JP2016024631A - Presentation device, presentation method, and presentation program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform presentation in order to enhance advertising effect.SOLUTION: A presentation device includes an estimation section and a presentation section. The estimation section estimates feeling of a user in an interaction with an interactive agent system that interacts with the user. The presentation section presents feeling information indicated by the feeling of the user estimated by the estimation section, to an advertiser of an advertisement to be output during the interaction. For example, the presentation section 133 presents a determination tree including feeling information estimated by the estimation section 132, as feeling of a user who uses the interactive agent system 1, to an advertiser terminal 60.SELECTED DRAWING: Figure 4

Description

  The present invention relates to a presentation device, a presentation method, and a presentation program.

  2. Description of the Related Art Conventionally, there is known an interactive agent system that outputs a message corresponding to a message to the user terminal in response to receiving the message from the user terminal. In such an interactive agent system, there are provided technologies for outputting various messages to messages received from user terminals and outputting messages suitable for users using the user terminals.

  Also, in the above interactive agent system, when a predetermined keyword is included in a message from the user terminal, a technique for outputting an advertisement corresponding to the keyword to the user terminal as a message is known.

JP 2013-205523 A JP 2002-297626 A

  However, the above-described conventional technology does not always improve the advertising effect. Specifically, when an advertisement is output based only on a keyword included in a message from the user terminal, the advertisement is output regardless of the state of the user, and therefore, an advertisement that is highly appealing to the user may not be output. is there.

  An object of the present application is to provide a presentation device, a presentation method, and a presentation program that can make a presentation for improving the effect of an advertisement.

  A presentation device according to the present application includes: an estimation unit that estimates an emotion of the user when a conversation is performed with the dialog agent system in an interactive agent system that interacts with the user; and emotion information that indicates an emotion of the user estimated by the estimation unit A presentation unit that presents the advertisement to the advertiser of the advertisement output during the conversation.

  According to one aspect of the embodiment, there is an effect that a presentation for improving the effect of the advertisement can be performed.

FIG. 1 is a diagram illustrating an example of a bid process according to the first embodiment. FIG. 2 is a diagram illustrating an example of a determination tree according to the first embodiment. FIG. 3 is a diagram illustrating a configuration example of the interactive agent system according to the first embodiment. FIG. 4 is a diagram illustrating a configuration example of the bidding apparatus according to the first embodiment. FIG. 5 is a diagram illustrating an example of an advertiser page according to the first embodiment. FIG. 6 is a diagram illustrating an example of a determination tree according to the first embodiment. FIG. 7 is a sequence diagram showing a bidding process procedure according to the first embodiment. FIG. 8 is a diagram illustrating an example of an advertiser page according to a modified example of the first embodiment. FIG. 9 is a diagram illustrating an example of a determination tree according to a modification of the first embodiment. FIG. 10 is a diagram illustrating an example of an advertiser page according to a modified example of the first embodiment. FIG. 11 is a diagram illustrating an example of a bid process according to the second embodiment. FIG. 12 is a diagram illustrating a configuration example of the bidding apparatus according to the second embodiment. FIG. 13 is a sequence diagram illustrating a bidding process procedure according to the second embodiment. FIG. 14 is a diagram illustrating an example of a bid process according to the third embodiment. FIG. 15 is a diagram illustrating a configuration example of a bidding apparatus according to the third embodiment. FIG. 16 is a sequence diagram showing a bidding process procedure according to the third embodiment. FIG. 17 is a hardware configuration diagram illustrating an example of a computer that realizes the function of the bidding apparatus.

  Hereinafter, modes for realizing a presentation device, a presentation method, and a presentation program according to the present application (hereinafter referred to as “embodiments”) will be described in detail with reference to the drawings. In addition, the presentation apparatus, the presentation method, and the presentation program which concern on this application are not limited by this embodiment. Moreover, in the following each embodiment, the same site | part is attached | subjected the same code | symbol and the overlapping description is abbreviate | omitted. Further, in the following description, a bidding apparatus that executes a predetermined bidding process will be described as an example of a presentation apparatus.

(First embodiment)
[1. Bidding process)
First, an example of a bidding process according to the first embodiment will be described with reference to FIG. FIG. 1 is a diagram illustrating an example of a bid process according to the first embodiment. The bidding apparatus 100 shown in FIG. 1 provides a bidding service for acquiring the right to output an advertisement based on emotion information. Specifically, the bid device 100 acquires a determination tree (for example, a corpus) from the response generation server 30 as determination information that is information about the rules used in the interactive agent system 1, and the user's user at each node in the determination tree. Estimate emotions. Thereafter, the bidding device 100 presents the judgment tree together with the emotion information to the advertiser, and accepts a bid from the advertiser for the presented judgment tree. Further, the response generation server 30 illustrated in FIG. 1 determines a response in the dialog with the user using the determination tree. In the following description, an example is described in which it is estimated whether the user's emotion is “positive”, which is a positive state, or the user's emotion is “negative”, which is a backward state. However, as will be described later, the bidding apparatus 100 may also estimate other arbitrary emotions that the user has.

  Here, the determination tree used by the response generation server 30 will be described with reference to FIG. FIG. 2 is a diagram illustrating an example of a determination tree according to the first embodiment. In other words, the dialog agent system 1 realizes a dialog with the user by using a decision tree as shown in FIG. Specifically, the response generation server 30 determines a response message to be returned to the user in response to the input message that is the user's utterance based on the determination tree as illustrated in FIG.

  In the example shown in FIG. 2, the broken line block indicates a detection node, and the solid line block indicates an operation node. Further, node IDs N101 to N112 are assigned to these blocks, respectively. For example, the broken line block ““ weather ”detection” of the node ID “N101” indicates a detection node, and the solid line block ““ Today is raining ”of the node ID“ N103 ”indicates an operation node. The detection node defines a processing procedure corresponding to the input message from the user, and the operation node defines a processing procedure corresponding to the response message. Therefore, the decision tree shown in FIG. 2 includes a detection node and an operation node.

  Further, the nodes shown in FIG. 2 are associated by being connected by an arrow line. Specifically, the node on the starting point side of the arrow line is associated as the connection source node, and the node on the arrow side is associated as the connection destination node. For example, the arrow line that associates the node with the node ID “N104” and the node with the node ID “N106” indicates that the connection source node is the detection node with the node ID “N104” and the connection destination node is the operation with the node ID “N106”. Indicates a node. The node with the node ID “N104” is also associated with the operation node with the node ID “N107” as the connection source node. Here, the response generation server 30 detects a keyword from the user's utterance, and transitions to the detection node if the detection node corresponding to the keyword is associated as the connection destination node. Specifically, after the response generation server 30 responds to the response message “Today is raining” corresponding to the operation node having the node ID “N103”, “happy” is included in the user's utterance. If it has been detected, it is determined that “delighted detection” corresponding to the node ID “N104” has been performed, and a transition is made to either the node ID “N106” or the node ID “N107” that is the connection destination node. . Then, the response message of the corresponding operation node is returned to the user. Specifically, when transitioning to the node ID “N106”, the response generation server 30 responds to the user with a response message “It was good” corresponding to the operation node of the node ID “N106”. In this way, the dialog agent system 1 advances the dialog with the user. The response generation server 30 may make a transition from each connection source node to the connection destination node using a predetermined transition probability, or may make a random transition.

  Here, the action nodes in the determination tree shown in FIG. 2 include an action node related to advertisement and an action node related to conversation. For example, the operation node with the node ID “N106” is a response message that does not include a specific store name or product name ““ It was good ”, and therefore corresponds to an operation node related to conversation. In addition, the operation node of the node ID “N107” is “How about the store A? Since it is a response message including a specific store name “”, it corresponds to an operation node related to an advertisement. The dialogue agent system 1 provides the user with an advertisement in a natural conversation by interweaving such a node related to the advertisement and a node related to the conversation.

  Returning to FIG. 1, the bidding process according to the first embodiment will be described. As shown in FIG. 1, the bid device 100 acquires a determination tree from the response generation server 30 (step S11). At this time, the bidding apparatus 100 may acquire the entire determination tree or a part of the determination tree. In addition, the bidding device 100 may acquire a determination tree from the response generation server 30 every predetermined period such as every other day.

  Subsequently, based on the acquired determination tree, the bid device 100 estimates the user's emotion at each node in the determination tree. Then, the bidding device 100 presents a judgment tree including emotion information estimated as emotions of the user who uses the interactive agent system 1 to the advertiser (step S12). In the example shown in FIG. 1, the bidding apparatus 100 surrounds a region in which a user's emotion is estimated to be positive and a region determined to be negative in the determination tree with dotted lines, together with characters representing the emotion. Present to advertisers. That is, in the example shown in FIG. 1, characters representing emotions such as “positive” and “negative” are emotion information estimated as emotions of the user who uses the interactive agent system 1. The emotion information estimated as the emotion of the user who uses the interactive agent system 1 is in any form as long as it indicates the correspondence between the user's emotion and the node so as to be distinguishable from other nodes. Also good. In addition, the bid device 100 may present the entire determination tree to the advertiser or may present a part of the determination tree including emotion information. In the case of the example in FIG. 1, the bidding apparatus 100 presents a determination tree so that a region indicating that the user's emotion is positive and a region indicating that the user's emotion is negative can be identified. In addition, although estimation of a user's emotion is performed based on a detection node or an action node, details will be described later.

  Subsequently, the bid device 100 receives bid information from the advertiser (step S13). At this time, the accepted bid information may include information on an operation node related to an advertisement, information on a bid price, and the like, which are new rules to be added.

  Then, the bidding device 100 updates the determination tree based on the information of the operation node related to the advertisement, which is a new rule included in the bid information received from the advertiser determined to make a successful bid, and the updated determination tree is transmitted to the response generation server. 30 (step S14). Thereafter, the response generation server 30 determines a response message to be returned to the user in response to the input message that is the user's utterance based on the updated determination tree.

  Thus, the bid device 100 according to the first embodiment can present a determination tree including emotion information indicating the user's emotion estimated to the advertiser. Thereby, the bid apparatus 100 can accept an advertisement bid from the advertiser for the estimated emotion of the user. In other words, the advertiser can add an advertisement in accordance with the emotion of the user suitable for its own advertisement by receiving the presentation of emotion information indicating the estimated emotion of the user. Thereby, the bid apparatus 100 can generate a determination tree that outputs an advertisement in accordance with the user's emotion. That is, the bid device 100 can output a highly appealing advertisement to the user at an appropriate timing. The bid process may be executed in cooperation between the presentation device and the bid device.

[2. Dialogue agent system configuration]
As illustrated in FIG. 3, the dialog agent system 1 according to the present embodiment includes a user terminal 10, a speech recognition server 20, a response generation server 30, a speech synthesis server 40, and a service API (Application Program Interface). A server 50, an advertiser terminal 60, and a bidding device 100 are included. The user terminal 10, the speech recognition server 20, the response generation server 30, the speech synthesis server 40, the service API server 50, the advertiser terminal 60, and the bidding device 100 transmit and receive information via the communication network 70. I do.

  From here, the outline | summary of the process in which the dialogue agent system 1 provides a voice service to a user is demonstrated. The user terminal 10 is an information processing apparatus such as a mobile phone, a smartphone, a PDA (Personal Digital Assistant), a tablet PC, a notebook PC, or a desktop PC. When the user terminal 10 detects the user's speech after the application is started, the user terminal 10 transmits speech data of the speech to the speech recognition server 20.

  When the speech recognition server 20 receives the speech data from the user terminal 10, the speech recognition server 20 converts the speech data into text data, and transmits the speech text data to the user terminal 10. The user terminal 10 that has received the utterance text data from the voice recognition server 20 transmits the utterance text data to the response generation server 30.

  The response generation server 30 is a computer that transmits response generation information for the user's utterance to the user terminal 10 in accordance with the utterance text data transmitted by the user terminal 10. When the response generation server 30 receives the text data of the utterance from the user terminal 10, the response generation server 30 designates data search conditions necessary for generating the response, and sends data to the service API server 50 corresponding to the application started by the user terminal 10. Make a request.

  The service API server 50 transmits data to the response generation server 30 in accordance with the search condition received from the response generation server 30. For example, the service API server 50 performs a process of acquiring response data, and transmits the acquired response data to the response generation server 30. For example, the service API server 50 transmits XML (Extensible Markup Language) data to the response generation server 30 as response data.

  When the response generation server 30 receives, for example, XML data as response data from the service API server 50, the response generation server 30 extracts the data from the XML data, and converts the XML data into HTML data. Then, the response generation server 30 transmits the response utterance display text data to the speech synthesis server 40. The speech synthesis server 40 transmits to the response generation server 30 an intermediate notation for response speech generated by performing speech synthesis processing for synthesizing speech from text data for response speech display. The response generation server 30 transmits the intermediate notation for response utterance, text data for display of response utterance, and HTML data to the user terminal 10.

  The user terminal 10 outputs the response voice using the received intermediate notation for response utterance, and displays the response content using text data for displaying the response utterance and HTML data. In this way, the dialogue agent system 1 realizes a voice service that makes an appropriate response to the user's utterance.

[3. Bidding equipment configuration]
Next, the configuration of the bidding apparatus 100 according to the first embodiment will be described with reference to FIG. FIG. 4 is a diagram illustrating a configuration example of the bidding apparatus 100 according to the embodiment. As illustrated in FIG. 4, the bid device 100 includes a communication unit 110, a storage unit 120, and a control unit 130. The bid device 100 includes an input unit (for example, a keyboard and a mouse) that receives various operations from an administrator of the bid device 100 and a display unit (for example, a liquid crystal display) for displaying various types of information. May be.

(Communication unit 110)
The communication unit 110 is realized by a NIC (Network Interface Card), for example. The communication unit 110 is connected to the communication network 70 in a wired or wireless manner. The communication unit 110 transmits and receives information to and from the response generation server 30 and the advertiser terminal 60 via the communication network 70.

(Storage unit 120)
The storage unit 120 stores, for example, the judgment tree acquired from the response generation server 30, emotion information estimated as a user's emotion using the interactive agent system 1, bid information registered by the advertiser, and the like. The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk.

  The control unit 130 is configured so that, for example, various programs (corresponding to an example of a bidding program) stored in a storage device inside the bidding device 100 are stored in the RAM by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like. This is realized by being executed as The control unit 130 is realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

  As shown in FIG. 4, the control unit 130 includes an acquisition unit 131, an estimation unit 132, a presentation unit 133, a reception unit 134, and a transmission unit 135, and functions and functions of information processing described below. Realize or execute. Note that the internal configuration of the control unit 130 is not limited to the configuration illustrated in FIG. 4, and may be another configuration as long as the information processing described below is performed. In addition, the connection relationship between the processing units included in the control unit 130 is not limited to the connection relationship illustrated in FIG. 4, and may be another connection relationship.

(Acquisition part 131)
The acquisition unit 131 acquires from the response generation server 30 a determination tree that is information related to the rules of the dialog that the response generation server 30 uses for dialog with the user. At this time, the acquisition unit 131 may acquire the entire determination tree or a part of the determination tree. In addition, the bidding device 100 may acquire a determination tree from the response generation server 30 every predetermined period such as every other day.

(Estimation unit 132)
The estimation unit 132 estimates the emotion of the user when having a conversation with the dialog agent system 1 based on the dialog rules used by the dialog agent system 1. For example, the estimation unit 132 estimates the user's emotion at each node in the determination tree acquired from the response generation server 30 by the acquisition unit 131. Here, the emotion estimated by the estimation unit 132 can be appropriately set in accordance with the classification of human emotions such as happy, angry, sad, and fun. Hereinafter, the emotion estimated by the estimation unit 132 will be described as two types, positive and negative. Note that the positive estimated by the estimation unit 132 includes emotions that indicate joy and enjoyment such as high tension, fun, comfortable, happy, and good tempo. Moreover, the negative estimated by the estimation unit 132 is an emotion including anger, sadness, guilt, fear, and disgust, such as low tension, unpleasantness, uncomfortable, unpleasant, and poor tempo.

  For example, the estimation unit 132 uses information on each node in the determination tree in order to estimate a user's emotion in the node. Referring to the example shown in FIG. 2, since the positive keyword “happy” is detected for the detection node “joyful detection” of N104, the estimation unit 132 determines that the user at the detection node of N104 Estimate that emotion is positive. Then, the estimation unit 132 determines that the operation node “Now it was good” of N106, which is the connection destination node of the detection node N104, and the operation node ““ A storefront isn't it? Also for “”, the user's emotion is estimated to be positive in the same manner as the detection node of N104. On the other hand, for the detection node “dislike” detection of N105, since the negative keyword “dislike” is detected, the estimation unit 132 estimates that the user's emotion at the detection node of N105 is negative. And the estimation part 132 estimates that a user's emotion is negative also about the node of connection destination node N108-N112 after the detection node of N104 similarly to the detection node of N104. Thus, the estimation unit 132 may estimate the user's emotion at each node using the user's emotion estimated for the connection source node. Moreover, the estimation part 132 may estimate the user's emotion in the node based only on the information of the node itself irrespective of a connection origin node or a connection destination node. Moreover, the estimation part 132 may estimate the user's emotion in a node based on the database which matched and memorize | stored the keyword for every classification of emotion, for example. In this case, for example, when there is a bias in the appearance of a keyword in a certain area in the determination tree, the estimation unit 132 estimates that the emotion in the node included in that area is an emotion corresponding to the keyword that appears biased. May be. For example, the estimation unit 132 determines the user's emotion in a node included in a region where a keyword associated with a specific emotion appears more than a predetermined ratio compared to a keyword associated with another emotion, You may presume that specific emotion.

  Moreover, the estimation part 132 may use the information regarding a user's utterance when the node was used in the past for the estimation of a user's emotion in each node in a determination tree, for example. In this case, for example, when the inflection of the user's utterance when the node has been used in the past is small or when the user's utterance tempo is low, the estimation unit 132 estimates that the user's emotion at the node is negative. May be. In addition, when using information related to past user utterances, the acquisition unit 131 acquires information related to past user utterances from the response generation server 30 in advance. At this time, the storage unit 120 stores information related to past user utterances acquired from the response generation server 30 by the acquisition unit 131. In addition, the process in which the estimation part 132 estimates a user's emotion is not limited to the process mentioned above. For example, the estimation unit 132 estimates the user's emotion based on the time until the conversation moves from the utterance corresponding to a certain node to the utterance corresponding to the other node, the route connecting the nodes corresponding to the conversation, and the like. May be. Moreover, the estimation part 132 may provide the score calculated for every classification of emotion with respect to each node, and may estimate a user's emotion based on this score. In addition, the estimation unit 132 performs language analysis on conversations corresponding to a plurality of connected nodes, and based on the results of the language analysis, estimates an emotion of the user who is performing the conversation, and estimates the estimated user May be the user's emotion at the node to be processed.

(Presentation unit 133)
The presentation unit 133 presents emotion information indicating the user's emotion estimated by the estimation unit to the advertiser of the advertisement output during the conversation. For example, the presentation unit 133 presents to the advertiser terminal 60 a determination tree that includes emotion information estimated by the estimation unit 132 as a user's emotion using the interactive agent system 1.

  Here, FIG. 5 shows an example of the advertiser page presented to the advertiser terminal 60 by the presentation unit 133 according to the first embodiment. The presentation unit 133 publishes the determination tree acquired from the response generation server 30 and emotion information estimated as the user's emotion at each node in the determination tree on the advertiser page W10. In the example shown in FIG. 5, from the detection block ““ weather ”detection” of the node ID “N101” to the operation block ““ tea in the store C? ”Of the node ID“ N112 ”. Twelve nodes up to “detect” are posted on the advertiser page W10. In the example shown in FIG. 5, three nodes N104, N106, and N107 are listed as nodes that are estimated to have positive user emotions, and six nodes N105 and N108 to N112 are negative for user emotions. It is posted as a node estimated to be. Specifically, in the advertiser page W10, the presentation unit 133 surrounds a region R11 including N104, N106, and N107 estimated to be positive in the user's emotion with a dotted line, and presents it together with the characters “positive”. In addition, the presentation unit 133 surrounds the region R12 including N105 and N108 to N112, which are estimated to be negative in the user's emotion, in the advertiser page W10, and presents it together with the character “negative”. In this way, the presentation unit 133 presents the nodes clustered so as to be identifiable for each corresponding emotion together with the emotion information. This makes it easier for the user to visually recognize which emotion each node corresponds to. Note that the presentation unit 133 only needs to be able to present a node for each corresponding emotion in an identifiable manner. For example, the presentation unit 133 may present the node with a different color for each corresponding emotion or may display emotion information for each node. Good. Although illustration is omitted here, in the advertiser page W10, when the advertiser selects a node in the region R11 or the region R12, a bid page for registering bid information is presented to the advertiser terminal 60. The advertiser performs a bid by inputting information such as a new node to be added and a bid price using the bid page.

(Reception unit 134)
The accepting unit 134 receives bid information from the advertiser terminal 60 via a predetermined bid page. Thereafter, the reception unit 134 updates the determination tree based on information such as an action node that is a new node related to the advertisement included in the bid information received from the advertiser whose bid is determined. Here, it is determined whether a new node included in the bid information received by the reception unit 134 is added to the determination tree. As a result, since the reception unit 134 adds only new nodes that are determined to be added to the determination tree, it is possible to prevent an inappropriate node from including an expression that should not be used. . For example, the reception unit 134 may determine whether to add a new node of the received bid information. Further, the administrator of the bid device 100 may determine a new node to be added to the determination tree from the plurality of registered bid information, and add the determination node to the determination tree. At this time, the reception unit 134 and the administrator may make the determination based on the new node, or make the determination based on the various information included in the bid information, even if the determination is based on the bid price. You may go. Note that an advertiser who is permitted to register bid information may be notified of a result regarding whether or not a new node added by the advertiser can be added by an appropriate means.

  FIG. 6 shows the determination tree updated in the example shown in FIG. 5, and N113 is newly added as a connection destination node of N104 in region R11 in which the user's emotion is estimated to be positive in advertiser page W10. Has been. Specifically, how about the operation block ““ Novel “XYZ” ”of node ID“ N113 ”? "" Is added. By using this decision tree, “How about a novel“ XYZ ”in the flow of dialogue that presumes that the user's emotion is positive? "Can be output.

(Transmitter 135)
The transmission unit 135 transmits the updated determination tree to the response generation server 30. And the response production | generation server 30 which received the determination tree from the transmission part 135 determines the response message returned to a user with respect to the input message which is a user's utterance based on the received determination tree.

[4. (Bid processing procedure)
Next, referring to FIG. 7, the procedure of the bidding process by the interactive agent system 1 according to the first embodiment will be described. FIG. 7 is a sequence diagram showing a bidding process procedure by the interactive agent system 1 according to the first embodiment.

  As illustrated in FIG. 7, the bid device 100 acquires a determination tree, which is information related to the rules of the dialogue used by the response generation server 30 for the dialog with the user, from the response generation server 30 (step S101). In this case, the bid device 100 stores the acquired determination tree in the storage unit 120.

  Then, the bidding apparatus 100 estimates the user's emotion at each node in the determination tree based on the acquired determination tree (step S102). Then, the bidding apparatus 100 presents a determination tree including emotion information estimated as the emotion of the user who uses the interactive agent system 1 (step S103).

  Subsequently, the bid device 100 receives bid information including a new node from the advertiser terminal 60 via a predetermined bid page (step S104).

  Subsequently, the bidding apparatus 100 generates a determination tree including the new node received from the advertiser terminal 60 (step S105), and transmits the determination tree including the new node to the response generation server 30 (step S106). ). Thereafter, the response generation server 30 determines a response message to be returned to the user in response to the input message that is the user's utterance based on the determination tree including the new node received from the bid device 100, and provides a voice service. (Step S107).

[5. (Modification)
The dialog agent system 1 according to the first embodiment described above may be implemented in various different forms other than the first embodiment. Therefore, in the following, an embodiment in which the above-described dialog agent system 1 is modified will be described.

[5-1. (Changes in the presentation content according to the type of business of the advertiser)
In the first embodiment, the presentation unit 133 of the bid device 100 presented to the advertiser terminal 60 a determination tree that includes emotion information estimated by the estimation unit 132 as the emotion of the user who uses the interactive agent system 1. However, the presentation unit 133 varies the conversation area in which the user is estimated to have a predetermined emotion according to the business type of the advertiser.

  FIG. 8 shows an example of an advertiser page presented to the advertiser terminal 60 of an advertiser whose business type is a pharmaceutical company by the presentation unit 133. Here, in the bidding apparatus 100, it is assumed that the business type “pharmaceutical company” is associated with the emotion “negative”. The presentation unit 133 publishes the determination tree acquired from the response generation server 30 and emotion information estimated as the user's emotion at each node in the determination tree on the advertiser page W20. In the example shown in FIG. 8, 12 nodes from N101 to N112 are posted on the advertiser page W20, similarly to the advertiser page W10 shown in FIG.

  Here, in the example illustrated in FIG. 8, six nodes N105 and N108 to N112 are listed as nodes that are estimated to have negative user emotions. Specifically, in the advertiser page W20, the presentation unit 133 surrounds a region R21 including N105 and N108 to N112, which is estimated to be negative for the user's emotion, with a dotted line and presents it together with the character “negative”. In this way, the presentation unit 133 clusters the nodes corresponding to the user's emotion “negative” and displays the nodes together with the emotion information in an identifiable manner. On the other hand, in the example shown in FIG. 8, it is not estimated that the user's emotion is positive in the three nodes N104, N106, and N107. Specifically, in the advertiser page W10, the presentation unit 133 displays other nodes in which the user's emotion is also estimated in the same display manner in the N104, N106, and N107 that are estimated that the user's emotion is positive. Present. In this way, depending on the advertiser's industry, the area of conversation in which the user is presumed to have a predetermined emotion is changed, and the areas of the advertiser's business are categorized and presented separately for each advertiser's industry. It is possible to make a presentation for improving the effectiveness of the advertisement according to the type of business.

  FIG. 9 shows the determination tree updated in the example shown in FIG. 8, and N114 is newly added as a connection destination node of N109 in region R21 in which the user's emotion is estimated to be negative in advertiser page W20. Has been. Specifically, how is the operation block ““ cold medicine M? ”Of the node ID“ N114 ”. "" Is added. By using this determination tree, it is possible to output an advertisement “How about the cold medicine M?” In the flow of dialogue in which the user's emotion is estimated to be negative. In the example illustrated in FIG. 8, the bidding apparatus 100 presents to the advertiser, which is a pharmaceutical company, a node in which the user's emotion is estimated to be negative, but presents which emotion can be identified. This may be selected by the advertiser. Further, the bidding apparatus 100 may estimate which emotion is desired to be identifiably presented according to the business type of the advertiser. In this case, if the advertiser's business type is a hospital, a funeral company, or the like, the bidding apparatus 100 is estimated to be an advertiser of a business type that desires to output an advertisement in a conversation in which the user's emotion is estimated to be negative. May be.

[5-2. (Bid price presentation)
In the first embodiment, the presentation unit 133 of the bid device 100 presented to the advertiser terminal 60 a determination tree that includes emotion information estimated by the estimation unit 132 as the emotion of the user who uses the interactive agent system 1. However, the presentation unit 133 presents information related to a bid price for outputting an advertisement during a conversation included in a predetermined area, together with a determination tree including emotion information estimated as emotions of a user who uses the interactive agent system 1. May be.

  Here, FIG. 10 shows an example of the advertiser page presented to the advertiser terminal 60 by the presentation unit 133. The presenting unit 133 publishes a bid price together with a determination tree including emotion information estimated as user emotion on the advertiser page W30. In the example shown in FIG. 10, 12 nodes from N101 to N112 are posted on the advertiser page W30, similarly to the advertiser page W10 shown in FIG. In the example shown in FIG. 10, three nodes N104, N106, and N107 are listed as nodes that are estimated to have positive emotions, and six nodes N105 and N108 to N112 are negative in terms of user emotions. It is posted as a node estimated to be. Specifically, in the advertiser page W30, the presentation unit 133 surrounds a region R31 including N104, N106, and N107 estimated to be positive for the user's emotion with a dotted line, and presents it together with the characters “positive”. In addition, the presentation unit 133 surrounds the region R32 including N105 and N108 to N112, which are estimated to be negative in the user's emotion, on the advertiser page W30, and presents it together with the character “negative”. In this way, the presentation unit 133 presents the nodes clustered so as to be identifiable for each corresponding emotion together with the emotion information.

  In the example illustrated in FIG. 10, the presentation unit 133 includes information on the bid price such as “reference registration fee: ¥ 10,000 yen” and “reference fee: ΔΔ yen / once” together with the characters “positive” in the region R31. Present. Note that the information regarding the bid price is a reference price necessary for adding a new rule to the region R31, and the advertiser registers bid information based on this information. In addition, the presentation unit 133 presents information on the bid price such as “reference registration fee: □ 10,000 yen” and “reference fee: ☆☆ yen / once” together with the character “negative” in the region R32. Thereby, the advertiser can determine whether or not to register bid information based on the bid price. In addition, the information regarding the bid price presented by the presentation unit 133 is not limited to the above. For example, when accepting a bid with a limited period, it can be appropriately changed depending on the bid conditions such as presenting “XX million yen / month”. It is. In addition, the presentation unit 133 may present a registration fee and a fee that are decided immediately as a bid price. In this case, the accepting unit 134 accepts bid information including an intention to bid and information related to the advertisement from the advertiser via a predetermined bid page.

[5-3. Bid price)
Here, the bid device 100 may calculate the bid price presented by the presentation unit 133 according to the value of the conversation included in the area. For example, the bidding apparatus 100 may set the bid price of an area including a rule that has been used many times higher than the bid price of another area.

  Further, the bid device 100 may calculate the bid price presented by the presentation unit 133 according to the user's emotion in the area. For example, the bid device 100 may set a bid price for an area where the user's emotion is estimated to be positive to be higher than a bid price for an area where the user's emotion is estimated to be positive. In addition, for the advertiser who desires to output an advertisement in a region where the user's emotion is estimated to be negative, the bid device 100 is more than the other advertisers regarding the region where the user's emotion is estimated to be negative. You may also present an expensive bid price.

[5-4. (Advertising based on keywords)
In the first embodiment, the bid device 100 updates the determination tree and transmits it to the response generation server 30. However, the bid device 100 may transmit the advertisement information to the response generation server 30 each time. In this case, the bidding apparatus 100 may accept an emotion that a bid is desired, a keyword that is desired to output an advertisement, and advertisement information that is desired to be output, for example, bid information that includes a bid / a bid price per time (FIG. 11). reference). That is, the bid device 100 accepts an advertisement bid for each keyword for each emotion. In this case, the bid device 100 may include an extraction unit that extracts bid information corresponding to the specified emotion from the registered bid information.

  Here, a process after the reception unit 134 of the bid device 100 receives bid information including a keyword for which an advertisement output is desired will be described. First, when the user is conversing with the dialog agent system, the acquisition unit 131 acquires, from the response generation server 30, information related to the node used in the dialog at the time of acquisition and information related to the user. In this case, the information regarding the user includes information regarding the content of the user's utterance, for example, text information. Then, the extracting unit (not shown) extracts bid information corresponding to the emotion information that matches the emotion estimated in the node used in the conversation at the time of acquisition from the bid information received by the receiving unit 134. Here, when the keyword of the extracted bid information is included in the information related to the content of the user's utterance, the transmission unit 135 transmits the advertisement information in the extracted bid information to the response generation server 30. Thereafter, the response generation server 30 outputs an advertisement based on the received advertisement information. Thereby, the advertiser can output an advertisement in a dialogue in which both his / her desired emotion and keyword correspond, and can output an advertisement with high appeal to the user. Moreover, the transmission part 135 may transmit the advertisement information in the extracted bid information to the response generation server 30 when the information about the content of the user's utterance is related to the extracted bid information keyword. As used herein, “related to a keyword” means that when the keyword is “Kima”, a word having a meaning related to the meaning of the keyword, such as “Indian cuisine”, “Curry”, or “Lassi”, is spoken by the user. It is included in the information related to the contents of The keyword of the bid information may be associated with a word used for detection at the detection node. In this case, the bidding apparatus 100, when the keyword used for the bid information and the word used for detection in the detection node used in the dialog at the time of acquisition satisfy the predetermined condition, for example, when they match or relate, Processing may be performed.

[5-5. Device configuration〕
Further, the bidding apparatus 100 illustrated in FIG. 4 may not include the storage unit 120. Specifically, the bidding device 100 may be connected to a database server that holds the storage unit 120.

[5-6. effect〕
As described above, the bid device 100 according to the first embodiment includes the estimation unit 132 and the presentation unit 133. The estimation unit 132 estimates a user's emotion when having a conversation with the dialog agent system in the dialog agent system that interacts with the user. The presentation unit 133 presents emotion information indicating the user's emotion estimated by the estimation unit 132 to the advertiser of the advertisement output during the conversation.

  Thereby, the bid apparatus 100 according to the first embodiment can make a presentation for improving the effect of the advertisement.

  In addition, the bid device 100 according to the first embodiment includes a reception unit 134. The reception unit 134 receives an advertisement corresponding to the emotion information presented by the presentation unit 133 from the advertiser.

  Thereby, the bid apparatus 100 according to the first embodiment can accept an advertisement according to emotion information from the advertiser.

  In addition, in the bidding device 100 according to the first embodiment, the presentation unit 133 presents, to the advertiser, the conversation area in which the user is presumed to have a predetermined emotion in the conversation with the conversation agent system together with the emotion information. .

  Thereby, the bid apparatus 100 according to the first embodiment can make a presentation for improving the effect of the advertisement.

  In addition, in the bid device 100 according to the first embodiment, the presentation unit 133 presents to the advertiser an area in which the user's emotion is estimated to be positive and another area that can be identified.

  Thereby, the bid apparatus 100 according to the first embodiment makes it easier for the user to visually recognize which region of the node corresponds to a positive emotion, and can make a presentation for improving the effectiveness of the advertisement.

  In addition, in the bidding apparatus 100 according to the first embodiment, the presentation unit 133 presents to the advertiser an area in which the user's emotion is estimated to be negative and another area that can be identified.

  Thereby, the bid apparatus 100 according to the first embodiment makes it easier for the user to visually recognize which region of the node corresponds to the negative emotion, and can make a presentation for improving the effectiveness of the advertisement.

  In addition, in the bid device 100 according to the first embodiment, the presentation unit 133 presents a region corresponding to a predetermined emotion by clustering regions where the user is estimated to be the predetermined emotion.

  Thereby, the bid apparatus 100 according to the first embodiment makes it easier for the user to visually recognize which emotion each node corresponds to, and can make a presentation for improving the effect of the advertisement.

  Further, in the bid device 100 according to the first embodiment, the presentation unit 133 varies the conversation area in which the user is estimated to have a predetermined emotion according to the business type of the advertiser.

  Thereby, the bid apparatus 100 according to the first embodiment performs the presentation for improving the effectiveness of the advertisement according to the business type of the advertiser by dividing the area to be presented identifiable for each advertiser. Can do.

  In the bid device 100 according to the first embodiment, the presentation unit 133 presents a bid price for outputting an advertisement during the conversation included in the region.

  Thereby, since the bidder 100 according to the first embodiment can determine whether or not the advertiser bids based on the bid price, the bid apparatus 100 presents to improve the effectiveness of the advertisement according to the business type of the advertiser. be able to.

  Further, in the bidding device 100 according to the first embodiment, the presentation unit 133 presents a bid price calculated according to the value of the conversation included in the area.

  Thereby, since the bidder 100 according to the first embodiment can determine whether or not the advertiser bids based on the bid price, the bid apparatus 100 presents to improve the effectiveness of the advertisement according to the business type of the advertiser. be able to.

  Further, in the bid device 100 according to the first embodiment, the presentation unit 133 presents a bid price calculated according to the emotion of the user corresponding to the region.

  Thereby, since the bidder 100 according to the first embodiment can determine whether or not the advertiser bids based on the bid price, it can make a presentation for improving the effectiveness of the advertisement.

(Second Embodiment)
As described in the first embodiment, in the interactive agent system 1, the emotion of the user at each node in the determination tree acquired from the response generation server 30 by the acquisition unit 131 is estimated. In other words, the dialog agent system 1 provides a bidding service based on the user's emotion estimated from information generated by a dialog with a past user, such as a decision tree. Therefore, in the second embodiment, an example will be described in which a bidding service based on the user's emotion estimated from information related to the user having a conversation based on the dialog agent system and the determination tree is provided. Note that the configuration of the interactive agent system according to the second embodiment is the same as the configuration shown in FIG.

[1. Bidding process)
First, the bidding process according to the second embodiment will be described using the example shown in FIG. As shown in FIG. 11, the bid device 200 presents bid list information for each estimated emotion of the user to the advertiser (step S <b> 21). In the example shown in FIG. 11, for the emotion “positive”, the bidding apparatus 200 has the advertisement information desired to be output “Delicious curry at A store” and the current bid price is “XX yen per charge”. ”. In addition, the bid device 200 has the advertisement information desired to be output for the emotion “negative” as “Please consult with F company if it is a funeral”, and the current bid price is “□□ ¥” per charge. Present that. The bid price is not limited to the above. For example, when accepting a bid for a limited time period, the bid price can be appropriately changed depending on the bid conditions, such as presenting “XX million yen / month”.

  Subsequently, the bid device 200 receives bid information from the advertiser (step S22). At this time, the accepted bid information includes the emotions of the user who wishes to bid, the advertisement information desired to be output, for example, the bid price per charge / time, and the like. Then, when the accepted bid price exceeds the current bid price, the bid device 200 updates the corresponding emotion information in the advertisement list information with the accepted bid information.

  Then, when the user is conversing with the dialog agent system (step S23), the response generation server 30 transmits information regarding the user who has a conversation based on the determination tree with the dialog agent system to the bid device 200 (step S24). . Here, the information regarding the user includes voice information acquired from the user's utterance. Thereafter, the bid device 200 estimates the user's emotion based on the information about the user acquired from the response generation server 30. At this time, for example, when the user's utterance is small or when the user's utterance is slow, the bidding apparatus 200 estimates that the emotion of the user who has a conversation based on the dialogue tree and the determination tree is negative. May be. Then, the bid device 200 extracts bid information corresponding to the emotion that matches the emotion estimated from the information related to the user from the bid list information (step S25). If there is an emotion that matches the emotion estimated from the information about the user in the bid list information, the bid device 200 transmits the advertisement information in the extracted bid information to the response generation server 30 (step S26). Thereafter, the response generation server 30 outputs the advertisement of the received advertisement information to the user. Specifically, when the emotion estimated from the information related to the user is positive, the bid apparatus 200 has the bid information corresponding to the emotion “positive” in the bid list information. “The curry of store A is delicious” is transmitted to the response generation server 30. Then, the response generation server 30 outputs “Delicious curry at store A” to the user at an appropriate timing.

  In this way, the bid information 200 according to the second embodiment accepts a bid for each estimated user emotion, and the advertisement information that matches the user emotion estimated from the information about the user acquired from the response generation server 30 Is transmitted to the response generation server 30. Therefore, the dialog agent system can output to the user an advertisement based on the user's emotion estimated from information regarding the user who has a conversation based on the determination tree. Specifically, the bid device 200 can present emotion information indicating the estimated user emotion to the advertiser. Thereby, the bid apparatus 200 can accept an advertisement bid from the advertiser for the estimated user's emotion. Thereby, the bid apparatus 200 can accept a bid for improving the effect of the advertisement. That is, the bid device 200 can output a highly appealing advertisement to the user at an appropriate timing. The bid device 100 may transmit advertisement information in the bid information similar to or in a predetermined correspondence with the user's emotion estimated from the information about the user acquired from the response generation server 30 to the response generation server 30.

[2. Bidding equipment configuration]
Next, the configuration of the bidding apparatus 200 according to the second embodiment will be described with reference to FIG. FIG. 12 is a diagram illustrating a configuration example of the bidding apparatus 200 according to the embodiment. As illustrated in FIG. 12, the bid device 200 includes a communication unit 110, a storage unit 220, and a control unit 230. In addition, below, the description which overlaps with the bid apparatus 100 which concerns on 1st Embodiment shown in FIG. 4 is abbreviate | omitted.

(Storage unit 220)
The storage unit 220 stores, for example, bid list information based on bid information registered by the advertiser as illustrated in FIG. 11, information about the user acquired from the response generation server 30, and the like.

  As shown in FIG. 12, the control unit 230 includes an acquisition unit 231, an estimation unit 232, a presentation unit 233, a reception unit 234, a transmission unit 235, and an extraction unit 236. Information described below Implement or execute a processing function or action. Note that the internal configuration of the control unit 230 is not limited to the configuration illustrated in FIG. 12, and may be another configuration as long as information processing described later is performed. Further, the connection relationship between the processing units included in the control unit 230 is not limited to the connection relationship illustrated in FIG. 12, and may be another connection relationship.

(Acquisition unit 231)
The acquisition unit 231 acquires information about a user who has a conversation based on the dialog agent system and the determination tree from the response generation server 30. Here, the acquisition unit 231 may acquire information related to a user who has a conversation based on the dialog agent system and the determination tree, for example, from the response generation server 30 every time the conversation progresses, that is, every time a node transitions.

(Estimation unit 232)
The estimation unit 232 estimates the emotion of the user when having a conversation with the dialog agent system in the dialog agent system that interacts with the user. For example, the estimation unit 232 estimates the user's emotion based on information about the user acquired from the response generation server 30 by the acquisition unit 231. Here, the emotion estimated by the estimation unit 232 can be appropriately set in accordance with the classification of human emotions such as happy, angry, sad, and fun. Hereinafter, the emotion estimated by the estimation unit 232 will be described as two types, positive and negative. In this case, the estimation unit 232 estimates that the emotion of the user having a conversation based on the dialogue agent system and the determination tree is positive, for example, when the user's utterance inflection is large or the user's utterance is good. May be. In addition, when the acquisition part 231 acquires the emotion of the user who has a conversation based on a dialog agent system and a determination tree, the estimation part 232 may use the information regarding the user's emotion which the acquisition part 231 acquired. At this time, the storage unit 220 stores information on the user's emotion acquired from the response generation server 30 by the acquisition unit 231.

(Presentation unit 233)
The presenting unit 233 presents emotion information indicating the user's emotion estimated by the estimating unit to the advertiser of the advertisement output during the conversation. For example, the presentation unit 233 presents bid list information for each estimated emotion of the user to the advertiser terminal 60. In the example shown in FIG. 11, for the emotion “positive”, the presentation unit 233 indicates that the advertisement information desired to be output is “Delicious curry at store A”, and the current bid price is “XX yen per charge”. ”. In addition, the presentation unit 233 has the advertisement information that is desired to be output for the emotion “negative” as “Please consult with F company if it is a funeral”, and the current bid price is “□□ ¥” per charge. Present that. The bid price is not limited to the above. For example, when accepting a bid for a limited time period, the bid price can be appropriately changed depending on the bid conditions, such as presenting “XX million yen / month”.

  Although illustration is omitted here, the presentation unit 233 presents a bid page for registering bid information to the advertiser terminal 60. The advertiser performs a bid by inputting bid information including the emotion of the user who wishes to bid, the advertisement information desired to be output, for example, a bid / a bid price per time.

(Reception unit 234)
The accepting unit 234 receives bid information from the advertiser terminal 60 via a predetermined bid page. In the example shown in FIG. 11, the bid information received by the receiving unit 234 includes emotions for which bids are desired, advertisement information for which output is desired, for example, a bid price per charge. When the received bid price exceeds the current bid price, the entry receiving unit 234 updates the corresponding emotion column in the advertisement list information with the received bid information.

(Transmitter 235)
The transmission unit 235 transmits the advertisement information in the bid information extracted by the extraction unit 236 described below to the response generation server 30 in the advertisement list information in the example illustrated in FIG. And the response production | generation server 30 outputs the advertisement of the advertisement information received from the transmission part 235 at a suitable timing with respect to a user.

(Extractor 236)
Based on the comparison result between the emotion of the user estimated based on the information about the user acquired by the acquisition unit 231 and the emotion estimated by the estimation unit 232 for the conversation the user is performing, Extract advertisements that are output to. That is, in the present embodiment, the extraction unit 236 extracts bid information corresponding to an emotion that matches the emotion estimated by the estimation unit 232 from the bid list information in the example illustrated in FIG. Therefore, when there is an emotion in the bid list information that matches the emotion estimated from the information related to the user, as described above, the transmission unit 235 sends the advertisement information in the bid information extracted by the extraction unit 236 to the response generation server. To 30. Specifically, when the emotion estimated from the information about the user is negative, the extraction unit 236 extracts bid information corresponding to the emotion “negative” from the bid list information. Then, the transmitting unit 235 transmits the emotion “negative” advertisement information “please consult company F for funeral” to the response generation server 30. Thereafter, the response generation server 30 outputs “consult F company for funeral” to the user at an appropriate timing.

[3. (Bid processing procedure)
Next, with reference to FIG. 13, a procedure of bidding processing by the interactive agent system according to the second embodiment will be described. FIG. 13 is a sequence diagram showing a bidding process procedure by the dialog agent system according to the second embodiment.

  As shown in FIG. 12, the bidding apparatus 200 presents bid list information including emotion information estimated as emotions of the user who uses the interactive agent system to the advertiser (step S201). Subsequently, the bid device 200 receives bid information including advertisement information from the advertiser terminal 60 (step S202).

  Then, when the user is conversing with the dialog agent system (step S203), the response generation server 30 transmits information regarding the user who has a conversation based on the dialog agent system and the determination tree to the bid device 200 (step S204). . Here, the information regarding the user includes voice information acquired from the user's utterance.

  Subsequently, the bid device 200 estimates the user's emotion based on the information about the user acquired from the response generation server 30. Then, the bid device 200 extracts bid information that matches the emotion estimated from the information related to the user from the bid list information (step S205).

  If there is an emotion in the bid list information that matches the emotion estimated from the information about the user, the bid device 200 transmits the advertisement information in the bid information corresponding to the emotion to the response generation server 30 (step S206). Thereafter, the response generation server 30 reflects the received advertisement information and provides a voice service (step S207). Specifically, the response generation server 30 outputs an advertisement of advertisement information received from the bid device 200 to the user at an appropriate timing.

[4. (Modification)
The dialog agent system according to the second embodiment described above may be implemented in various different forms other than the second embodiment. Therefore, in the following, an embodiment in which the dialog agent system according to the second embodiment is modified will be described.

[4-1. (Ad output according to keyword)
In the second embodiment, the accepting unit 234 of the bidding apparatus 200 accepts the bid information including the emotion that the user wishes to bid and the advertisement information that he / she wants to output, for example, the fee / a bid price per time. However, the accepting unit 234 may accept a keyword for which an advertisement output is desired in addition to the above emotion and advertisement information. In this case, the keyword corresponding to each bid is included in the bid list information in the example shown in FIG. That is, the reception unit 234 can receive an advertisement bid for each keyword for each emotion.

  Here, a process after the reception unit 234 receives bid information including a keyword for which an advertisement output is desired will be described. First, the acquisition unit 231 acquires information about a user who has a conversation based on the dialog agent system and the determination tree from the response generation server 30. In this case, the information on the user acquired from the response generation server 30 by the acquisition unit 231 includes information on the content of the user's utterance, for example, text information. And the estimation part 232 estimates a user's emotion based on the information regarding the user which the acquisition part 231 acquired. Then, the extraction unit 236 extracts bid information corresponding to emotion information that matches the emotion estimated by the estimation unit 232 from the bid list information. Here, when the keyword of the extracted bid information is included in the information regarding the content of the user's utterance, the transmission unit 235 transmits the advertisement information in the extracted bid information to the response generation server 30. Thereafter, the response generation server 30 outputs an advertisement based on the received advertisement information. Thereby, the advertiser can output an advertisement in a dialogue in which both his / her desired emotion and keyword correspond, and can output an advertisement with high appeal to the user. Moreover, the transmission part 235 may transmit the advertisement information in the extracted bid information to the response generation server 30 when the information about the content of the user's utterance is related to the extracted keyword of the bid information. The information related to the content of the user's utterance here is related to the keyword, for example, when the keyword is “Kima”, it is related to the meaning of the keyword such as “Indian cuisine”, “Curry”, “Lassi”, etc. This means that a meaningful word is included in the information related to the content of the user's utterance.

[4-2. effect〕
As described above, the bidding apparatus 200 according to the second embodiment includes the acquisition unit 231, the estimation unit 232, the presentation unit 233, and the extraction unit 236. The acquisition unit 231 acquires information regarding a user who is having a conversation with the dialog agent system. The estimation unit 232 estimates the emotion of the user when having a conversation with the dialog agent system in the dialog agent system that interacts with the user. The presentation unit 233 presents emotion information indicating the user's emotion estimated by the estimation unit 232 to the advertiser of the advertisement output during the conversation. The extraction unit 236 is based on a comparison result between the emotion of the user estimated based on the information about the user acquired by the acquisition unit 231 and the emotion estimated by the estimation unit 232 regarding the conversation the user is performing. The advertisement output during the conversation is extracted.

  Thereby, the bidding apparatus 200 according to the second embodiment can output an advertisement based on the user's emotion estimated from the information about the user who is using it to the user, and improve the effect of the advertisement. You can make a presentation.

(Third embodiment)
In the third embodiment, an example of providing a bidding service based on a user's emotion estimated from a determination tree and a user's emotion estimated from information on a user who has a conversation based on the dialog agent system and the determination tree explain. Note that the configuration of the interactive agent system according to the third embodiment is the same as the configuration shown in FIG.

[1. Bidding process)
First, the bidding process according to the third embodiment will be described using the example shown in FIG. As shown in FIG. 14, the bid apparatus 300 acquires a determination tree from the response generation server 30 (step S31). In this case, the bidding apparatus 300 may acquire the determination tree from the response generation server 30 every time the determination tree is updated, for example. Subsequently, based on the acquired determination tree, the bid device 300 estimates the user's emotion at each node in the determination tree. Then, the bidding device 300 presents a determination tree including emotion information estimated as the emotion of the user who uses the dialog agent system (step S32). In the example shown in FIG. 14, the bidding apparatus 300 surrounds an area in which a user's emotion is estimated to be positive and an area determined to be negative in the determination tree with dotted lines, together with characters representing the emotion. Present. At this time, the bidding apparatus 300 may present the entire determination tree to the advertiser, or may present a part of the determination tree including emotion information. In the case of the example in FIG. 14, the bidding apparatus 300 presents a determination tree so that an area indicating that the user's emotion is positive and an area indicating that the user's emotion is negative can be identified.

  Subsequently, the bid device 300 receives bid information from the advertiser (step S33). At this time, the accepted bid information may include information on an operation node related to an advertisement, information on a bid price, and the like, which are new rules to be added.

  When the user terminal 10 is talking to the dialog agent system (step S34), the response generation server 30 is based on the information about the node used in the dialog at the time of acquisition, the dialog agent system, and the determination tree. Information regarding the user who has a conversation is transmitted to the bid device 300 (step S35). Here, the information regarding the user includes voice information acquired from the user's utterance. Thereafter, the bid device 300 estimates the user's emotion based on the information about the user acquired from the response generation server 30. At this time, the bidding apparatus 300 estimates that, for example, when the user's utterance inflection is small or when the user's utterance is slow, the emotion of the user who has a conversation based on the determination tree with the dialog agent system is negative. May be.

  If the node corresponding to the accepted bid information matches the node used in the conversation at the time of acquisition, the bid device 300 is estimated from the user's emotion estimated from the determination tree and the information about the user. The bid information corresponding to the result of comparing with the emotion is extracted (step S36). When the result of the comparison satisfies a predetermined condition, the bidding apparatus 300 updates the determination tree based on the information of the operation node, which is a new rule related to the advertisement included in the bid information received from the advertiser determined to make a successful bid. Then, the updated determination tree is transmitted to the response generation server 30 (step S37). Here, the predetermined condition may be that the emotion of the user estimated from the determination tree matches the emotion estimated from the information about the user.

  As described above, the bidding apparatus 300 according to the third embodiment can present a determination tree including emotion information to the advertiser. Therefore, the advertiser can add an advertisement in accordance with the flow of conversation suitable for his / her advertisement and the emotion of the user. Thereby, the bid apparatus 300 can accept a bid for improving the effect of the advertisement.

[2. Bidding equipment configuration]
Next, the configuration of the bidding apparatus 300 according to the third embodiment will be described with reference to FIG. FIG. 15 is a diagram illustrating a configuration example of the bidding apparatus 300 according to the embodiment. As shown in FIG. 15, the bid device 300 includes a communication unit 110, a storage unit 320, and a control unit 330. In addition, below, the description which overlaps with the bid apparatus 100 which concerns on 1st Embodiment shown in FIG. 4, and the bid apparatus 200 which concerns on 2nd Embodiment shown in FIG. 12 is abbreviate | omitted.

(Storage unit 320)
The storage unit 320 includes, for example, the determination tree acquired from the response generation server 30, information about the user, emotion information estimated as a user's emotion using the interactive agent system, bid information registered by the advertiser, and the like. Remember.

  As illustrated in FIG. 15, the control unit 330 includes an acquisition unit 331, an estimation unit 232, a presentation unit 333, a reception unit 334, a transmission unit 335, and an extraction unit 336. Information processing described below Realize or execute the functions and operations of Note that the internal configuration of the control unit 330 is not limited to the configuration illustrated in FIG. 15, and may be another configuration as long as information processing described later is performed. Further, the connection relationship between the processing units included in the control unit 330 is not limited to the connection relationship illustrated in FIG. 15, and may be another connection relationship.

(Acquisition unit 331)
The acquisition unit 331 acquires from the response generation server 30 a determination tree that is information relating to the rules of the dialog that the response generation server 30 uses for dialog with the user. In this case, the acquisition unit 331 may acquire the determination tree from the response generation server 30 every time the determination tree is updated, for example. Further, the acquisition unit 331 acquires from the response generation server 30 information related to the node used in the conversation at the time of acquisition and information related to the user who has a conversation based on the dialog agent system and the determination tree. Here, the acquisition unit 331 obtains information on the node used in the dialog at the time of acquisition and information on the user who has a conversation based on the dialog agent system and the determination tree. You may acquire from the response production | generation server 30 whenever it changes.

(Estimation unit 332)
The estimation unit 332 estimates the emotion of the user when having a conversation with the dialog agent system in the dialog agent system that interacts with the user. For example, the estimation unit 332 estimates the user's emotion at each node in the determination tree acquired from the response generation server 30 by the acquisition unit 331. Further, the estimation unit 332 estimates the user's emotion based on the information about the user acquired from the response generation server 30 by the acquisition unit 331. That is, the estimation unit 332 estimates the user's emotion at each node in the determination tree and the user's emotion based on the information about the user.

(Presentation part 333)
The presentation unit 333 presents emotion information indicating the user's emotion estimated by the estimation unit to the advertiser of the advertisement that is output during the conversation. For example, as shown in FIG. 14, the presentation unit 333 presents to the advertiser terminal 60 a determination tree that includes emotion information estimated as the user's emotion at each node in the determination tree. The presentation unit 333 has the same function as the presentation unit 133 illustrated in FIG.

  Although illustration is omitted here, the presentation unit 333 presents a bid page for registering bid information to the advertiser terminal 60. The advertiser performs a bid by inputting information such as a new node to be added and a bid price using the bid page.

(Reception unit 334)
The accepting unit 334 receives bid information from the advertiser terminal 60 via a predetermined bid page. The reception unit 334 has the same function as the reception unit 134 illustrated in FIG.

(Transmitter 335)
The transmission unit 335 transmits the determination tree updated by the extraction unit 336 described below to the response generation server 30. And the response production | generation server 30 which received the determination tree from the transmission part 335 determines the response message returned to a user with respect to the input message which is a user's utterance based on the received determination tree.

(Extractor 336)
Based on the comparison result between the user's emotion estimated based on the information about the user acquired by the acquisition unit 331 and the emotion estimated by the estimation unit 332 for the conversation the user is performing, the extraction unit 336 Extract advertisements that are output to. In other words, in the present embodiment, the extraction unit 336, when the node for which the advertiser has bid for the addition of a new operation node matches the used node acquired by the acquisition unit 331, the bid for which the advertiser has bid. Extract information. Then, the extracting unit 336 compares the user's emotion estimated from the determination tree with the emotion estimated from the information related to the user. When the result of the comparison between the user's emotion estimated from the determination tree and the emotion estimated from the information related to the user satisfies a predetermined condition, the extraction unit 336 is included in the bid information received from the advertiser determined to be successful The decision tree is updated based on the information of the operation node related to the advertisement, which is a new rule. Then, the transmission unit 335 transmits the updated determination tree to the response generation server 30.

  Here, the predetermined condition regarding the result of comparison between the emotion of the user estimated from the determination tree and the emotion estimated from the information about the user may be that both emotions are similar. In addition, the extraction unit 336 executes the determination tree update process when the used node acquired by the acquisition unit 331 is close to the node for which the advertiser has bid for the addition of a new operation node. May be. Moreover, it is good also as a case where both emotions have the opposite relationship as a predetermined condition in a result for comparison of both emotions. For example, when the emotion of the user estimated from the determination tree is positive and the emotion estimated from the information about the user is negative, the dialogue agent system may output an advertisement or the like regarding the hospital to the user. In this case, the bid information received by the receiving unit 334 from the advertiser may include information on the predetermined condition.

[3. (Bid processing procedure)
Next, with reference to FIG. 16, a procedure of bidding processing by the interactive agent system according to the third embodiment will be described. FIG. 16 is a sequence diagram showing a bidding process procedure by the dialog agent system according to the third embodiment.

  As illustrated in FIG. 16, the bid device 300 obtains from the response generation server 30 a determination tree that is information relating to the rules of the dialog that the response generation server 30 uses for the dialog with the user (step S301). In this case, the bid device 300 stores the acquired determination tree in the storage unit 320. Then, the bidding apparatus 300 estimates the user's emotion at each node in the determination tree based on the acquired determination tree (step S302). Thereafter, the bidding device 300 presents a determination tree including emotion information estimated as the emotion of the user who uses the dialog agent system (step S303).

  Subsequently, the bid apparatus 300 receives bid information including a new node from the advertiser terminal 60 via a predetermined bid page (step S304).

  When the user terminal 10 is talking to the dialog agent system (step S305), the response generation server 30 is based on the information about the node used in the dialog at the time of acquisition, the dialog agent system, and the determination tree. Information related to the user having the conversation is transmitted to the bid device 300 (step S306). Here, the information regarding the user includes voice information acquired from the user's utterance. Thereafter, the bid device 300 estimates the user's emotion based on the information about the user acquired from the response generation server 30. When the node on which the advertiser terminal 60 bids for the addition of a new operation node matches the node used by the user terminal 10, the bidding device 300 determines the user's emotion estimated from the determination tree, and the user Is compared with the emotion estimated from the information related to (step S307).

  When the result of the comparison between the user's emotion estimated from the determination tree and the emotion estimated from the information related to the user satisfies a predetermined condition, for example, both emotions match, the bidding device 300 determines the advertiser whose bid has been determined. The decision tree is generated based on the information of the operation node related to the advertisement, which is a new rule included in the bid information received from (Step S308), and the determination tree including the new rule is transmitted to the response generation server 30 (Step S308). S309). Thereafter, the response generation server 30 determines a response message to be returned to the user in response to the input message that is the user's utterance based on the determination tree including the new node received from the bid device 300, and provides a voice service. (Step S310).

[4. (Modification)
The dialog agent system according to the third embodiment described above may be implemented in various different forms other than the third embodiment. Therefore, a modified embodiment of the dialog agent system according to the third embodiment will be described.

[4-1. (Advertising based on keywords)
In the third embodiment, the bid device 300 updates the determination tree and transmits it to the response generation server 30. However, the bid device 300 may transmit the advertisement information to the response generation server 30 each time. In this case, the bidding apparatus 300 may accept an emotion that a bid is desired, a keyword that is desired to output an advertisement, and advertisement information that is desired to be output, for example, bid information including a bid / a bid price per time (FIG. 11). reference). That is, the bid device 300 accepts an advertisement bid for each keyword for each emotion.

  Here, a process after the accepting unit 334 of the bid device 300 accepts a keyword for which an advertisement output is desired will be described. First, when the user is conversing with the dialog agent system, the acquisition unit 131 acquires, from the response generation server 30, information related to the node used in the dialog at the time of acquisition and information related to the user. In this case, the information regarding the user includes information regarding the content of the user's utterance, for example, text information.

  Here, the extraction unit 336 satisfies the predetermined result of the comparison between the emotion estimated at the node used in the acquired dialogue and the emotion estimated from the information about the user, for example, both emotions match. If so, the following processing is performed. The extraction unit 336 extracts bid information corresponding to emotion information that matches emotion estimated from information about the user, from among the bid information received by the reception unit 334. Here, when the keyword of the bid information corresponding to the extracted emotion is included in the information regarding the content of the user's utterance, the transmission unit 335 sends the advertisement information corresponding to the extracted emotion to the response generation server 30. Send to. Thereafter, the response generation server 30 outputs an advertisement based on the received advertisement information. Thereby, the advertiser can output an advertisement in a dialogue in which both his / her desired emotion and keyword correspond, and can output an advertisement with high appeal to the user. In addition, since the bid device 300 estimates the user's emotion using both the emotion estimated in the node used in the conversation at the time of acquisition and the emotion estimated from the information about the user, The user's emotion can be estimated with high accuracy. Further, the transmission unit 335 may transmit the advertisement information corresponding to the issued emotion to the response generation server 30 when the information about the content of the user's utterance is related to the keyword of the bid information corresponding to the extracted emotion. Good. As used herein, “related to a keyword” includes words having a meaning related to the meaning of the keyword, such as “Indian cuisine”, “Curry”, and “Lassi” when the keyword is “Kima”. It means being. The keyword of the bid information may be associated with a word used for detection at the detection node. In this case, the bidding apparatus 300 performs the above processing when the keyword of the bid information and the word used for detection in the detection node used in the dialog at the time of acquisition satisfy a predetermined condition, for example, match or relate to each other. May be performed.

[4-2. effect〕
As described above, the bidding apparatus 300 according to the third embodiment includes the acquisition unit 331, the estimation unit 332, the presentation unit 333, and the extraction unit 336. The acquisition unit 331 acquires information about a user who is having a conversation with the dialog agent system. The estimation unit 332 estimates the emotion of the user when having a conversation with the dialog agent system in the dialog agent system that interacts with the user. The presentation unit 333 presents emotion information indicating the user's emotion estimated by the estimation unit 332 to the advertiser of the advertisement output during the conversation. The extraction unit 336 is based on a comparison result between the user's emotion estimated based on the information about the user acquired by the acquisition unit 331 and the emotion estimated by the estimation unit 332 for the conversation the user is performing. The advertisement output during the conversation is extracted.

  Thereby, the bid apparatus 300 according to the third embodiment can output the advertisement to the user in accordance with the conversation flow suitable for the advertisement and the user's emotion, and improve the effect of the advertisement. Can be presented.

(About the first to third embodiments)
[1. Judgment information)
In the first to third embodiments, a determination tree as shown in FIG. 2 is used as the determination information. However, the determination information is not limited to a tree shape such as a determination tree, and may be, for example, a node map shape. When such node-map determination information is used, the root node (corresponding to N101 in the example shown in FIG. 2) is the node at the start of the conversation.

[2. User emotions)
In the above-described embodiments 1 to 3, the bidding devices 100 to 300 (hereinafter simply referred to as bidding devices) have the user's emotion “positive” or “negative” based on the content of the determination tree. I estimated. However, the embodiment is not limited to this. That is, the bidding device can estimate an arbitrary emotion of the user and present the estimated emotion.

  For example, the bidding device may estimate the user's emotion based on whether it is “comfortable” or “uncomfortable”. In addition, the bidding apparatus may set an arbitrary emotion such as “joy”, “anger”, “sadness”, “fun”, or the like as an estimation target. In addition, the user's emotion changes as the conversation continues. Therefore, the bidding apparatus may score each emotion to be estimated for each node along the predicted conversation flow, and may estimate the user's emotion at each node based on the scoring. . In addition, the bidding device estimates the user's emotion by indicating the user's emotion as a vector having an arbitrary number of dimensions, associating each dimension with the emotion to be estimated, and performing a vector operation for each node. May be. In this way, the bidding device may estimate the user's emotions while deciding (that is, in a top-down manner) or may be estimated according to the flow of conversation (ie, in a bottom-up manner).

  Specifically, the bidding device may preset an evaluation function corresponding to each emotion. In this case, for example, the bidding device may employ a normal distribution function or the like as each evaluation function, or may employ an arbitrary distribution based on a predetermined emotion model. The emotions estimated by the bid device include, for example, “joy”, “acceptance”, “anger”, “disgust”, “fear”, “surprise”, “expectation”, “sadness”, etc. The bidding device is not limited to the enumerated emotions and may estimate any classification of emotions as long as they are human emotions.

  In addition, the bidding device may digitize each “emotion” using each evaluation function. For example, the bid device may calculate a numerical value of pleasantness / discomfort corresponding to each emotion from the response registered in the response node. In this case, the bidding device may employ, for example, a morphological analysis of response text or a vectorization technique of words. In addition, the bidding device makes corrections based on the correspondence between the responding node and the detection node connected, the relationship with each node existing within a predetermined number of hops, the history of pleasant and unpleasant values so far, etc. You may go.

  In addition, the bidding apparatus may determine the current “emotion” using a weighting function. For example, the bid device may integrate numerical values of pleasantness and unpleasantness using a predetermined formula. In this case, for example, the bidding device may use not only the sum of numerical values of pleasant and unpleasant but also weights for emotions such as pleasant and unpleasant. In addition, the bidding device may use a forgetting curve in which emotions fade with time. In this case, for example, the bidding device may gradually reduce the weighting for each emotion such as pleasant discomfort in the weighting function over time. Further, the bidding device may estimate the user's emotion (for example, pleasant and uncomfortable) based on the integration result of the pleasant and unpleasant numerical values. In addition, the bid device may display the numerical value of each emotion as it is, or may present the emotional pattern of the user by outputting the numerical value of each emotion in association with the color. In this case, the bidding device may present a part with a high numerical value of each emotion in red and a part with a low numerical value in blue as in a heat map, for example. In this case, the bidding device may present, for example, character information of emotions such as “joy” and “sadness” in each colored part. In addition, the bidding device, for example, displays red if the emotion is “joy”, increases the saturation of the portion with a high numerical value, shows low saturation in the portion with a low numerical value, and blue if the emotion is “sad” As an alternative, the saturation at a high numerical value may be increased and the saturation at a low numerical value may be presented low.

[3. Application example)
Moreover, although the said 1-3 embodiment demonstrated based on the form which utilizes a dialog agent system using the user terminal 10, it is not restricted to the said 1-3 embodiment, For example, a user is the user terminal 10 The conversation agent system may be used by having a conversation with a robot having the same function as the above.

[4. program〕
The bid device according to each embodiment described above is realized by a computer 1000 configured as shown in FIG. 17, for example. FIG. 17 is a hardware configuration diagram illustrating an example of a computer 1000 that implements the functions of each bidding device. The computer 1000 includes a CPU 1100, RAM 1200, ROM 1300, HDD 1400, communication interface (I / F) 1500, input / output interface (I / F) 1600, and media interface (I / F) 1700.

  The CPU 1100 operates based on a program stored in the ROM 1300 or the HDD 1400 and controls each unit. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started up, a program depending on the hardware of the computer 1000, and the like.

  The HDD 1400 stores programs executed by the CPU 1100, data used by the programs, and the like. The communication interface 1500 receives data from other devices via the communication network 70 and sends the data to the CPU 1100, and transmits the data generated by the CPU 1100 to other devices via the communication network 70.

  The CPU 1100 controls an output device such as a display and a printer and an input device such as a keyboard and a mouse via the input / output interface 1600. The CPU 1100 acquires data from the input device via the input / output interface 1600. In addition, the CPU 1100 outputs the generated data to the output device via the input / output interface 1600.

  The media interface 1700 reads a program or data stored in the recording medium 1800 and provides it to the CPU 1100 via the RAM 1200. The CPU 1100 loads such a program from the recording medium 1800 onto the RAM 1200 via the media interface 1700, and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase change rewritable disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. Etc.

  For example, when the computer 1000 functions as the bidding apparatus 100 according to the first embodiment, the CPU 1100 of the computer 1000 implements the function of the control unit 130 by executing a program loaded on the RAM 1200. Further, the HDD 1400 stores data in the storage unit 120 in the bidding apparatus according to the first embodiment. The CPU 1100 of the computer 1000 reads these programs from the recording medium 1800 and executes them, but as another example, these programs may be acquired from other devices via the communication network 70.

[5. Others]
In addition, among the processes described in the above embodiments, all or a part of the processes described as being automatically performed can be manually performed, or the processes described as being manually performed All or a part of the above can be automatically performed by a known method. In addition, the processing procedures, specific names, and information including various data and parameters shown in the document and drawings can be arbitrarily changed unless otherwise specified. For example, the various types of information illustrated in each drawing is not limited to the illustrated information.

  Further, each component of each illustrated apparatus is functionally conceptual, and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution / integration of each device is not limited to that shown in the figure, and all or a part thereof may be functionally or physically distributed or arbitrarily distributed in arbitrary units according to various loads or usage conditions. Can be integrated and configured. For example, the acquisition unit 131 and the reception unit 134 illustrated in FIG. 4 may be integrated.

  Moreover, each embodiment mentioned above can be combined suitably in the range which does not contradict a process content.

  As described above, some of the embodiments of the present application have been described in detail with reference to the drawings. However, these are merely examples, and various modifications based on the knowledge of those skilled in the art, including the aspects described in the disclosure line of the invention. It is possible to implement the present invention in other forms with improvements.

  In addition, the “section (module, unit)” described above can be read as “means” or “circuit”. For example, the acquisition unit can be read as acquisition means or an acquisition circuit.

DESCRIPTION OF SYMBOLS 1 Dialogue agent system 10 User terminal 20 Speech recognition server 30 Response generation server 40 Speech synthesis server 50 Service API server 60 Advertiser terminal 100 Bid device 120 Storage part 131 Acquisition part 132 Estimation part 133 Presentation part 134 Reception part 135 Transmission part

Claims (13)

In an interactive agent system that interacts with a user, an estimation unit that estimates an emotion of the user when talking with the interactive agent system;
A presentation unit that presents emotion information indicating a user's emotion estimated by the estimation unit to an advertiser of an advertisement output during the conversation;
A presentation device comprising: A receiving unit that receives an advertisement corresponding to the emotion information presented by the presenting unit from the advertiser;
The presentation device according to claim 1, further comprising: The presenting unit
3. The presentation device according to claim 1, wherein, in the conversation with the dialog agent system, a region of conversation in which the user is estimated to have a predetermined emotion is presented to the advertiser together with the emotion information. . The presenting unit
Presenting to the advertiser an area that is presumed to be positive for the user's emotion and other areas,
The presentation device according to claim 3. The presenting unit
Presenting the advertiser with an area that is presumably negative for the user's emotion and other areas,
The presentation device according to claim 3 or 4, wherein The presenting unit
By clustering areas where the user is presumed to be a predetermined emotion, an area corresponding to the predetermined emotion is presented.
The presentation device according to any one of claims 3 to 5, wherein The presenting unit
According to the advertiser's business type, the conversation area where the user is presumed to have a predetermined emotion is changed.
The presentation device according to any one of claims 3 to 6, wherein The presenting unit
Along with the area, present a bid price for outputting an advertisement during a conversation included in the area.
The presenting device according to any one of claims 3 to 7, wherein The presenting unit
Presenting a bid price calculated according to the value of the conversation included in the area;
The presentation device according to claim 8. The presenting unit
Presenting a bid price calculated according to the emotion of the user corresponding to the region,
The presentation device according to claim 8 or 9, wherein An acquisition unit for acquiring information on a user who is talking to the dialog agent system;
Based on the comparison result between the emotion of the user estimated based on the information about the user acquired by the acquisition unit and the emotion estimated by the estimation unit for the conversation being performed by the user, output during the conversation An extractor for extracting advertisements to be generated,
The presentation device according to claim 1, further comprising: A computer-implemented bidding method,
In a dialog agent system that interacts with a user, an estimation step for estimating a user's emotion when talking with the dialog agent system;
Presenting step of presenting emotion information indicating the user's emotion estimated in the estimation step to an advertiser of the advertisement output during the conversation;
A presentation method characterized by comprising: In a dialogue agent system that interacts with a user, an estimation procedure for estimating a user's emotion when talking with the dialogue agent system;
A presentation procedure for presenting emotion information indicating a user's emotion estimated in the estimation procedure to an advertiser of an advertisement output during the conversation;
A program for causing a computer to execute.

Images