JP2012022453A - Interactive apparatus and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To extract interactive service data allowing a plurality of users to cooperatively perform decision making.SOLUTION: An interactive apparatus 100 for extracting interactive data with users includes: a personal data management unit 51 which manages private data for every user extracted from an interaction with a user; a reference management unit 60 which controls a permission/non-permission to refer the personal data on another user in the interaction with an object user; a domain knowledge management unit 30 which manages ontology as an interaction object candidate with the user; and an interaction control unit 11 which generates interactive data with the object user based on the ontology managed by the domain knowledge management unit 30, the personal data on the object user managed by the personal data management unit 51, and the personal data on another user permitted by the reference management unit 60.

Description

  The present invention relates to an interactive device and a program for generating data for interaction with a search target user by using personal data of other users that can be referred to.

  Conventionally, a dialogue control system in which a user makes some decision through dialogue with the system has been proposed. In the dialogue control system, for example, the user's needs and values are extracted by gradually delving into questions while repeating the dialogue with the target user, and the user's needs can be extracted from a wide variety of services and content information. Extract what matches your needs and values.

  For example, in the spoken dialogue control system of Patent Document 1, a feature word that is a search key is extracted from a user's utterance (that is, input speech), and the feature word is associated with items and sub-items associated with the search target database. The database is searched based on the correspondence relationship, and the dialogue content suitable for the feature word is extracted and presented to the user.

Japanese Patent Laid-Open No. 2001-100807

  However, in the conventional dialogue control system, it is assumed that the user interacts with the system alone, and the opinions and intentions of a specific third party are not automatically reflected during the dialogue. . On the other hand, when a certain user makes a decision, it is often desirable to reflect the opinions of a specific third party. For example, there are cases where it is better to refer to the opinions of the family when acquiring a house for a couple or deciding to change jobs. Therefore, it is often preferable to use the dialogue control system while reflecting the opinions of a specific third party, because a dialogue in accordance with the actual situation of the user is established.

  With respect to the above problems, the object of the present invention is that a plurality of users can make a decision in cooperation, or another user can make a decision based on a certain user's opinion (dialog history). Is to propose an interactive apparatus and program.

  In order to solve the above-described problem, according to an aspect of the present invention, there is provided an interactive device that generates data for interaction with a user, and manages personal data for each user extracted from a user's utterance. A domain management unit that manages permission / non-permission of referring to personal data related to other users in a dialogue with the target user, a domain knowledge management unit that manages an ontology as a candidate for a dialogue with the user, and the domain knowledge For interaction with the target user based on the ontology managed by the management unit, the personal data of the target user managed by the personal data management unit, and the personal data of other users permitted by the reference management unit There is provided a dialogue apparatus comprising a dialogue control unit for generating data.

  According to this, it is possible to refer to related personal data of other users through an interaction with the target user who is the current user. At that time, by using only the personal data of other authorized users, it is possible to ensure the security (security) of the information when incorporating the personal data of the other users into the dialogue with the target user.

  The personal data management unit may store information acquired through interaction with the target user in the personal information database as personal data for each user.

  The information acquired by the dialogue with the target user may be an analysis result of a user answer, a user intention, and information related to the user.

  The reference management unit stores a reference permission item and a reference disapproval item of personal data related to one or more other users related to the target user in a reference condition database, and the dialog control unit is stored in the reference condition database. Only the contents of personal data of other users with respect to the items permitted to be referenced may be reflected based on the reference permitted items and the reference not permitted items of the personal data.

  Each of the personal information database and the reference condition database may be provided inside the dialogue apparatus or outside the dialogue apparatus.

  When there are a plurality of other users who are permitted to refer to personal data, the dialogue control unit determines whether the dialogue data is in accordance with the similarity calculated according to a given importance among the personal data of the plurality of other users. Personal data of other users to be reflected in the generation may be selected.

  The interactive device displays a question message for the target user on a user-side terminal according to the ontology, receives a user's answer to the displayed question message, and receives a user's answer based on an analysis result of the received user's answer You may use for the dialog control system which advances the dialog according to the intention of this.

  In order to solve the above problem, according to another aspect of the present invention, there is provided a program for causing a computer to implement a function of an interactive device that generates data for interaction with a user, from a user's utterance. Personal data management processing for managing the extracted personal data for each user, reference management processing for managing permission / non-permission of reference to personal data related to other users in the dialog with the target user, and candidates for interaction with the user Domain knowledge management process for managing the ontology of the user, the ontology managed by the domain knowledge management unit, the personal data of the target user managed by the personal data management unit, and other users permitted by the reference management unit A dialogue control process for generating dialogue data with the target user based on personal data of Program for executing is provided.

  As described above, according to the present invention, a dialog in which a plurality of users can make a decision in cooperation or another user can make a decision based on a certain user's opinion (dialog history). Type service can be provided.

1 is an overall configuration diagram of a dialog control system and a dialog device according to first and second embodiments of the present invention. FIG. It is the flowchart which showed operation | movement of the dialogue apparatus which concerns on 1st Embodiment. It is an example of the block diagram which shows the structure of property definition. FIG. 4 is a configuration diagram in which a part of the property definition shown in FIG. 3 is expanded in a network format. It is an example of the block diagram which expanded the extended personal data in the network format. It is an example of the reference condition data stored in the reference condition database. It is a figure for demonstrating designation | designated of the reference permission / non-permission item of a high-order node, and the state of a low-order node. It is a flowchart which shows operation | movement of the reference control part which concerns on 1st Embodiment. It is an example of the extended personal data when another user gives the reference permission shown to D621 of FIG. 6 with respect to the target user. It is an example of the property definition using reference information. It is an example of the matching table which concerns on 1st Embodiment. It is a flowchart which shows the flow of the matching process which concerns on 1st Embodiment. It is an example of the matching table for demonstrating similarity calculation processing. It is a figure for demonstrating similarity calculation processing. It is a figure for demonstrating similarity calculation processing. The configuration example of extended personal data referring to a plurality of user data according to the second embodiment is described in a network format. It is an example of the matching table which concerns on 2nd Embodiment.

  Hereinafter, a dialog control system according to an embodiment of the present invention and a dialog device used in the system will be described with reference to the accompanying drawings. In addition, in this specification and drawing, about the component which has the substantially same function structure, duplication description is abbreviate | omitted by attaching | subjecting the same code | symbol.

  In the dialog control system described below, a question sentence is generated based on domain knowledge (ontology) and presented to the user, the answer from the user is interpreted at a semantic level, and information and usage extracted from the user are used. The data for interaction with the user is generated by matching the analysis result interpreted from the user's answer with the search target data. Furthermore, by using a multidimensional ontology for domain knowledge, it is possible to support personalization, such as responses that give a close feeling to the user's remarks, switching of the flow of dialogue, changes in importance according to changes in the user's situation, etc. Realize.

<First Embodiment>
[overall structure]
First, the overall configuration of a dialog control system (interactive analysis / search system) including the dialog device according to the first embodiment of the present invention will be described with reference to FIG. In the dialogue control system Sys according to the first embodiment, the user interacts with the dialogue apparatus 100 using, for example, the user's PC 90 (Personal Computer). Actually, the dialog is established by displaying the dialog data transmitted from the dialog device 100 on the browser of the PC 90 via the Web server 70 in response to the user's request. The interactive device 100 is a device that generates data for interaction with a target user in order to provide an interactive service to the user.

  The dialogue apparatus 100 includes a dialogue management unit 10, an intention analysis unit 20, a domain knowledge management unit 30, a matching unit 40, a matching target analysis unit 50, and a reference management unit 60. The interactive device 100 is connected to a data providing server 80 arranged outside the interactive device 100. The dialogue apparatus 100 extracts the target data that can be provided to the target user during the dialogue from the provided data held by the data providing server 80, and stores it in the storage area (target information database 54) of the matching target analysis unit 50.

  The user starts the dialog control system Sys by accessing the Web server 70 using the browser 90. The user interface can be widely applied to a PC having a communication function, a portable terminal, a dedicated terminal, or the like, but may be provided with a voice synthesizing / recognizing unit that converts a voice uttered by the user into a text. Good.

  The dialog management unit 10 manages the progress of a dialog with a user who wants to search, and includes a dialog control unit 11 that controls the progress of the dialog. The dialogue control unit 11 receives the ontology from the domain knowledge management unit 30 and creates a question sentence (also referred to as a system utterance sentence) related to the progress of the dialogue based on the ontology. That is, the domain knowledge management unit 30 has an ontology for progressing the dialogue, and the dialogue control unit 11 presents a question sentence to the target user according to the progress of the dialogue based on the predetermined ontology.

  The dialogue control unit 11 gives an answer sentence (also referred to as a user utterance sentence) from the user to the question sentence to the intention analysis unit 20 to analyze the contents of the user utterance sentence. When the dialogue control unit 11 receives the user utterance sentence, the dialogue control unit 11 instructs the personal data management unit 51 to store the user utterance sentence as personal data of the user in the personal information database 53 of the matching target analysis unit 50.

  Further, the dialogue control unit 11 instructs the matching unit 20 to match the personal data with the target data. Various timings for instructing the matching are set, for example, when a matching instruction is received from the user, when an answer sentence for each question sentence is obtained, or when all question items are completed. be able to. Further, the dialogue control unit 11 instructs the confirmation of the presence / absence of a user (another user) who allows the target user to refer to data based on the reference condition database 62 of the reference management unit 60. The personal data to be matched includes personal data of the target user and personal data of other users permitted by the reference management unit 60. In addition, the dialogue control unit 11 generates data for dialogue with the target user based on the ontology managed by the domain knowledge management unit, the personal data of the target user, and the personal data of other users permitted to be referred to. . Here, the dialogue data is a sentence that is generated from the ontology and later becomes an utterance sentence of the system. The dialogue control unit 11 generates dialogue data, which is displayed to the user and becomes a system utterance sentence. Therefore, as the data, the dialog data and the system utterance are the same, and the data generated by the dialog control unit 11 is called dialog data. When the dialog data is displayed on the user's screen, it is called a system utterance.

  The intention analysis unit 20 receives a user utterance sentence answered by the user in response to the question sentence acquired from the dialogue control unit 11, and analyzes the contents of the user utterance sentence. In addition, the intention analysis unit 20 gives the analyzed intention analysis result to the dialogue control unit 11 in an ontology format. Specifically, the intention analysis unit 20 includes an intention analysis unit 21 and an intention analysis dictionary 22.

  The intention analysis unit 21 refers to the intention analysis dictionary 22 and parses the user utterance sentence with reference to the morpheme analysis unit 211 that performs morphological analysis on the user utterance sentence and the intention analysis dictionary 22. And a syntax analysis unit 212.

  The intention analysis unit 21 obtains a class defined in the property definition of the domain knowledge database (domain knowledge DB) 32 by performing morphological analysis and syntax analysis on the user utterance sentence, and domain knowledge management Part 30 is given. The property definition will be described later.

  The intention analysis dictionary 22 is a dictionary group for analyzing the contents of a user utterance sentence. For example, a morpheme dictionary (for example, a Japanese morpheme dictionary), a syntax dictionary (for example, a Japanese syntax dictionary), a property definition, and a class This applies to domain knowledge dictionaries that store definitions.

  The domain knowledge management unit 30 manages domain knowledge using the ontology as knowledge. The domain knowledge management unit 30 includes an ontology management unit 31 and a domain knowledge DB 32. The domain knowledge DB 32 stores property definitions and class definitions, and various information (referred to as option information) associated with the classes and properties.

  The domain knowledge management unit 30 receives the intention analysis result from the intention analysis unit 20, refers to the ontology of the domain knowledge DB 32, and returns personal data corresponding to the intention analysis result to the dialogue control unit 11.

  The matching unit 40 performs matching between the personal data managed by the personal data management unit 51 and the target data managed by the target data management unit 52, and matches the conditions desired by the user.

  The matching unit 40 gives the ontology received from the dialogue control unit 11 to the matcher 42, and gives the matched information to the dialogue control unit 11 to the matcher 42, and personal data stored in the personal information DB 53. And a matcher 42 that performs matching processing with the target data stored in the target information DB 54. As a result, the matching unit 40 matches the target data corresponding to the target user's personal data corresponding to the ontology acquired by the domain knowledge management unit 30, and the predetermined other user related to the target user for the acquired ontology. Extract target data that matches personal data. The predetermined other users will be described later.

  The matching target analysis unit 50 includes a personal data management unit 51, a target data management unit 52, a personal information database (personal information DB) 53, and a target information database (target information DB) 54. The beta bases 53 and 54 of the matching target analysis unit 50 store personal data and target data to be matched. Each beta base 53, 54 may store personal data and target data expanded in a form that facilitates matching.

  The personal information DB 53 holds a user data path indicating a storage location (storage position) of personal data and a user data class name indicating a value of personal data. The personal information DB 53 is managed by the personal data management unit 51. The personal data management unit 51 stores the personal data obtained from the user's speech under the control of the dialogue control unit 11 in the personal information DB 53 for each user. The information acquired from the utterance of the target user includes, for example, analysis results of user responses, user intentions, and information related to the users.

  The target data management unit 52 acquires target data from the provided data stored in the data providing server 80 provided outside and stores the target data in the target information DB 54. The target data acquired by the target data management unit 52 is accumulated in the target information DB 54. The target data is, for example, job information data.

  The reference management unit 60 manages permission / non-permission of referring to personal data regarding other users in the dialog with the target user. The reference management unit 60 includes a reference control unit 61 and reference condition data 62. The reference condition database (reference condition DB) 62 stores information on whether reference to personal data regarding other users is permitted or not permitted in the dialog with the target user. The reference control unit 61 refers to personal data of a certain user (user A; other user) based on the contents of the reference condition DB 62 when the system interacts with another user (user B; target user). If possible, the user B can refer to the personal data of the user A.

  In this way, the dialogue apparatus 100 according to the present embodiment displays a question message for the target user on the user-side terminal according to the ontology, receives a user's answer to the displayed question message, and is received This is used in a dialog control system Sys that advances a dialog according to the user's intention based on the analysis result of the user's answer.

  A command to each unit of the interactive apparatus 100 is executed by a dedicated control device or a CPU (not shown) that executes a program. Each operation of the interactive device 100 described below can realize each function of the interactive device 100 by the CPU reading and executing the program.

  Next, after briefly describing the property definition with reference to FIGS. 3 and 4, the operation of the interactive apparatus 100 according to the present embodiment using the property definition will be described with reference to the flowchart of FIG.

(Property definition)
FIG. 3 is a configuration diagram showing the configuration of the property definition, and FIG. 4 is a diagram showing a part of the property definition in a network format. The property definition shown in FIG. 3 describes, for example, the property relationship of each class with “human” as a vertex. In addition, it describes option information related to dialogue strategies and system utterances that each property has, and defines what questions the system asks in what order.

  In FIG. 3, the property definition has items of a domain, a property, a value range, and option information. In addition, the option information is provided with an importance indicating the degree of determining whether the question is an essential question and a priority indicating the degree of determining the order of questions (that is, the flow of dialogue).

  Further, the basic question sentence generated by the dialog control system Sys is defined in the option information. Note that the option information can define various functions for the system to smoothly proceed with the conversation, such as a connection sentence generated before the question sentence and various received sentences generated based on the intention analysis result.

  The option information describes the type of property, for example, a group and an acquisition target. Although the case of two types of properties has been illustrated here, the types of properties are not limited to this, and one type or a plurality of types can be set.

  When the property type is “acquisition target”, the domain knowledge management unit 30 gives the basic question sentence set in the item of the basic question sentence to the dialogue control unit 11, and the dialogue control unit 11 regards this as the system utterance sentence. To the user side through the Web server 70.

  On the other hand, when the property type is “group”, the domain knowledge management unit 30 searches and selects a property having the range class in the domain. At this time, when there are a plurality of properties having this range class as the domain, the domain knowledge management unit 30 looks at the priority of the option information and selects the property with the highest priority. If the selected property is “acquisition target”, the domain knowledge management unit 30 gives the basic question sentence to the dialogue control unit 11, and the dialogue control unit 11 transmits this to the user side as a system utterance sentence. To do. As described above, when the property type is a group, it is possible to realize the transition to the property having the range class in the domain by using the priority.

  FIG. 4 is a configuration diagram in which a part of the property definition shown in FIG. 3 is expanded in a network format. In FIG. 4, each circled node indicates a class (that is, a domain and a value range in FIG. 3), and a line connecting between the classes indicates a property. Information enclosed by a square on the line indicates various option information.

  For example, in FIG. 4, when “human” is the apex and the property is “personality”, the value range is “personality”, and when the property is “a chance to change jobs”, the value range is “change jobs” The reason is “reason”, and the value range when the property is “work experience” is “current work”. Further, for example, when “personality” is defined as a domain, as in the above, for example, a value range when a property is “site visit purpose” has a relationship such as “site visit purpose”.

[Operation of interactive device]
Now, the overall flow of dialogue control will be described below using the flowchart of FIG. First, a user who uses the dialog control system Sys accesses the designated URL using the browser 90 and activates the dialog control system Sys through the Web server 70. When the dialog control system Sys is activated, the dialog control unit 11 refers to the property definition stored in the domain knowledge DB 32 through the ontology management unit 31, and has the highest priority among the properties having the current pointer in the domain. Select.

  When the system is started, the definition area to be selected as the initial value is set in the property definition. For example, a pointer is set for the first domain domain class “human” in FIG. Accordingly, the domain knowledge management unit 30 looks at the range class “personality” in the case of setting the property “personality” of the domain class “human” pointed to by the initial pointer. At this time, since the type of the property of the range class “personality” is “group”, the domain knowledge management unit 30 has the domain range having the highest priority among the properties having the range class “personality” as the domain. Select "Personality"-Property "Site Visit Purpose"-Range Class "Site Visit Purpose" Property.

  Then, the domain knowledge management unit 30 gives the basic question sentence of this property “Please tell us the purpose of coming to this site” to the dialogue control unit 11, and the dialogue control unit 11 sends this basic question sentence to the system utterance. To the user side. In this way, a question from the dialogue control system Sys is started (step S101).

  The user who has received this system utterance sentence replies, for example, “I want to know what my job is suitable”. This answer sentence is transmitted to the dialogue control unit 11 through the Web server 70 (step S102).

  When the dialog control unit 11 receives the user utterance, the dialogue control unit 11 gives the user utterance to the intention analysis unit 20. In the intention analysis unit 20, the intention analysis unit 21 uses the intention analysis dictionary 22 to perform morphological analysis and syntax analysis on the user utterance sentence. As a result, for example, the intention of the user “I want to know the right job” is obtained. Analyzing and giving the intention analysis result to the dialogue control unit 11 (step S103).

  Upon receiving the intention analysis result from the intention analysis unit 20, the dialogue control unit 11 stores the information, that is, personal data (including extended personal data) in the personal information DB 53 of FIG. 1 for each user. The dialogue control unit 11 inquires of the domain knowledge management unit 30 about the next question and determines the next question (step S104).

  At this time, the domain knowledge management unit 30 refers to the property definition and sets the domain class “personality” —the property “current job type” —the range class “job type”, which has the next highest priority after the range class “site visit”. Then, the basic question sentence “What is your current job type?” Is given to the dialogue control unit 11, and the dialogue control unit 11 determines this as a system utterance and transmits it to the user side (step S105). . In this way, the dialogue between the dialogue control system Sys and the user is realized by repeatedly presenting the next question sentence and receiving the answer sentence from the user. In step S105, when the next question is not determined, that is, when all the questions are finished, the system operation is finished. The above is the basic dialog control operation.

  The intention analysis result acquired by the intention analysis is accumulated in the extended personal data by the dialogue control. FIG. 5 is an example of a configuration diagram in which the extended personal data stored in the personal information DB 53 is expanded in a network format.

  In FIG. 5, the property “personality” is a group property and there is no information to be acquired. In FIG. 5, this state is indicated by shading. For the properties "site visit purpose", "job type", and "life event" beyond that, the information obtained from user utterances "I want to know the right job", "General office work", "Child was born" is stored in each range . This indicates that these values have been acquired from the user response through interaction with the system.

  In addition, the property “customer”, “experienced business category”, “affiliated company size”, and “client industry” ahead of the property “current job” have a blank node in the value range. This indicates that it is a storage location of a value range that is not acquired at the present time but may be acquired in a future dialogue.

(Reference method of other users)
Next, a case will be described in which reference data of another user related to the target user during the dialogue is used during the dialogue. FIG. 6 is an example of reference condition data stored in the reference condition database 62. The reference condition data in FIG. 6 includes a user name, a reference permitted user name, a reference permission item, and a reference disapproval item. The user name is a place for storing information identifying the user, such as the user name and user ID. The reference-permitted user name is a place for storing information on a user who is permitted to refer to all or a part of personal data such as extended personal data for the user specified by the user name. The reference permission item is a place for designating which item of the extended personal data is permitted to be referred to. The reference non-permission item is a place for designating which part of the extended personal data is not permitted to be referred to. The designation of the reference permission item and the reference non-permission item may be performed by individually specifying an item name, or information specified by the upper node may be copied to the lower node.

  FIG. 7 explains what happens to the state of the lower node by designating the reference permission item and the reference non-permission item to the upper node. FIG. 7A shows the state of the extended personal data in the case where “top” “human” is designated as the reference permission item and the reference non-permission item is not designated, that is, in the case of D621 shown in FIG. Show. The information of the top node is copied to all nodes below “human”, and the reference is permitted.

  FIG. 7B shows the case where the top node “human” is designated as a reference permission item, and “stage of job change activity” and “life event” are designated as reference non-permission items, that is, in the case of D622 in FIG. Indicates the status of extended personal data. With the exception of “Stage of change of job activity” and “Life event” specified as non-permitted items for reference, each item under “Human” is permitted for reference.

  In FIG. 7B, classes are not defined under “Stage for change of job activity” and “Life event”, but when defined, “Stage of job change” and “Life event” are defined. As with, it cannot be referenced.

  As described above, the reference condition DB 62 stores the reference permission items and the reference disapproval items of personal data related to one or more other users related to the target user, and the dialog control unit 11 is stored in the reference condition DB 62. Based on the personal data reference permission item and the reference non-permission item, only the content of the personal data of other users for the reference permitted item is reflected in the dialogue data. The target user refers to a user who is interacting (a user with a user name in FIG. 6), and the other user is a user other than the target user related to the target user (another user: a reference-permitted user name in FIG. 6). User).

  Next, an operation when the reference data of another user is used during the dialogue between the target user and the dialogue control system Sys will be described with reference to FIG. FIG. 8 is a flowchart showing the operation of the reference management unit 60. When a system start question is made in S101 of FIG. 2 and the user inputs information (ID or the like) for identifying an individual (S801), the reference control unit 61 refers to the reference condition DB 62 according to an instruction from the dialogue control unit 11. Then, it is confirmed whether or not there is a user who is permitted to refer to the extended personal data for the target user during the conversation (S802).

  If there is no reference-permitted user, this process ends. On the other hand, if there is a reference-permitted user in S802, the reference control unit 61 checks whether the reference-permitted item of the reference-permitted user is specified (S803), and the extended personal stored in the personal information DB 53 for the reference-permitted user. Of the data, the reference permission item and its subordinate items are copied to the buffer (S804). If there is no reference permission item in S803, the process returns to S802.

  Next, the reference control unit 61 refers to the reference disapproval item in the reference condition DB 62 (S805), and if there is no reference disapproval item, stores the contents of the buffer in the personal information DB 53 of the interactive user (S807). If there is a reference non-permitted item, the lower items of the reference non-permitted item and the non-referenceable item are deleted from the buffer (S806), and the contents of the deleted buffer are stored in the personal information DB 53 of the interactive user (S807). If there are other reference-permitted users, the process returns to S802, and the processes of S802 to S807 are repeated. If there is no other reference-permitted user, this process ends. The operation of the reference management unit 60 has been described above.

  FIG. 9 is an example of the extended personal data when user B (another user related to the target user) gives the reference permission indicated by D621 in FIG. The shaded portion is user information of user B. The referenced user information may be handled separately from other user information, for example, by making it impossible to overwrite.

  Next, the flow of processing when a question (dialogue) using information of a reference-permitted user is performed will be described with reference to FIG. FIG. 10 is an example of property definition using reference information. As optional information of domain class “personality” -property “life event” -range class “life event”, in addition to priority D322, importance D323, basic question text D324, conditional question text condition D325, conditional A question sentence D326 is defined.

  Here, the conditional question sentence condition D325 indicates a condition for the state of the personal data at the time when the question is made, and the conditional question sentence D326 indicates a question sentence when the conditional question sentence condition is satisfied. The basic question sentence is a default question sentence when none of the conditions is satisfied, and a question sentence used only when the condition is satisfied is a conditional question sentence.

  When the conditional question sentence condition D325 is satisfied, the dialogue control unit 11 generates the conditional question sentence D326 as an output sentence of the dialogue control system Sys. On the other hand, if the conditional question sentence condition D325 is not satisfied, the basic question sentence D324 is generated as an output sentence of the dialog control system Sys. In addition, the conditional question sentence condition D325 and the character string enclosed in {} in the conditional question sentence D326 are obtained by connecting the domain-property relationship with “/”. It shall represent. For example, {user A / reference user data} indicates “user B” (D321) in FIG.

  In S104 of FIG. 2, it is assumed that the ontology management unit 31 refers to the property definition of FIG. Here, if the conditional question sentence condition D325 is satisfied, that is, if the life event ahead of the user data to be referred to is not empty, according to the conditional question sentence “user B”, “a child was born. “That is,” is given to the dialogue control unit 11 (“” is a character string quoted from the reference destination). On the other hand, if the conditional question sentence condition D325 is not satisfied, that is, if the life event ahead of the user data to be referred to is empty, the basic question sentence "Please tell me about your current situation." The dialogue control unit 11 determines this as a system utterance and transmits it to the user side (S105).

  As described above, when the dialogue control unit 11 advances the dialogue with the target user according to the given ontology managed by the domain knowledge management unit 30, the content of the personal data of other users who are the reference-permitted users By reflecting, the question text can be changed.

(Search using personal data of other users who are authorized users)
Next, a flow of processing for performing a search using information of a reference-permitted user will be described. FIG. 11 is an example of a matching table that defines which information in the personal information DB 53 is matched with which information in the target information DB 54. The matching table is stored in the domain knowledge DB 322 as part of the ontology, and is referenced from the matching unit 40 via the domain knowledge management unit 30. FIG. 11A shows the matching table in a table format, and FIG. 11B shows the matching table in a network format.

  The matching table in FIGS. 11A and 11B includes “personal data”, “importance”, and “target data” as items. The item “personal data” indicates a property (condition) in the extended personal data. The item “target data” indicates a property (condition) in the extension target data. The item “importance” indicates a weight for a property in matching. In addition, “/ person / personality / current job type” in the personal data and “/ recruitment / experience job type” in the target data correspond to each other, and the importance is “0.5”. The property name differs depending on the relationship between the personal data and the target data. This is because the condition name (property name) of the job offer data (data) and the condition name (property name) extracted from the corresponding user are Because it is different. In addition, a plurality of properties of personal data items can correspond to the same target data item and vice versa.

  FIG. 12 is a flowchart showing the flow of the matching process. First, when a matching request is output from the dialogue control unit 11 to the matching unit 40 (S1201), the matching unit 40 reads the matching table in the domain knowledge DB 32 via the domain knowledge management unit 30 (S1202), and the matching table. The value of the item of personal data defined in is acquired from the personal information DB 53 (S1203). Next, the target information DB 54 is read one by one (S1204), and the data value stored in the personal information DB 53 acquired in S1203 is collated with the data value stored in the corresponding target information DB 54 according to the definition of the matching table. Then, the similarity is calculated (S1206). When the similarity calculation processing for all the target information DBs 54 is completed, the matching unit 40 rearranges the target data in descending order of similarity and outputs this as a matching result (S1207).

  Next, taking the search for real estate properties as an example, the similarity calculation processing will be specifically described with reference to FIGS. 13 and 14. The matching unit 40 receives the matching table shown in FIG. 13 and calculates the degree of similarity between the personal data in FIG. 14A and the target data in FIG. First, according to the matching table of FIG. 13, the values of “desired floor plan”, “desired price”, and “desired place” are acquired from the expanded personal data in FIG. 14A stored in the personal information DB 53. Next, from the target data in FIG. 14B stored in the target information DB 54, the values of “room layout”, “property price”, and “location of the property” are acquired, and the values of the corresponding extended personal data and Collate. “Property 1” is “Property layout = 4LDK”, “Property price = 28 million yen”, and “Property location = Saitama city”, so it does not match User A ’s “Scheduled layout = 3LDK”. , "Desired price = 30 million yen or less" and "desired place = Saitama City".

Here, the formula for calculating the similarity is defined as follows.
[Formula 1]

  In this case, the similarity between “User A” and “Property 1” is “(0 × 1.0) + (1 × 1.0) + (1 × 1) as shown in FIG. .0) = 2.0 ". As shown in FIG. 15A, when the reference user data is present together with the extended personal data and the information of the reference user data is defined in the user data item of the matching table as in D3284 to D3286 of FIG. For the reference user data “desired floor plan = 4LDK”, “desired price = up to 30 million yen”, and “desired place = Tokyo”, the same degree of similarity is calculated as the user's own information (D3281-D3283). . That is, the similarity between the desired condition of “User A” in FIG. 15A and “Property 1” in FIG. 14B is “(0 × 1) + (1 × 1) + (1 × 1)”. + (1 × 0.5) + (1 × 0.5) + (0 × 0.5) = 3.0 ”. The above is the description of the matching process when reference user data exists.

As described above, according to the dialog control system Sys according to the first embodiment, the following effects (1) to (3) can be obtained.
(1) It is possible to refer to personal data (for example, a conversation history) of another user who is a specific third party in a dialog with a target user who is the current user.
(2) The system utterance sentence shown to the target user can be changed by personal data of other users.
(3) By using only the personal data of other authorized users, it is possible to solve the security problem when incorporating the personal data of the other users into the dialogue with the target user.

  Specifically, when the user A who is the target user during the conversation interacts with the dialog control system Sys, the information of the user B who is another user can be used. As a result, the following dialog and search can be realized. "User B says" I want to know about the job change market ", but please tell me the purpose of your visit to this site." ("" Is the quotation of information that user B is allowed to see) In this way, personalized question answering and information provision can be provided to the user.

  Further, according to the dialogue control system Sys according to the first embodiment, when the situation is estimated from the relationship between the user A and the user B, the question content can be simplified or omitted. For example, if user A and user B are a couple, and the user information of user B is the property “life event”-“you have a child”, the question is “Tell me about your current situation” As a result, the user A is not asked to confirm that the child is born, or the user A is not asked a question. The value of the property “life event” of the user A is “ It becomes possible to store “children made”.

  Furthermore, according to the dialogue control system Sys according to the first embodiment, it is possible to perform a search in consideration of information of another user other than the target user. Specifically, when searching for the job change destination of the target user A, the search of the other user B (for example, the wife of the user A) is also performed in the search condition, or the user A and the user B each desired condition It is possible to search for and compare job change destinations that satisfy.

<Second Embodiment>
Next, a dialogue control system Sys using the dialogue device according to the second embodiment of the present invention will be described. In the dialogue control system Sys according to the second embodiment, the configuration of extended personal data when there are a plurality of reference-permitted users will be described. Here, a case where a plurality of reference-permitted users, that is, user B (user A's wife) and user C (parent of user A) exist in user A is taken as an example.

  However, in addition to this, as an example of referring to information of a plurality of reference users, reference is made to information of a user set that makes a decision together with a user, such as a family, or a user set having the same specific attribute such as occupation and age There are cases.

  FIG. 16 shows a configuration example of extended personal data referring to a plurality of user data in a network format. Under the top node “human (user A)” and property “reference user data”, information of a plurality of reference-permitted users (user B / user C) is set. At this time, the property name connecting the reference user, the user B, and the user C may be a name indicating that it is another reference user data such as “reference user data 1” and “reference user data 2”. However, as shown in FIG. 16, the name may indicate the attribute of each user. In addition to the method of FIG. 16, a method of describing information of a plurality of reference-permitted users (that is, other users who can be referred to) directly connected to the top node “human” is conceivable.

  When matching processing is performed with reference to information on a plurality of user data, as shown in FIG. 17, the user data path and the corresponding target data items and importance levels are described in the matching table. Just do it.

  Also in the dialog control system Sys according to the second embodiment, the dialog control unit 11 is calculated according to a given importance stored in the matching table among the personal data of a plurality of other users permitted to refer to the personal data. Depending on the similarity, any other user's personal data to be reflected in the generation of the dialog data is selected and used to generate the dialog data.

  According to the dialog control system Sys according to the second embodiment, it is possible to perform a search in consideration of information of a plurality of other users different from the target user, and the dialog with the target user is performed using the information of the plurality of other users. Can be used. For example, generation of dialogue data considering the desired location of the “parent” who is the higher-priority user C shown in FIG. 17 among a plurality of other users, or displaying search results This makes it possible to provide personalized question responses and information to the target user.

  For example, taking FIG. 16 as an example, if there are two wives and parents as reference users, a desired location included in reference user data (user B: wife) with low importance (that is, low similarity). For the dialogue with the target user (user A) in consideration of the desired location “Saitama City” included in the reference user data (user C: parent) with higher importance (ie, higher similarity) than “Tokyo” As data, for example, “Are you sure you want to get a home in“ Saitama City ”?” Is generated. In addition, as a search result, target data corresponding to “Saitama City”, which is the reference user C's wish, is displayed as a search result higher than “Tokyo”, which is the request of the reference user B. Further, the target data corresponding to “Saitama City”, which is the request of the reference user C, is displayed as a higher-order search result than “Tokyo”, which is the request of the reference user B.

  Also in the dialogue control system Sys according to the second embodiment, the question content can be simplified or omitted when the situation is estimated from the relationship between the target user and a plurality of other users. For example, if user A and user B are a couple, and user A and user C are parents and children, as user information of user B, a confirmation question is given to user A, “Your child was born”. Or “You are hoping to live with your parents”. As a result, it is possible to make the target user feel close and to reduce the load of the target user's answer and to establish a smooth conversation.

  In the first and second embodiments, the operations of the respective units are related to each other, and can be replaced as a series of operations and a series of processes in consideration of the relationship between each other. Thereby, the embodiment of the interactive device can be an embodiment of a program for causing a computer to realize the functions of the interactive device.

  In addition, this is a program for causing a computer to realize the function of an interactive device that generates data for interaction with a user, and manages personal data for each user extracted from the user's speech A reference management process for managing permission / non-permission of reference to personal data regarding other users in a dialog with the target user, a domain knowledge management process for managing an ontology as a candidate for a dialog with the user, and the domain knowledge management Data for interaction with the target user based on the ontology managed by the personal data management unit, the personal data of the target user managed by the personal data management unit, and the personal data of other users permitted by the reference management unit And a program for causing a computer to execute interactive control processing for generating Can.

  The preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to such examples. It is obvious that a person having ordinary knowledge in the technical field to which the present invention pertains can come up with various changes or modifications within the scope of the technical idea described in the claims. Of course, it is understood that these also belong to the technical scope of the present invention.

  For example, in the first and second embodiments, the dialog device 100 is applied to the dialog control system Sys. However, not only a dialog control system that extracts user needs and opinions by dialog, but also an interactive analysis / search system and dialog. It can be applied to a type document creation support system. In other words, the dialogue apparatus 100 extracts the user's needs and values by repeating the dialogue with the target user and gradually digging up the questions, and the user's needs from a wide variety of services and content information. It can be applied to an information search / analysis system that searches for things that match the values.

  Each of the personal information DB 53, the target information DB 54, and the reference condition DB 62 may be provided inside the interactive device 100 according to the present invention, or may be provided outside the interactive device 100.

  In the first and second embodiments, the example in which information of one or a plurality of users permitted to refer to a certain user is used in the user's dialogue has been described. Not limited to this, when a plurality of users interact with the system at the same time, it is possible to refer to each other's information and to use it for interaction and search.

  In the first and second embodiments, an item “reference permitted user name” is set in the reference condition data, and the user information is permitted to be referred to individual users described in the reference permitted user name. However, the present invention is not limited to this, and a condition based on information given to each user may be described, and reference may be permitted to a user who matches the condition. For example, it is possible to allow reference to a user within a certain degree of kinship, a user with the same group ID, a user with a specific item (for example, a job type, etc.), etc., according to the relationship information.

  Further, in the first and second embodiments, whether or not reference is possible is determined based on information described in the reference condition data. However, the present invention is not limited to this, and identity verification is performed through general authentication processing. Further, a means for obtaining the approval of the reference-permitted user in real time may be taken.

  In the first and second embodiments, the personal information DB 53, the target information DB 54, the domain knowledge DB 32, and the reference condition DB 62 that store user information to be referenced are stored inside the dialogue apparatus 100. However, the present invention is not limited to this, and each of the above DBs may exist outside the dialog control system, and each information may be acquired by communicating with the dialog control system via a network. Good. Further, the user information of the extended personal data may also be external information (such as various service usage histories and behavior histories) other than the information acquired from the dialogue.

  In the first and second embodiments, the description has been given of the mode in which the information of the user permitted to be referenced is copied and used in the extended personal data of the user who is currently interacting. However, the present invention is not limited to this. May be directly referred to, and information on items for which reference is permitted may be used.

DESCRIPTION OF SYMBOLS 10 Dialogue management part 11 Dialogue control part 20 Intention analysis part 21 Intention analysis part 22 Intent analysis dictionary 30 Domain knowledge management part 31 Ontology management part 32 Domain knowledge DB
40 Matching Unit 41 Matching Manager 42 Matcher 50 Matching Target Analysis Unit 51 Personal Data Management Unit 52 Target Data Management Unit 53 Personal Information DB
54 Target information DB
60 Reference Management Unit 61 Reference Control Unit 62 Reference Condition DB
70 Web Server 80 Data Providing Server 90 Browser 100 Dialogue Device Sys Dialogue Control System

Claims (8)

An interactive device that generates data for interaction with a user,
A personal data management unit for managing personal data for each user extracted from the user's utterance;
A reference management unit that manages permission / non-permission of referring to personal data related to other users in a dialogue with the target user;
A domain knowledge management unit that manages the ontology as a candidate for interaction with the user;
Based on the ontology managed by the domain knowledge management unit, personal data of the target user managed by the personal data management unit, and personal data of other users permitted by the reference management unit, A dialogue control unit for generating the dialogue data of
A dialogue apparatus comprising:   The interactive apparatus according to claim 1, wherein the personal data management unit stores information acquired by the utterance of the target user in a personal information database as personal data for each user.   The dialogue apparatus according to claim 2, wherein the information acquired through the dialogue with the target user is a user answer analysis result, a user intention, and information related to the user. The reference management unit stores a reference permission item and a reference disapproval item of personal data related to one or more other users related to the target user in a reference condition database,
The interactive control unit reflects only the contents of personal data of other users with respect to an item for which reference is permitted based on a reference permission item and a reference disapproval item of personal data stored in the reference condition database. An interactive device according to any one of the above.   The interactive device according to claim 4, wherein each of the personal information database and the reference condition database is provided in the interactive device or outside the interactive device.   When there are a plurality of other users who are permitted to refer to personal data, the dialogue control unit determines whether the dialogue data is in accordance with the similarity calculated according to a given importance among the personal data of the plurality of other users. 6. The interactive apparatus according to claim 1, wherein personal data of any other user reflected in generation is selected.   The interactive device displays a question message for the target user on a user-side terminal according to the ontology, receives a user's answer to the displayed question message, and receives a user's answer based on an analysis result of the received user's answer The dialogue apparatus according to any one of claims 1 to 6, which is used in a dialogue control system that advances a dialogue according to a user's intention. A program for causing a computer to realize a function of a dialog device that generates data for dialog with a user,
Personal data management processing for managing personal data for each user extracted from the user's utterance;
A reference management process for managing permission / non-permission of referring to personal data regarding other users in the dialogue with the target user;
Domain knowledge management processing for managing the ontology as a candidate for interaction with the user;
Based on the ontology managed by the domain knowledge management unit, personal data of the target user managed by the personal data management unit, and personal data of other users permitted by the reference management unit, Dialog control processing for generating data for dialogue,
A program that causes a computer to execute.

Images