FIGS.

<First Embodiment>
FIG. 2 is a flowchart showing a flow of operation of the emotion response generation apparatus according to the first embodiment.

  First, in step 100, the input unit 11 inputs a user's utterance and an emotion estimated when the user utters.

  In step 110, the emotional state estimating unit 12 estimates an emotional state corresponding to the user's emotion input by the input unit 11. For example, when “happy” is input as the user's emotion, the emotional state estimating unit 12 estimates the emotional state of “happy”. Specifically, the emotional state estimation unit 12 determines whether the emotional state of “happy” is positive, negative, surprise, or neutral using the emotional state dictionary 15 shown in FIG. In this case, it is estimated from FIG. 3 that the emotional state of “happy” is “positive”.

  In step 120, the emotional state estimation unit 12 determines whether or not the emotional state of the user estimated in step 110 is negative. If the emotional state is other than negative, the process proceeds to step 130. If the emotional state is negative, the process ends without generating an emotional response.

  In step 130, the emotion response generation unit 13 generates an emotion response according to the emotion state estimated by the emotion state estimation unit 12. In the above example, since the emotional state of “happy” is “positive”, an emotional response is generated according to the emotional state “happy”. Specifically, the emotion response generation unit 13 generates an emotion response corresponding to “happy” using the emotion response dictionary 16 shown in FIG. 4. In this case, “I'm happy” is generated as an emotional response that is synchronized with the emotional state.

  In step 140, the output unit 14 outputs the emotion response generated by the emotion response generation unit 13 and ends the process. The emotion response is output by, for example, an operation by a robot that can output an emotion response by a speaker or express emotion response contents.

  In step 150, the input unit 11 determines whether or not there is a next input from the user utterance. When there is a user utterance, the input unit 11 returns to step 100 and repeats the above processing, and when there is no user utterance, the processing is terminated.

  The case where the user's emotion is “happy” has been described as an example. However, when the user's emotion is “sad”, the operation is as follows.

  When the input unit 11 inputs “sad”, the emotion state estimation unit 12 estimates “negative” as the emotion state based on the emotion state dictionary 15. In this case, no emotional response is generated, and the next user utterance is awaited.

  As described above, the emotion response generation device according to the first embodiment can take the user's emotional state into consideration, and can keep the user quiet without performing an emotional response when the emotional state is negative.

<Second Embodiment>
FIG. 5 is a flowchart showing a flow of operation of the emotion response generation apparatus according to the second embodiment. The flowchart of FIG. 5 is obtained by adding step 200 to the flowchart according to the first embodiment of FIG. The same steps as those in the first embodiment are denoted by the same reference numerals, detailed description thereof is omitted, and steps different from those in the first embodiment are mainly described.

  Specifically, when the emotion state estimation unit 12 determines in step 120 that the user's emotion state is negative, the process proceeds to step 200.

  In step 200, the emotion response generation unit 13 does not generate an emotion response but generates a conflict. In the generation of the conflict, the conflict as shown in FIG. 6 is selected at random. In addition, the context (what the emotional response generation device has generated in the utterance history so far) or the like can also be used for selection of the conflict.

  As described above, the emotion response generation apparatus according to the second embodiment can take the user's emotional state into consideration, and when the emotional state is negative, can generate a conflict and respond to the user.

<Third Embodiment>
FIG. 7 is a flowchart showing a flow of operation of the emotion response generation apparatus according to the third embodiment. The flowchart in FIG. 7 is obtained by changing step 130 in the flowchart according to the second embodiment in FIG. Note that the same steps as those of the second embodiment are denoted by the same reference numerals, detailed description thereof is omitted, and steps different from those of the second embodiment are mainly described.

  Specifically, when the emotion state estimation unit 12 determines in step 120 that the user's emotion state is not negative, the process proceeds to step 300.

  In step 300, the emotion response generation unit 13 generates a response that digs deeply into the user's speech and emotional state. In the generation of the deep digging response, the emotional response generation unit 13 is a response such as “So then?” Which is a standard expression that further extracts information on the emotional state of the user based on the template shown in FIG. If the emotional state is “happy”, generate a response that asks the reason, time, or location of the emotional state, such as “Why are you happy?” “When are you happy?” “Where are you happy?” To do. In selecting the deep digging response, the context (what the emotion response generating device has generated in the utterance history so far) or the like is used.

  The emotion response generation apparatus configured as described above can realize, for example, the following interaction example with the user. Hereinafter, the utterance of the apparatus in response to the user utterance is referred to as “system utterance”.

(Dialogue example)
User utterance 1: “Grandchild will come” (Estimated user feeling “happy”)
System utterance 1: “Why are you happy?”
User utterance 2: "I can only meet occasionally." (Estimated user feeling "sad")
System utterance 2: “That's right.”
User utterance 3: “When I was fine, I was glad to go swimming together.” (Estimated user emotion “fun”)
System utterance 3: “When did you enjoy?”
User utterance 4: “Someday, long ago.”

  As described above, the emotion response generation device according to the third exemplary embodiment considers the emotional state of the user, and when the emotional state is negative, the emotional state is kept quiet by generating a response and responding. When it is not negative, it is possible to continue the dialogue smoothly by responding deeply to the user's emotions.

  Note that the present invention is not limited to the above-described embodiment, and can be applied to a design modified within the scope of the claims.

  For example, as a means for outputting an emotional response, it may be displayed on a display capable of outputting an image in addition to a speaker or a robot.

It is a block diagram which shows the structure of the emotion response production | generation apparatus which concerns on embodiment of this invention. It is a flowchart which shows the flow of an effect | action of the emotion response production | generation apparatus which concerns on 1st Embodiment. It is a figure which shows the structure of an emotion state dictionary. It is a figure which shows the structure of an emotion response dictionary. It is a flowchart which shows the flow of an effect | action of the emotional response production | generation apparatus which concerns on 2nd Embodiment. It is a figure which shows the candidate of a match. It is a flowchart which shows the flow of an effect | action of the emotion response production | generation apparatus which concerns on 3rd Embodiment. It is a figure which shows the template which produces | generates a deep digging response.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Input part 12 Emotion state estimation part 13 Emotion response production | generation part 14 Output part 15 Emotion state dictionary 16 Emotion response dictionary

Claims (7)

An input means for inputting the user's utterance and the user's emotion corresponding to the utterance;
Emotion state estimation means for estimating whether or not the state of the user's emotion is negative based on the user's emotion input by the input means;
When the emotional state of the user estimated by the emotional state estimating means is negative, no emotional response is generated, and when it is not negative, an emotional response that synchronizes with the user's emotion input by the input means is generated. Emotion response generation means;
Output means for outputting the emotion response generated by the emotion response generation means;
An emotional response generation device comprising: An input means for inputting the user's utterance and the user's emotion corresponding to the utterance;
Emotion state estimation means for estimating whether or not the state of the user's emotion is negative based on the user's emotion input by the input means;
When the user's emotional state estimated by the emotional state estimating means is negative, no emotional response is generated, and when it is not negative, an emotional response that digs deeply into the user's emotion input by the input means is generated Emotion response generation means to
Output means for outputting the emotion response generated by the emotion response generation means;
An emotional response generation device comprising:   When the emotional state of the user estimated by the emotional state estimating unit is other than negative, the emotional response generation device uses an emotional response based on a fixed expression that further extracts information about the emotional state, or the input unit The emotion response generation apparatus according to claim 2, wherein an emotion response that asks any one of the reason, time, and place of the input user's emotion is generated.   The emotional response generation apparatus according to claim 1, wherein when the emotional state of the user estimated by the emotional state estimation means is negative, an emotional response generation device is generated.   The emotional response generation device according to claim 2 or 3, wherein when the emotional state of the user estimated by the emotional state estimation means is negative, a conflict is generated. Computer
Based on the user's emotion input by the input means, the emotion state estimation means for estimating whether the user's emotion state is negative, and the user's emotion state estimated by the emotion state estimation means is negative An emotional response generating means for generating an emotional response that does not generate an emotional response and that is synchronized with the user's emotion input by the input means when not negative,
Emotional response generation program to function as. Computer
Based on the user's emotion input by the input means, the emotion state estimation means for estimating whether the user's emotion state is negative, and the user's emotion state estimated by the emotion state estimation means is negative An emotion response generating means for generating an emotion response that does not generate an emotion response when it is other than negative and digs deeply into the user's emotion input by the input means
Emotional response generation program to function as.