JP6960579B2 - Information processing methods, information processing devices and programs - Google Patents

Description

The present disclosure relates to information processing methods, information processing devices and programs, and more particularly to information processing methods, information processing devices and programs for interacting with users to perform information processing.

Conventionally, a dialogue system that interacts with a user by a device such as a computer is known.

For example, Patent Document 1 discloses a technique in which a chatbot and a user interact with each other on a chat application. Here, a chatbot is a technology in which a system plays the role of at least one of the participants on a platform for a chat application initially provided for humans to exchange messages.

Japanese Unexamined Patent Publication No. 2013-117938

However, in the technology disclosed in Patent Document 1, since the chatbot message and the human message are presented in a unified manner, the question to be answered by the user is buried in the chatbot message. There is also. In such a case, the chatbot cannot completely obtain the conditions related to the information that the user requests to search from the user.

The present disclosure has been made in view of the above circumstances, and an object of the present disclosure is to provide an information processing method, an information processing device, and a program capable of acquiring all the conditions related to the information requested by the user through dialogue with the user. do.

In order to solve the above problem, the information processing method and the like according to one aspect of the present disclosure are for information requested by the user to be searched by repeating the display of the question and the acquisition of the answer from the user to the question. An information processing method for advancing the narrowing down, in which a step of displaying a question regarding the search target in the first aspect on a display provided in the terminal and an answer to the question input by the user through the terminal are displayed. In the step of displaying the question, including the step of acquiring, the first information regarding the progress of the narrowing down is displayed in the second aspect at the upper part of the question, and the answer guidance to the question is displayed at the lower part of the question. The second information is displayed in the third aspect.

It should be noted that these general or specific embodiments may be realized in a device, system, method, integrated circuit, computer program or recording medium, and any of the device, system, method, integrated circuit, computer program and recording medium. It may be realized by various combinations.

According to the information processing method and the like of the present disclosure, it is possible to obtain all the conditions related to the information requested by the user through dialogue with the user.

It is a schematic block diagram of the information processing system which concerns on embodiment. It is a figure which shows the display mode of the information processing apparatus which concerns on embodiment. It is a figure which shows the detailed structure of the dialogue control part and a database which concerns on embodiment. It is a flowchart which shows an example of the processing of the information processing apparatus which concerns on embodiment. It is a figure which shows an example of the dialogue model which concerns on embodiment. It is a figure which shows an example of the display screen which concerns on the comparative example. It is a figure which shows an example of the display screen which concerns on embodiment. It is a figure which shows the display screen which concerns on the modification of embodiment. It is a figure which shows an example of the content data in Example 1 of the Embodiment. It is a figure which shows an example of the sentence generation data which concerns on Example 1 of Embodiment. It is a figure which shows an example of the display screen which concerns on Example 1 of Embodiment. It is a figure which shows an example of the content data in Example 2 of Embodiment. It is a figure which shows an example of the sentence generation data which concerns on Example 2 of Embodiment. It is a figure which shows an example of the display screen which concerns on Example 2 of Embodiment 2. It is a figure which shows an example of the content data in Example 3 of Embodiment. It is a figure which shows an example of the sentence generation data which concerns on Example 3 of Embodiment. It is a figure which shows an example of the display screen which concerns on Example 3 of Embodiment. It is a figure which shows an example of the content data which concerns on Example 4 of Embodiment. It is a figure which shows an example of the sentence generation data which concerns on Example 4 of Embodiment. It is a figure which shows an example of the display screen which concerns on Example 4 of Embodiment.

The information processing method according to one aspect of the present disclosure is an information processing method in which the information processing requested by the user to be searched is narrowed down by repeating the display of the question and the acquisition of the answer from the user to the question. The present invention includes a step of displaying a question regarding the search target in the first aspect on a display provided in the terminal, and a step of obtaining an answer to the question input by the user through the terminal. In the step of displaying the question, the first information regarding the progress of the narrowing down is displayed in the second aspect at the upper part of the question, and the second information regarding the answer guidance to the question is displayed at the lower part of the question in the third aspect. Display it.

With this configuration, it is possible to present reference dialogue information in a state in which it is easy to distinguish it from the system utterance display, so that it is possible for the user to easily input a response to the system utterance. As a result, it is possible to acquire all the conditions related to the information that the user requests to search by interacting with the user.

Here, for example, the information processing method further causes the display to display information indicating candidates for the search target narrowed down by the progress of the narrowing down when the progress of the narrowing down satisfies a predetermined condition. It may include steps.

For example, the predetermined condition may be that the number of candidates to be searched is equal to or less than a predetermined number.

Further, for example, in the step of displaying the question, in the frame area which is a balloon-shaped, rectangular or rounded rectangular area surrounded by a frame, characters indicating the contents of the question are displayed in a predetermined font name, font size, and the like. The question may be displayed in the first aspect by displaying the text color, the background color, or the specifications specified by the typeface.

Further, for example, in the step of displaying the question, a character indicating the first information is designated in a predetermined font name, font size, text color, background color or typeface in an area outside and above the frame area. The first information may be displayed in the second aspect by displaying the information according to the specifications.

Further, for example, the character specifications indicating the content of the question in the first aspect and the character specifications indicating the first information in the second aspect are the same.

Further, for example, in the step of displaying the question, a character indicating the second information is designated in a predetermined font name, font size, text color, background color or typeface in an area outside and below the frame area. The second information may be displayed in the third aspect by displaying the second information according to the specifications.

Further, for example, the character specifications indicating the content of the question in the first aspect and the character specifications indicating the second information in the third aspect may be the same.

Further, for example, the specifications of the characters indicating the first information in the second aspect and the specifications of the characters indicating the second information in the third aspect may be the same.

Further, for example, the first information may include the number of items to be searched when displaying the question.

Further, for example, the first information may include the remaining number of times when the question is displayed out of the predetermined total number of times the question can be asked.

Further, for example, the first information may include a degree of progress indicating the cumulative number of times when the question is displayed with respect to a predetermined total number of times the question can be asked.

Further, for example, the second information may indicate a response method based on an example response including a character indicating affirmation and a character indicating denial.

Further, for example, the second information may indicate an answer method based on an answer example including words and phrases that can narrow down the search target.

Further, for example, the target of the search may be a plurality of cooking recipes or menus, and the question may indicate a content asking for a request for the plurality of cooking recipes or menus.

Further, for example, the target of the search may be a plurality of travel plans, and the question may indicate the content of asking a request for the plurality of travel plans.

Further, for example, the target of the search may be a plurality of real estate properties, and the question may indicate the content of asking a request for the plurality of real estate properties.

Further, for example, the target of the search may be a plurality of video contents, and the question may indicate a content asking for a request for the plurality of video contents.

Further, for example, in the step of displaying the question, the first question, which is the content of asking the request for the recipe or the menu of the plurality of dishes, is displayed in the first aspect, and is included in the database at the upper part of the first question. The total number of recipes or menus for the plurality of dishes is displayed as the first information, and at the bottom of the first question, one ingredient included in the recipes or menus for the plurality of dishes based on the database. Alternatively, after obtaining the first answer to the first question in the first step of displaying the answer example including the phrase indicating one classification as the second information and the step of obtaining the answer, the recipes of the plurality of dishes or The second question, which is the content of further asking the request for the menu, is displayed in the first aspect, and the recipes of the plurality of dishes included in the database, which are narrowed down by the first answer, are displayed at the top of the second question. The number of menu items is displayed as the first information, and at the bottom of the second question, one ingredient or one included in the recipes or menus of the plurality of dishes included in the database narrowed down by the first answer. The database includes recipes for a plurality of dishes, classification information indicating the classification of each of the recipes for the plurality of dishes, and classification information indicating the classification of each of the recipes of the plurality of dishes. Ingredient information indicating the ingredients used for each of the plurality of cooking recipes may be associated with the ingredient information.

Further, for example, in the step of displaying the question, if a predetermined phrase is included in the second answer to the second question acquired in the step of acquiring the answer, further, one of the items included in the classification information. The third question, which is a content asking whether to request one ingredient included in the classification or the ingredient information, is displayed in the first aspect, and is acquired in the step of acquiring the answer at the upper part of the third question. The number of recipes or menus of the plurality of dishes included in the database, which is narrowed down by the second answer to the second question, is displayed as the first information, and the above one is displayed at the bottom of the third question. The second information may be displayed as an example of an answer including a phrase that affirms and a phrase that denies the classification or the above-mentioned one ingredient.

Further, for example, the plurality of dishes included in the database, which include the affirmative phrase in the third answer to the third question acquired in the acquisition step for the answer and are narrowed down by the third answer. When the number of recipes or menus in the above is less than or equal to the threshold value, the display includes a step of displaying information on the recipes or menus of the plurality of dishes included in the database, which is narrowed down by the third answer. In the step of displaying the question, if the number of recipes or menus of the plurality of dishes included in the database does not fall below the threshold value due to the third answer to the third question acquired in the step of acquiring the answer, the first The fourth question, which is a content asking whether to request one classification different from the above classification or one ingredient different from the one ingredient, is displayed in the first aspect, and the first question is displayed above the fourth question. The number of recipes or menus of the plurality of dishes included in the database, narrowed down by the three answers, is displayed as the first information, and the affirmative phrase and the negative phrase are displayed at the bottom of the third question. The including answer example may be displayed as the second information.

Further, the information processing apparatus according to one aspect of the present disclosure is information processing that advances the narrowing down of the information requested by the user by repeating the display of the question and the acquisition of the answer from the user to the question. An input process for acquiring an answer to the question input by the user through the dialogue control unit, which is a device and displays a question about the search target in the first aspect on a display provided in the terminal, and the terminal. The dialogue control unit is provided with a unit, and the dialogue control unit displays the first information regarding the progress of the narrowing down in the second aspect at the upper part of the question, and at the lower part of the question, the second information regarding the answer guidance to the question. The information is displayed in the third aspect.

With this configuration, it is possible to present reference dialogue information in a state in which it is easy to distinguish it from the system utterance display, so that it is possible for the user to easily input a response to the system utterance. As a result, it is possible to acquire all the conditions related to the information that the user requests to search by interacting with the user.

In addition, the program according to one aspect of the present disclosure provides an information processing method that advances the narrowing down of information requested by the user by repeating the display of the question and the acquisition of the answer from the user to the question. A program to be executed by a computer, in which a step of displaying a question regarding the search target in the first aspect on a display provided in the terminal and an answer to the question input by the user are acquired through the terminal. In the step of executing the step and displaying the question, the first information regarding the progress of the narrowing down is displayed in the second aspect at the upper part of the question, and the answer guidance for the question is displayed at the lower part of the question. 2 Information is displayed in the third aspect.

With this configuration, it is possible to present reference dialogue information in a state in which it is easy to distinguish it from the system utterance display, so that it is possible for the user to easily input a response to the system utterance. As a result, it is possible to acquire all the conditions related to the information that the user requests to search by interacting with the user.

It should be noted that these comprehensive or specific embodiments may be realized in a system, method, integrated circuit, computer program or recording medium such as a computer-readable CD-ROM, system, method, integrated circuit, computer. It may be realized by any combination of a program and a recording medium.

Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. Each of the embodiments described below is a specific example of the present disclosure. The numerical values, shapes, components, steps, order of steps, etc. shown in the following embodiments are examples, and are not intended to limit the present disclosure. Further, among the components in the following embodiments, the components not described in the independent claims indicating the highest level concept are described as arbitrary components. Moreover, in all the embodiments, each content can be combined.

(Embodiment)
Hereinafter, the information processing method and the like in the embodiment will be described with reference to the drawings.

[Information processing system]
FIG. 1 is a schematic configuration diagram of an information processing system according to the present embodiment.

The information processing system shown in FIG. 1 includes an information processing device 100 and a plurality of terminals including at least one of a terminal 200 and a terminal 300. The information processing device 100 and the plurality of terminals are connected via a network 400.

The terminal 200 and the terminal 300 are examples of a plurality of terminals, and may be mobile terminals such as smartphones, tablets, and mobile phones, or terminals such as personal computers. The terminal 200 or the terminal 300 is used by the user and is responsible for input / output when interacting with the information processing apparatus 100. In the example shown in FIG. 1, two terminals are shown, but the number is not limited to two and may be three or more.

The network 400 is a general-purpose network including the Internet, a wireless LAN, and the like, and connects them by relaying communication between the information processing device 100 and the terminal 200 and the like.

The information processing device 100 is realized by a computer including a CPU, an input / output device, a communication device, and the like, and is, for example, a server. The information processing apparatus 100 repeats the display of the question and the acquisition of the answer from the user to the question, thereby proceeding with the narrowing down of the information requested by the user to search.

Hereinafter, the configuration of the information processing device 100 will be described.

[Configuration of Information Processing Device 100]
The information processing device 100 shown in FIG. 1 includes an input processing unit 11, a dialogue control unit 12, and a database 13.

<Input processing unit 11>
The input processing unit 11 acquires an answer to a question input by the user through the terminal 200 or the like. In the present embodiment, the input processing unit 11 is controlled by the dialogue control unit 12, and outputs the understanding result of processing the input of the dialogue by the user and the input of the dialogue by the user itself to the dialogue control unit 12. The expressions included in the answers obtained by the input processing unit 11 differ depending on the user, but expressions having similar meanings are mapped to one label by a generally known language understanding technique and understood as the same content.

The display format of the dialogue by the user may be, for example, not text but voice. Further, when the user answers the closed question described later, in addition to the above, the answer may be a line-of-sight answer or a gesture answer.

<Dialogue control unit 12>
The dialogue control unit 12 has a function of determining an output sentence for dialogue with the user, such as a question to the user, by using the understanding result output by the input processing unit 11 and the database 13. Further, the dialogue control unit 12 not only has a function of determining an output statement, but also manages the number of candidates for information desired by the user, that is, information requested by the user to search, and as an example of the user's response to the determined output statement. It also has a function to manage the information to be presented as management information. The dialogue control unit 12 converts the output sentence into a display format presented to the user and outputs the output sentence. The display format includes, for example, a voice or text display format.

FIG. 2 is a diagram showing a display mode of the information processing device 100 according to the present embodiment.

Further, as shown in FIG. 2, the dialogue control unit 12 displays a question regarding the search target in the first mode on the display screen D11 of the display provided in the terminal 200 or the like, and at the top of the question, the question is displayed. The first information regarding the progress of narrowing down is displayed in the second aspect, and the second information regarding the answer guidance to the question is displayed in the third aspect at the bottom of the question.

Here, displaying the question in the first aspect means that, for example, as shown in the display screen D11 of FIG. 2, a predetermined frame area, which is a rectangular area surrounded by a frame, is designated by a display setting or the like. It is to be displayed in the specifications. In the frame area, which is a balloon-shaped, rectangular or rounded rectangular area surrounded by a frame, characters indicating the content of the question are specified by a predetermined font name, font size, text color, background color or typeface. The question may be displayed in the first aspect by displaying it in the specifications.

Further, the first information may be any information regarding the progress of narrowing down, and may be, for example, the number of search targets when displaying a question. Further, the first information may be, for example, the remaining number of times when the question is displayed out of the predetermined total number of times the question can be asked, or the cumulative number of times when the question is displayed with respect to the predetermined total number of times the question can be asked. The degree of progress indicating the number of times may be used.

Further, displaying the first information in the second mode means, for example, displaying the first information in the dotted square area outside and above the frame area, which is a rectangular area surrounded by the frame of the display screen D11 in FIG. 2, with a predetermined specification. To let. The first information can be displayed by displaying the characters indicating the first information in the outer and upper areas of the frame area with the specifications specified by the predetermined font name, font size, text color, background color, or typeface. It may be displayed in the second aspect. Of course, the character specifications indicating the content of the question in the first aspect and the character specifications indicating the first information in the second aspect may be the same or different.

In addition, the answer guidance is to present an example of an answer or instruct an answer method such as "input with a button". Then, the second information may be any information regarding the answer guidance to the question, and may indicate, for example, an answer method using an answer example including a character indicating affirmation and a character indicating denial. Further, for example, the second information may indicate, for example, an answer method using an answer example including a phrase that can narrow down the search target.

Further, displaying the second information in the third mode means, for example, displaying the second information in a dotted square area outside and below the frame area, which is a rectangular area surrounded by the frame of the display screen D11 in FIG. 2, with a predetermined specification. To let. The second information is displayed outside the frame area and in the lower area with the specifications specified by the predetermined font name, font size, text color, background color, or typeface. May be displayed in the third aspect. Of course, the character specifications indicating the content of the question in the first aspect and the character specifications indicating the second information in the third aspect may be the same or different. Further, the specifications of the characters indicating the first information in the second aspect and the specifications of the characters indicating the second information in the third aspect may be the same or different.

Further, when the progress of the narrowing down satisfies a predetermined condition, the dialogue control unit 12 displays information indicating the candidate of the search target narrowed down by the progress of the narrowing down on the display. Here, the predetermined condition is, for example, that the number of candidates to be searched is equal to or less than a predetermined number.

FIG. 3 is a diagram showing a detailed configuration of the dialogue control unit 12 and the database 13 according to the present embodiment. As shown in FIG. 3, the dialogue control unit 12 includes a dialogue management unit 121, a system utterance display generation unit 122, a dialogue information display generation unit 123, and a display integration unit 124. The dialogue control unit 12 according to the present embodiment refers to the dialogue model 131, the content data 132, and the statement generation data 133 included in the database 13, and cooperates with the input processing unit 11 to obtain information requested by the user to search. To proceed with the narrowing down of.

<< Dialogue Management Department 121 >>
The dialogue management unit 121 updates the management information managed to display the first information and the second information by using the understanding result output by the input processing unit 11 and the database 13, and asks a question to the user. Determine the output statement, or system behavior, for interaction with the user.

More specifically, the dialogue management unit 121 follows the dialogue state transition model such as the state transition diagram stored in the dialogue model 131, and responds to the action of the immediately preceding system and the understanding result output by the input processing unit 11. Then, the current dialogue state is updated, and the action of the system is decided based on it. Here, the information processing device 100 or the dialogue control unit 12 determines the behavior of the system by using keywords related to the type of output sentence and the content of the output sentence, such as "open question" and "closed question". Is to determine.

The management information is information managed by the information processing device 100 or the dialogue control unit 12, and does not have to be in a human-readable format. The management information may include the narrowed down information candidates from the user input history in the task of proceeding with the narrowing down of the information requested by the user from the information included in the content data 132. .. Further, as for the management information, when the dialogue management unit 121 pre-reads the state transition diagram in order to determine the behavior of the system and also acquires the information of the input example assumed next, the information of the input example is used. It may be included. The management information is not limited to these, and may include information indicating the degree of progress, information indicating the upper limit of the number of questions, and / or the number of hits in the task of proceeding with the narrowing down of the information requested by the user to search. ..

<< System utterance display generator 122 >>
Based on the behavior of the system determined by the dialogue management unit 121, the system utterance display generation unit 122 generates a system utterance display that displays a question regarding the search target in the first aspect with reference to the sentence generation data 133.

More specifically, the system utterance display generation unit 122 acquires the system behavior determined by the dialogue management unit 121, refers to the sentence generation data 133, and expresses the system behavior as a concrete natural language expression. It converts into a system utterance such as a question sentence and generates an utterance display to be displayed in the first mode such as enclosing it in a balloon.

<< Dialogue information display generation unit 123 >>
The dialogue information display generation unit 123 generates a dialogue information display in which the management information managed by the dialogue management unit 121 is converted into a format presented to the user.

More specifically, the dialogue information display generation unit 123 refers to the management information managed by the dialogue management unit 121, and simply presents the first information, which is, for example, the number of narrowed-down information candidates, to the user. It was converted into the second aspect which is the display format of the above, and was converted into the third aspect which is the display format for simply presenting the second information which is the information about the input example of the user assumed next to the user. Generate a dialogue information display.

≪Display integration unit 124≫
The display integration unit 124 is generated by the understanding result output by the input processing unit 11, the input itself of the dialogue by the user, the system utterance display generated by the system utterance display generation unit 122, and the dialogue information display generation unit 123. Integrate with dialogue information display.

More specifically, the display integration unit 124 generates a user utterance display in which the input itself of the dialogue by the user output by the input processing unit 11 is surrounded by a balloon. Then, the display integration unit 124 integrates the generated user utterance display, the system utterance display generated by the system utterance display generation unit 122, and the dialogue information display generated by the dialogue information display generation unit 123, that is, the same display screen. Place in to generate an integrated display. As described above, in the display integration unit 124, for example, the first information such as the number of narrowed-down information candidates is arranged at the upper part of the system utterance, and the information of the next assumed user input example is arranged. The second information of is placed at the bottom of the system utterance.

<Database 13>
The database 13 is composed of, for example, a rewritable non-volatile memory such as a hard disk drive or a solid state drive, and stores information regarding a search target. More specifically, the database 13 is referred to for the dialogue management unit 121 to determine the output content of the dialogue and to generate the output statement.

In the present embodiment, as shown in FIG. 3, the database 13 is composed of the dialogue model 131, the content data 132, and the sentence generation data 133.

≪Dialogue model 131≫
The dialogue model 131 stores, for example, a dialogue state transition model represented in the form of a state transition diagram. The dialogue model 131 stores the plurality of dialogue states, the behavior of the system in each of the plurality of dialogue states, and the user's response in association with each other. The dialogue model 131 is associated with the behavior of the system based on the dialogue state according to the behavior of the immediately preceding system and the understanding result output by the input processing unit 11, and is referred to by the dialogue management unit 121.

≪Content data 132≫
The content data 132 is, for example, an example of a knowledge base, for example, a relationship between a plurality of candidates to be searched such as a relationship between a disease and a subjective symptom and element information (keyword) indicating a feature, an attribute, a condition, or the like. It is a database that describes knowledge based on a specific expression format. The content data 132 is stored, for example, in association with a plurality of candidates to be searched, classification information indicating the classification of the plurality of candidates, and element information constituting the plurality of candidates.

≪Sentence generation data 133≫
The sentence generation data 133 is composed of, for example, a list of templates for which a sentence is completed when a keyword is embedded, and conditions for applying the template. More specifically, the sentence generation data 133 stores an open question template, a closed question template, a question template for acquiring knowledge, and a template used for answering a user question, which the dialogue control unit 12 refers to. Has been done.

[Processing of information processing device 100]
Next, the processing of the information processing apparatus 100 configured as described above will be described.

FIG. 4 is a flowchart showing an example of processing of the information processing apparatus according to the present embodiment.

First, the information processing device 100 initializes (S1).

Next, the information processing apparatus 100 updates the dialogue state based on the initial conditions (S2).

Next, the information processing apparatus 100 causes the display provided in the terminal 200 or the like to display a question regarding the search target in the first aspect, and at the upper part of the question, displays the first information regarding the progress of narrowing down in the second aspect. (S3), the question and the like display process including the process of displaying the second information regarding the answer guidance to the question in the third mode at the lower part of the question is performed (S3).

More specifically, in step S3, first, the information processing apparatus 100 generates a system utterance display for displaying, for example, a question regarding a search target in the first aspect (S31). Next, the information processing apparatus 100 generates a dialogue information display that displays the first information in the second aspect and displays the second information in the third aspect (S32). Next, the information processing device 100 integrates the system utterance display generated in step S31 and the dialogue information display generated in step S32, that is, generates and outputs an integrated display for dialogue with a user arranged on the same display screen. (S33).

Next, the information processing apparatus 100 performs an answer acquisition process including a process of acquiring an answer to a question input by the user (S4).

More specifically, in step S4, first, the information processing apparatus 100 acquires an input from the user who responds to the content of the system utterance display included in the integrated display output in step S3, and displays the input. Output the user utterance display (S41). Next, the information processing apparatus 100 understands the input from the user acquired in step S41 (S42).

Next, the information processing apparatus 100 returns to step S2, updates the dialogue state based on the understanding result of the user's input, and rotates the loop of steps S2 to S4 in the same manner as described above to work with the user. Continue the dialogue.

In FIG. 4, it is described that step S32 is performed after step S31, but the present invention is not limited to this, and vice versa. That is, the order of these does not matter.

[Example of dialogue model 131]
FIG. 5 is a diagram showing an example of the dialogue model 131 according to the present embodiment. FIG. 5 shows an interactive state transition model represented in the form of a state transition diagram as an example of the dialogue model 131. In FIG. 5, the dialogue state is represented by a circled state number, and the transition is represented by an arrow. Two events are linked on the arrow, the event closer to the tail of the arrow is linked to the action of the system, and the event closer to the head of the arrow is linked to the action of the user. .. The user's behavior corresponds to the understanding result of the user's input. State 1 indicates that the dialogue state of the information processing device 100 is the initial state, and state 5 indicates that the dialogue state of the information processing device 100 is the final state.

In the state 1, when the information processing device 100 asks an open question and the user provides information, the dialogue state transitions to the state 2.

In the state 2, when the information processing device 100 asks an open question and the user provides information, or when the information processing device 100 asks a knowledge acquisition question and the user provides information, the interactive state is in the state. Return to 2. On the other hand, in the state 2, when the information processing apparatus 100 asks an open question and the user denies it, the dialogue state transitions to the state 3.

In the state 3, when the information processing apparatus 100 asks a closed question and the user affirms or denies, the dialogue state returns to the state 3. Further, when the information processing apparatus 100 asks a closed question and the user asks a user question, the dialogue state shifts to the state 4. The other transition in the state 3 is when the information processing apparatus 100 presents a conclusion, and the dialogue state transitions to the state 5 by any input from the user represented by "*".

When the information processing device 100 gives an answer in the state 4, the dialogue state returns to the state 3 without waiting for the user's input. Here, "Φ" shown in FIG. 5 means a special symbol representing an immediate transition.

In addition, when there is a dialogue state, if the information processing apparatus 100 has one action to be taken, that action may be taken, but if a plurality of actions are set, it depends on the policy. You can select it. In the example shown in FIG. 5, for example, in state 2, there is an option of an open question or a knowledge acquisition question. Regarding this, if the keyword provided immediately before is known to the information processing device 100, an open question is selected, and if it is unknown to the information processing device 100, a knowledge acquisition question is selected. The policy, which is the rule for deciding, may be set in advance. Further, for example, in the state 3, there is an option of a closed question or a conclusion presentation. Regarding this, if the information obtained by the user just before is sufficiently narrowed down, the conclusion presentation is selected, and if not, the closed question is used to obtain information for further narrowing down the information candidates. Such policies may be set in advance. Then, it may be assumed that the policy is stored in the content data 132. That is, when a plurality of actions are set in, for example, the state 2, the information processing device 100 may determine by referring to the policy defined in the content data 132.

[Effects, etc.]
As described above, the information processing device or the like according to the present embodiment makes it easier for the user to input a response to the system utterance by presenting reference dialogue information in a state where it can be easily distinguished from the system utterance display. Therefore, it is possible to obtain all the conditions related to the information that the user requests to search by interacting with the user.

In the above description, it has been explained that the system utterance is displayed as text or the like, but the present invention is not limited to this. When the information processing device or the like according to the present embodiment includes the voice output means, only the text of the system utterance in the balloon may be read aloud.

Further, although the content data 132 has been described as being configured in the database 13 inside the information processing apparatus 100, it may be configured on an external server or the like.

Hereinafter, the effects of the information processing apparatus and the like according to the present embodiment will be described with reference to FIGS. 6 to 8. 6 to 8 show an example in which a task of narrowing down recipes satisfying a plurality of conditions is performed as a task of narrowing down information requested by the user through dialogue between the information processing device 100 and the user. It is shown.

FIG. 6 is a diagram showing an example of the display screen D91 according to the comparative example. In FIG. 6, as a comparative example, a display related to dialogue for four turns is shown as a display screen D91 in which a chatbot message and a human message are presented in a unified manner.

On the display screen D91, the area D911 indicates the utterance area of the system in the first turn. Area D912 shows the utterance area of the user in the second turn. Both have a balloon frame, and the content shown in the utterance area is shown in colloquial terms.

Specifically, the contents of area D911 include the sentence "The number of registered recipes is 1987.", The sentence "What kind of recipe are you looking for?" And "For example," Main dish is good "and" Japanese food. " Please answer like "I like it." The content of the area D912 is the sentence "Western style is good", which is a response to the second sentence of the area D911. In the area D911, the first sentence is an explanatory sentence based on the management information managed by the system, and the second sentence is a question sentence to the user. The third sentence is a supplementary sentence based on management information and is a supplementary sentence to the second sentence. Among these three sentences, the main expression for interactively interacting with the user is the question sentence of the second sentence, and the user is requested to respond to this question sentence. The other two sentences are sub-expressions and are classified as supplementary information. However, in the area D911, the three sentences are lined up without any sharpness, and the question to be answered by the user is buried. That is, it is difficult for the user to know which of the three sentences should be responded to.

Conventionally, there is an application for humans to chat with each other by text, and dialogue between humans and chatbots is also realized on the same platform. In other words, the chatbot remains a platform for text-based verbal communication between humans, and the message of the chatbot, which is the output of the machine, is placed on it. Therefore, it is output to the balloon as well as human input, and as shown in Fig. 6, when the explanation about the information managed by the machine and the supplementary information such as the usage managed by the machine are output, it is the main. Information is buried, and it is rather difficult for the user to answer.

As described above, in the comparative example shown in FIG. 6, the message from the chatbot and the message from the human are presented in a unified manner such as the text in the balloon. Therefore, all the information managed by the system is expressed as the dialogue of the chatbot, and the question to be answered by the user is buried in the long dialogue, which has a problem related to the ease of input by the user. As a result, the chatbot cannot completely obtain the conditions related to the information that the user requests to search from the user.

FIG. 7 is a diagram showing an example of the display screen D11 according to the present embodiment. FIG. 7 shows the display related to the dialogue for four turns as the display screen D11 according to the present embodiment.

On the display screen D11, the area D111 indicates the utterance area of the first turn by the information processing device 100, which is a system, and is surrounded by a balloon-shaped frame. The area D112 indicates the utterance area of the user in the second turn, and is surrounded by a balloon-shaped frame.

The area D1111 is a dialogue information display area arranged above the area D111, and first information regarding the progress of narrowing down such as "recipe registration: 1987" is displayed. The area D1112 is a dialogue information display area arranged in the lower part of the area D111, and displays second information regarding answer guidance for a question such as "input example: main dish is good / Japanese food is good". In FIG. 7, the area D1111 and the area D1112 are surrounded by a dotted line to clearly indicate the area, but they are not necessary for actual display.

When the user is trying to input the answer to the question from the information processing apparatus 100 in the utterance area shown in the area D112, the first information is in the area D111, the upper area D1111 of the area D111, and the second information is in the lower area D1112. Information is displayed. Therefore, it is easy for the user to input the answer "Western style is good" to the question in the utterance area shown in the area D111. Further, it can be easily understood that the first information shown in the area D1111 and the second information shown in the area D1112 are supplementary information for the question in the utterance area shown in the area D111.

Conventionally, auxiliary information such as time information has been presented on a human-to-human chat platform, but it is distinguished from dialogue information, that is, first information and second information, which are presented to promote the progress of dialogue. ..

FIG. 8 is a diagram showing a display screen D12 according to a modified example of the present embodiment. The same elements as those in FIG. 7 are designated by the same reference numerals, and detailed description thereof will be omitted. FIG. 8 shows a display relating to dialogue for four turns as a display screen D12 which is a modification of the display screen D12.

The display screen D12 shown in FIG. 8 differs from the display screen D11 shown in FIG. 7 in the contents shown in the upper part and the lower part of the area D111.

More specifically, the area D1221 is a dialogue information display area arranged above the area D111, and the second information regarding the answer guidance to a question such as "input example: main dish is good / Japanese food is good" is provided. Is displayed. The area D1222 is a dialogue information display area arranged in the lower part of the area D111, and displays the display of the first information regarding the progress of narrowing down such as "recipe registration: 1987". That is, the contents shown in the dialogue information display area shown in FIG. 7 are interchanged up and down.

However, on the display screen D12 according to this modification, it is not easy for the user to input the answer "Western style is good" to the question in the utterance area shown in the area D111, that is, the user thinks that it is difficult to answer. This is because the user wants to refer to the second information regarding the answer guidance to the question such as the input example after the question of the information processing apparatus 100. Further, the content of the first information regarding the progress of narrowing down managed by the information processing device 100 is a reference for the user to answer the question of the information processing device 100, contrary to the causal relationship of determining the content of the system utterance. Because it doesn't become much. As shown in the display screen D12 according to the present modification, displaying the dialogue information of the first information and the second information in the dialogue information display areas at the upper and lower parts of the area D111 simply means the comparison shown in FIG. Compared with displaying the display screen D91 in the example, that is, all three sentences in the utterance area D911, there is an effect of adding sharpness. However, the content that is appropriate to be displayed at the upper part of the area D111 and the content that should be displayed at the lower part of the area D111 are different.

As described above, the dialogue state is changed by the user's utterance, and the content of the question or the like of the information processing apparatus 100 is determined accordingly. Therefore, the content of the first information regarding the progress of narrowing down, which is the information regarding the dialogue state, can be naturally referred to by the user when it is displayed in the dialogue information display area above the area D111 which is the utterance area. On the other hand, in response to the content of the question or the like of the information processing device 100, the second information regarding the answer guidance to the question, which is the information on what kind of user's utterance is assumed by the information processing device 100, is the area D111 which is the utterance area. If it is displayed in the dialogue information display area at the bottom of, it is easier for the user to input answers, etc., and it can be used with peace of mind.

As described above, if management information that changes in response to the user's utterance, such as the number of hits of information, is presented at the upper part of the system utterance balloon, the user can easily understand that the dialogue has progressed. Further, when management information such as an input example when the user inputs such as the input waiting by the information processing device 100 is presented at the lower part of the balloon of the system utterance, the user can embody his / her own input. It can be used as a reference, and it also helps to fill the gap between the input that you have come up with and the input that the information processing device 100 is waiting for.

As described above, the information processing device or the like according to the present embodiment makes it easier for the user to input a response to the system utterance by presenting reference dialogue information in a state where it can be easily distinguished from the system utterance display. Therefore, it is possible to obtain all the conditions related to the information that the user requests to search by interacting with the user.

Hereinafter, a specific embodiment in which the narrowing down of the information requested by the user through the dialogue with the user is advanced will be described as an example.

(Example 1)
In the first embodiment, an example will be described in which a task of narrowing down recipes satisfying a plurality of conditions is performed as a task of proceeding with narrowing down of information requested by the user to search. That is, the case where the search target is a recipe or menu of a plurality of dishes and the question indicates the content of asking a request for the recipe or menu of the plurality of dishes will be described.

[Content data 132]
FIG. 9 is a diagram showing an example of content data 132 according to the first embodiment of the present embodiment.

FIG. 9 shows, as an example of the content data 132, a knowledge base expressing the relationship between dishes and their classification items, the relationship between dishes and ingredients, and the like in the form of a directed graph. Note that FIG. 9 corresponds to an example of a part of a huge knowledge base. In other words, the content data 132 is associated with the recipes of a plurality of dishes, the classification information indicating the classification of each of the recipes of the plurality of dishes, and the ingredient information indicating the ingredients used for each of the recipes of the plurality of dishes. There is.

Generally, in a directed graph expressing knowledge, a circle is called a node and an arrow is called an edge. A triad that connects one node and another node with an edge is called a triple. Concepts are connected to nodes, relationships between two concepts are connected by edges (line segments), and the relationships between concepts represented by triples are the building blocks of knowledge.

In the example shown in FIG. 9, dishes 1 and 2 correspond to a plurality of candidates to be searched, and Western-style, pottery, onions, etc. correspond to element information (keywords) indicating features, attributes, conditions, and the like. Further, FIG. 9 shows the relationship that, for example, dish 1 and dish 2 of node 1 are a kind of Western-style dish and a kind of roasted dish. Further, it is shown that the onions and the dioscorea opposita of the node 2 are used as the ingredients in the dishes 1 and 2 of the node 1. In addition, the relationship between the souvenirs, which are the superordinate concept of onions and dioscorea opposita, as the ingredients of dishes 1 and 2 is shown. In order to show the above relationship, two types of relatives, "IsA" and "HasA", are applied.

Therefore, information that satisfies a plurality of conditions is satisfied by increasing the AND search keywords of a plurality of keywords by using the node2 in [node1 IsA node2] or [node1 HasA node2] as a keyword and the node1 associated with the keyword as a search target. Can be narrowed down.

Further, the elliptical region R11 shown by the dotted line represents the relationship between the food material and its classification item. The area R11 is an example of knowledge used for answering a user question when it is assumed that the user has no knowledge about "earthenware". The area R11 is set as a separate frame because it is not directly referred to for searching for dishes.

[Sentence generation data 133]
FIG. 10 is a diagram showing an example of sentence generation data 133 according to the first embodiment of the present embodiment. FIG. 10 illustrates an open question, a closed question, a knowledge acquisition question, an answer template, and their application conditions referred to by the information processing apparatus 100.

That is, when the information processing device 100 outputs an open question regarding the search target, at the start of the dialogue with the user, the first question "What recipe are you looking for?" Of the open question shown in FIG. 10 is displayed. After that, the second and subsequent open questions shown in FIG. 10 are (repeatedly) generated asking "Are there any other conditions?".

Further, when the information processing apparatus 100 outputs a closed question regarding the search target, whether the relationship between the keyword (node2) specified by the dialogue management unit 121 and the search target (node1) is IsA or HasA. Is a condition, and a closed question is output using the template shown in FIG. In the case of IsA, the information processing device 100 outputs a closed question using the closed question template shown in FIG. 10 "Is (node2) okay?". On the other hand, in the case of HasA, the closed question is output using the closed question template shown in FIG. 10 "Can I use (node2)?".

Further, when the information processing apparatus 100 outputs a knowledge acquisition question, it always outputs a question sentence "What kind of food is it?" And asks the user to exemplify desired information.

Further, when the information processing device 100 outputs an answer to the user question, the user question is a WHAT question related to "X", and the relationship "Y" IsA "X" is a knowledge base from the content data 132. When found, the answer is output using the answer template shown in FIG. 10, "For example, list_up (" Y ")." Here, list_up (“Y”) in the template lists the knowledge of [“Y” IsA “X”] in the condition, which corresponds to “Y” when multiple “X” are found. Up, in Japanese, it means connecting via ",". If the number is large, it may be cut off at an appropriate number.

Further, when the information processing apparatus 100 outputs an answer to the user question, the user question is a WHAT question related to "X", and the relationship "X" IsA "Z" is a knowledge base from the content data 132. When it is found, the answer is output using another answer template, "It is a kind of" Z "." That is, the information processing device 100 generates an answer such as "for example, Y1 or Y2" or "for example, a kind of Z" according to the relationship existing in the content data 132 which is the knowledge base.

Although not shown, a template may be prepared in the sentence generation data 133 in order to generate an input example assumed as user input. For example, as the information providing template, "(keyword) is good", the affirmative answer template "yes", the negative answer template "no", and the like may be stored in the sentence generation data 133.

[Processing of information processing device 100]
FIG. 11 is a diagram showing an example of a display screen according to the first embodiment of the present embodiment.

The display screen D21 shows an example of a dialogue when the information processing device 100 asks an open question, and the display screen D31 shows an example of a dialogue when the information processing device 100 asks a closed question. The display screen D41 shows an example of a screen shown when the information processing apparatus 100 presents a conclusion.

Hereinafter, the processing described with reference to FIGS. 4 and 5 will be described.

In the initial state, that is, the state 1 of FIG. 5, it is assumed that the number of registered recipe information, that is, the number of candidates for a plurality of recipes or menus to be searched is 1987.

<Turn 1 and 2>
First, the information processing device 100 performs a question or the like display process in step S3. Specifically, first, the information processing apparatus 100 refers to the dialogue model 131 of FIG. 5, decides to ask an open question because it is in the initial state, that is, the state 1, and refers to the sentence generation data 133 of FIG. Generate the first open question "What recipe are you looking for?" System speech display. Here, the assumption of the next input shown in the state 1 of FIG. 5 is information provision. Next, the information processing apparatus 100 refers to the content data 132 of FIG. 9, extracts "main dish" and "Japanese style" as keyword candidates for narrowing down the recipe information candidates, and "main dish is good". Generate an interactive information display that displays an input example that says "Japanese style is good".

As a result, on the display screen D21, in the first turn, the question "what kind of recipe are you looking for?" In the area D211 and the first information "recipe registration: 1987" in the area outside and above the area D211. The second information "input example: main dish is good / Japanese food is good" is displayed in the outer and lower area of. In this way, while the question "What kind of recipe are you looking for?" Is expressed in spoken language, the first information "recipe registration: 1987" and the second information "input example: main dish is good / Japanese food is good". "Good" is simply expressed using symbols.

Next, the information processing apparatus 100 performs an answer or the like acquisition process in step S4. Specifically, first, the information processing apparatus 100 acquires the input "Western style is good" by the user and displays it as it is in the area D212 as the second turn on the display screen D21. When the information processing device 100 understands the user's answer, it corresponds to the information provision shown in FIG. 5, and further extracts the keyword "Western style".

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to the dialogue model 131 shown in FIG. 5, and when the open question is output in the state 1, the information provision is input, so that the dialogue state transitions from the state 1 to the state 2. It is determined that it is, and the dialogue state is updated.

<Turn 3 and 4>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 decides to ask an open question because it is in the state 2 with reference to FIG. 5, and the second open question “Are there any other conditions?” With reference to FIG. Generates a system utterance display. Here, the assumption of the next input shown in the open question of the state 2 of FIG. 5 is information provision or denial.

Next, the information processing apparatus 100 generates a dialogue information display that displays 158 items in which the candidates for the recipe information are narrowed down only by the keyword "Western style". Further, the information processing apparatus 100 extracts "main dish" and "pottery" as keyword candidates for narrowing down the candidates for recipe information with reference to FIG. 9, and "main dish is good" and "pottery is good". Generate an interactive information display that displays input examples of "good" and "no".

As a result, on the display screen D21, in the third turn, the question "Are there any other conditions?" In the area D213, the first information "recipe candidates: 158" in the area outside and above the area D213, the area. The second information "input example: main dish is good / baked food is good / no" is displayed in the outer and lower area of D213 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "with pottery" by the user and displays it as it is in the area D214 as the fourth turn on the display screen D21. When the information processing device 100 understands the user's answer, it corresponds to the information provision shown in FIG. 5, and further extracts the keyword "pottery".

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 determines that the dialogue state remains in the state 2 because the information provision is input when the open question is output in the state 2 with reference to FIG.

<Turn 5 and 6>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 decides to ask an open question because it is in the state 2 with reference to FIG. 5, and a third open question is asked with reference to FIG. Generate a system utterance display that says "Are you there?" Here, the assumption of the next input shown in the open question of the state 2 of FIG. 5 is information provision or denial.

Next, the information processing apparatus 100 generates a dialogue information display that displays 47 items in which recipe information candidates are narrowed down by the keywords "Western style" and "pottery". Further, the information processing apparatus 100 extracts "souvenirs" and "fish dishes" as keyword candidates for narrowing down the candidates for recipe information with reference to FIG. 9, and "wants to use souvenirs". Generate an interactive information display that displays input examples such as "I like fish dishes" and "No".

As a result, on the display screen D21, in the fifth turn, the question "Are there any other conditions?" In the area D215, the first information "recipe candidate: 47" in the area outside and above the area D215, the area. The second information "Input example: I want to use earthenware / I like fish dishes / No" is displayed in the outer and lower area of D215 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "Nothing in particular" by the user and displays it as it is in the area D216 as the sixth turn on the display screen D21. When the information processing device 100 understands the user's answer, it corresponds to the denial shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, referring to FIG. 5, the information processing apparatus 100 determines that the dialogue state has transitioned from the state 2 to the state 3 because a negative input is input when the open question is output in the state 2. Update the interactive state.

<Turn 7 and 8>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4. Up to this point, information has been narrowed down based on open questions, but from here on, information will be narrowed down based on closed questions. When the recipe information is narrowed down to 4 or less, the dialogue ends.

First, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, it is assumed that the information processing apparatus 100 extracts the keyword "earthenware" for narrowing down the candidates for recipe information with reference to FIG. In this case, as shown in FIG. 9, the relation between "earthenware" and cooking is "HasA". Therefore, can the information processing apparatus 100 refer to FIG. 10 and use "X"? Use the template "" to generate a system speech display with the closed question "Can I use earthenware?". Here, the assumption of the next input shown in the closed question of the state 3 of FIG. 5 is affirmative or negative. In addition, since the number of keywords has not increased in the sixth turn, which is the previous turn, the number of recipe information candidates remains 47.

Next, the information processing apparatus 100 generates a dialogue information display indicating that there are 47 recipe information candidates, and a dialogue information display displaying input examples of "Yes" and "No".

As a result, on the display screen D31, on the 7th turn, the question "Can I use earthenware?" The second information "input example: yes / no" is displayed in the outer and lower area of D311 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "what is it?" By the user and displays it as it is in the area D312 as the eighth turn on the display screen D31. When the information processing device 100 understands the user's answer, it corresponds to the user question (WHAT question) shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, referring to FIG. 5, the information processing apparatus 100 determines that the dialogue state has transitioned from the state 3 to the state 4 because the user question is input when the closed question is output in the state 3. , Update the dialogue state.

<Turn 9 and 10>
Next, the information processing device 100 performs a question or the like display process in step S3. Specifically, the information processing apparatus 100 refers to FIG. 5, and since it is in the state 4, it decides to give an answer. Here, the information processing apparatus 100 refers to the regions R11 and FIG. 10 of FIG. 9, and uses the template "for example, list_up (" Y ")." Generate a system speech display. Here, the assumption of the next input shown in the answer of the state 4 of FIG. 5 is “Φ”. Therefore, the information processing apparatus 100 determines that the dialogue state has transitioned to the state 3 as it is, and updates the dialogue state.

Next, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, since the keyword for narrowing down the recipe information candidates is still "earthenware", the information processing device 100 uses the same closed question as the 7th turn, "Can I use earthenware?" Generate a display. Here, as described above, the assumption of the next input shown in the open question of the state 3 of FIG. 5 is affirmative or negative, and the keywords have not increased even in the previous turn, so there are 47 recipe information candidates. There is up to. Therefore, the information processing apparatus 100 generates a dialogue information display indicating that there are 47 recipe information candidates, and a dialogue information display displaying input examples of "Yes" and "No".

As a result, on the display screen D31, on the 9th turn, the answer and question "For example, onions and dioscorea opposita. Can I use earthenware?" "Recipe candidates: 47" and the second information "Input example: Yes / No" are displayed in the outer and lower areas of the area D313 in a manner suitable for each.

Next, the information processing apparatus 100 acquires the input "OK" by the user and displays it as it is in the area D314 as the tenth turn on the display screen D31. When the information processing device 100 understands the user's answer, it corresponds to the affirmation shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5 and determines that the dialogue state remains in the state 3 because affirmation is input when the closed question is output in the state 3.

<Turn 11 and 12>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, it is assumed that the information processing apparatus 100 extracts the keyword "fish dish" (not shown) for narrowing down the candidates for recipe information. In this case, the information processing apparatus 100 refers to FIG. 10 and uses the template "Is it okay with" X "?" To generate a system utterance display of the closed question "Is it okay with fish dishes?". Here, the assumption of the next input shown in the closed question of the state 3 of FIG. 5 is affirmative or negative. In addition, it is assumed that the candidates for recipe information are narrowed down to 21 from the keywords "Western style", "pottery", and "earthenware".

Next, the information processing apparatus 100 generates a dialogue information display that displays that there are 21 recipe information candidates, and a dialogue information display that displays input examples of "yes" and "no."

As a result, on the display screen D31, in the 11th turn, the question "Is it okay to cook fish?" The second information "input example: yes / no" is displayed in the external and lower area in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "yes" by the user and displays it as it is in the area D316 as the twelfth turn on the display screen D31. When the information processing device 100 understands the user's answer, it corresponds to the affirmation shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5 and determines that the dialogue state remains in the state 3 because affirmation is input when the closed question is output in the state 3.

<Turn 13 and 14>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, it is assumed that the information processing apparatus 100 extracts the keyword "dairy product" (not shown) for narrowing down the candidates for recipe information. In this case, the information processing apparatus 100 refers to FIG. 10 and uses the template "Can I use" X "?" To display the closed question "Can I use dairy products?" To generate. Here, the assumption of the next input shown in the closed question of the state 3 of FIG. 5 is affirmative or negative. In addition, it is assumed that the recipe information is narrowed down to eight from the keywords "Western style", "pottery", "soil", and "fish dish".

Next, the information processing apparatus 100 generates a dialogue information display that displays that there are eight recipe information candidates, and a dialogue information display that displays input examples of "yes" and "no."

As a result, on the display screen D31, on the 13th turn, the question "Can I use dairy products?" The second information "input example: yes / no" is displayed in the outer and lower area of D317 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "yes" by the user and displays it as it is in the area D318 as the 14th turn on the display screen D31. When the information processing device 100 understands the user's answer, it corresponds to the affirmation shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5 and determines that the dialogue state remains in the state 3 because affirmation is input when the closed question is output in the state 3.

<Turn 15>
Subsequently, the information processing device 100 performs the process of step S3.

Here, if the candidates for recipe information are narrowed down from the keywords "Western style", "pottery", "soil", "fish dishes" and "dairy products", there are four cases.

In this case, the information processing apparatus 100 decides to present the conclusion because it is in the state 3 and the number of recipe information candidates is 4 or less with reference to FIG. Although the template for presenting the conclusion is not particularly shown, it may be stored in the sentence generation data 133 of FIG.

The information processing apparatus 100 refers to FIG. 10 and takes in the number of candidates for recipe information to generate, for example, a system utterance display saying "4 recipes have been found".

As a result, on the display screen D41, on the 15th turn, the conclusion was presented in the area D411 "4 recipes were found", and 4 recipe information candidates narrowed down to the outer and lower areas of the area D411, that is, found. A list of recipes is displayed.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5, outputs a conclusion presentation in the state 3, and determines that the dialogue state has transitioned from the state 3 to the state 5 when some input is obtained from the user. And update the dialogue state.

[Effects, etc.]
As described above, in the step of displaying the question, the information processing apparatus or the like according to the present embodiment displays the first question, which is the content of asking the request for the recipes or menus of a plurality of dishes, in the first aspect, and the first At the top of the question, the total number of recipes or menus for a plurality of dishes contained in the database 13 is displayed as the first information, and at the bottom of the first question, the recipes or menus for a plurality of dishes based on the database 13 are displayed. Recipes for multiple dishes after obtaining the first answer to the first question in the first step of displaying the answer example including one ingredient or a phrase indicating one classification as the second information and the step of obtaining the answer. Alternatively, the second question, which further asks for a request for the menu, is displayed in the first aspect, and the number of recipes or menus of a plurality of dishes contained in the database 13 narrowed down by the first answer at the top of the second question. Is displayed as the first information, and at the bottom of the second question, a phrase indicating one ingredient or one classification included in the recipes or menus of a plurality of dishes included in the database 13 narrowed down by the first answer is displayed. Includes a second step of displaying an example of an answer as second information. Here, in the database 13, the recipes of a plurality of dishes, the classification information indicating the classification of each of the recipes of the plurality of dishes, and the ingredient information indicating the ingredients used for each of the recipes of the plurality of dishes are associated with each other. ..

Further, in the information processing apparatus or the like according to the present embodiment, in the step of displaying the question, when a predetermined word or phrase is included in the second answer to the second question acquired in the step of acquiring the answer, the classification information is further added. In the step of obtaining an answer at the top of the third question, the third question, which is the content asking whether to request one of the categories included in the information or the one ingredient included in the ingredient information, is displayed in the first aspect. The number of recipes or menus of a plurality of dishes included in the database 13 narrowed down by the second answer to the acquired second question is displayed as the first information, and one classification or one is displayed at the bottom of the third question. An example of an answer including a phrase that affirms and a phrase that denies one ingredient is displayed as the second information.

Further, in the information processing apparatus and the like according to the present embodiment, the third answer to the third question acquired in the step of acquiring the answer contains affirmative words and phrases, and a plurality of information processing devices and the like included in the database 13 narrowed down by the third answer. When the number of cooking recipes or menus is less than or equal to the threshold value, the display includes a step of displaying information on a plurality of cooking recipes or menus contained in the database 13 narrowed down by the third answer. Then, in the step of displaying the question, if the number of recipes or menus of a plurality of dishes included in the database does not fall below the threshold value due to the third answer to the third question acquired in the step of acquiring the answer, one classification 13 The fourth question, which asks whether to request one category different from one or one ingredient different from one ingredient, is displayed in the first aspect, and is narrowed down by the third answer at the top of the fourth question. , The number of recipes or menus of a plurality of dishes included in the database 13 is displayed as the first information, and the answer example including the affirmative phrase and the negative phrase is included in the lower part of the third question as the second information. Display it.

In this way, the information processing device or the like according to the present embodiment can easily input a response to the system utterance by presenting reference dialogue information in a state where it can be easily distinguished from the system utterance display. Therefore, it is possible to obtain all the conditions related to the information that the user requests to search by interacting with the user.

(Example 2)
In the second embodiment, an example will be described in which a task of narrowing down a travel plan satisfying a plurality of conditions is performed as a task of proceeding with the narrowing down of information requested by the user. That is, a case where the search target is a plurality of travel plans and the question indicates the content of asking for a request for the plurality of travel plans will be described.

[Content data 132]
FIG. 12 is a diagram showing an example of content data 132 according to the second embodiment of the present embodiment.

FIG. 12 shows, as an example of the content data 132, a knowledge base expressing the relationship between the travel plan and the conditions in the form of a directed graph. Note that FIG. 12 corresponds to an example of a part of a huge knowledge base.

In the example shown in FIG. 12, plans 1 and 2 correspond to a plurality of candidates to be searched, and conditions such as Italy and 2 people correspond to element information (keywords) indicating features, attributes, conditions, and the like. Further, FIG. 12 shows a relationship that, for example, "Plan 1" of node1 satisfies the conditions of node2 such as Italy, 2 people, 5 nights, 150,000 or less, and "Italy" of node2 has plan 2. It is shown that there are other travel plans for mode1 that are less than 150,000. In order to show the above relationship, "HasProperty" is applied as the relative.

Therefore, information satisfying a plurality of conditions can be narrowed down by increasing the number of AND search keywords of a plurality of keywords by using the node2 of [node1 Hasproperty mode2] as a search keyword and the node1 associated with the search keyword. In this way, the travel plan can be narrowed down from a plurality of conditions.

Further, the elliptical region R12 shown by the dotted line represents the relationship that the superordinate concept for Italy and Germany, which is a condition, is for Europe. Region R12 is an example of knowledge used to answer a user question, assuming the user had no knowledge of "to Europe". The area R12 is set as a separate frame because it is not directly referred to for the search of the travel plan.

[Sentence generation data 133]
FIG. 13 is a diagram showing an example of sentence generation data 133 according to the second embodiment of the present embodiment. FIG. 13 illustrates an open question, a closed question, a knowledge acquisition question, an answer template, and their application conditions referred to by the information processing apparatus 100.

That is, when the information processing device 100 outputs an open question regarding the search target, the first open question shown in FIG. 13 "What kind of travel plan are you looking for?" At the start of the dialogue with the user. Is generated, and after that, the second and subsequent open questions shown in FIG. 13 are (repeatedly) generated asking "Are there any other conditions?".

Further, when the information processing apparatus 100 outputs a closed question regarding the search target, the closed question "Do you want to make a plan of" X "?" Is output using the template shown in FIG.

Further, when the information processing apparatus 100 outputs a knowledge acquisition question, it always outputs a question sentence "What kind of food is it?" And asks the user to exemplify desired information.

Further, when the information processing device 100 outputs an answer to the user question, the user question is a WHERE question related to "X", and the relationship "Y" IsA "X" is a knowledge base from the content data 132. When found, the answer is output using the answer template shown in FIG. 13, "For example, list_up (" Y ")."

Further, when the information processing apparatus 100 outputs an answer to the user question, the user question is a WHERE question related to "X", and the relationship "X" IsA "Z" is a knowledge base from the content data 132. When it is found, the answer is output using another answer template, "It is in" Z "." That is, the information processing device 100 generates an answer such as "for example, Y1 or Y2" or "is in Z" depending on the relationship existing in the content data 132 which is the knowledge base.

Although not shown, a template may be prepared in the sentence generation data 133 in order to generate an input example assumed as user input. For example, as the information providing template, "(keyword) is good", the affirmative answer template "yes", the negative answer template "no", and the like may be stored in the sentence generation data 133.

[Processing of information processing device 100]
FIG. 14 is a diagram showing an example of a display screen according to the second embodiment of the present embodiment.

The display screen D22 shows an example of a dialogue when the information processing device 100 asks an open question, and the display screen D32 shows an example of a dialogue when the information processing device 100 asks a closed question. The display screen D42 shows an example of a screen shown when the information processing apparatus 100 presents a conclusion.

Hereinafter, the processing described with reference to FIGS. 4 and 5 will be described.

In the initial state, that is, the state 1 of FIG. 5, it is assumed that the number of registered travel plan candidates, that is, plan information candidates, to be searched is 19870.

<Turn 1 and 2>
First, the information processing device 100 performs a question or the like display process in step S3. Specifically, first, the information processing apparatus 100 refers to the dialogue model 131 of FIG. 5, decides to ask an open question because it is in the initial state, that is, the state 1, and refers to the sentence generation data 133 of FIG. Generate the first open question "What kind of travel plan are you looking for?" System speech display. Here, the assumption of the next input shown in the state 1 of FIG. 5 is information provision. Next, the information processing apparatus 100 refers to the content data 132 of FIG. 12, extracts "Europe" and "North America" as keyword candidates for narrowing down the candidates for plan information, and "Europe is good" and "North America". Generate an interactive information display that displays an input example that says "is good".

As a result, on the display screen D22, the first turn asks the area D221 the question "What kind of travel plan are you looking for?", The first information "Plan candidates: 19870" in the area outside and above the area D221, the area. The second information "Input example: Europe is good / North America is good" is displayed in the outer and lower area of D221 in a mode suitable for each. In this way, while the question "What kind of travel plan are you looking for?" Is expressed in spoken language, the first information "plan candidates: 19870" and the second information "input example: Europe is good / North America is good". "Good" is simply expressed using symbols.

Next, the information processing apparatus 100 performs an answer or the like acquisition process in step S4. Specifically, first, the information processing apparatus 100 acquires the input "two people" by the user and displays it as it is in the area D222 as the second turn on the display screen D22. When the information processing device 100 understands the user's answer, it corresponds to the information provision shown in FIG. 5, and further extracts the keyword "2 persons".

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to the dialogue model 131 shown in FIG. 5, and when the open question is output in the state 1, the information provision is input, so that the dialogue state transitions from the state 1 to the state 2. It is determined that it is, and the dialogue state is updated.

<Turn 3 and 4>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 decides to ask an open question because it is in the state 2 with reference to FIG. 5, and the second open question “Are there any other conditions?” With reference to FIG. Generates a system utterance display. Here, the assumption of the next input shown in the open question of the state 2 of FIG. 5 is information provision or denial.

Next, the information processing apparatus 100 generates a dialogue information display that displays 1580 items in which the candidates for the plan information are narrowed down only by the keyword "2 persons". Further, the information processing apparatus 100 extracts "5 nights" and "4 people" as keyword candidates for narrowing down the candidates for plan information with reference to FIG. 12, and "5 nights is good" and "4". Generate an interactive information display that displays input examples such as "good name" and "no".

As a result, on the display screen D22, in the third turn, the question "Are there any other conditions?" In the area D223, the first information "plan candidates: 1580 cases" in the area outside and above the area D223, the area. The second information "input example: 5 nights is good / 4 people is good / no" is displayed in the outer and lower area of D223 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "at 5 nights" by the user and displays it as it is in the area D224 as the fourth turn on the display screen D22. When the information processing device 100 understands the user's answer, it corresponds to the information provision shown in FIG. 5, and further extracts the keyword "5 nights".

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 determines that the dialogue state remains in the state 2 because the information provision is input when the open question is output in the state 2 with reference to FIG.

<Turn 5 and 6>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 decides to ask an open question because it is in the state 2 with reference to FIG. 5, and a third open question is asked with reference to FIG. Generate a system utterance display that says "Are you there?" Here, the assumption of the next input shown in the open question of the state 2 of FIG. 5 is information provision or denial.

Next, the information processing device 100 generates a dialogue information display that displays 470 cases in which the candidates for the plan information are narrowed down by the keywords "2 people" and "5 nights". Further, the information processing apparatus 100 extracts "2 people" and "4 people" as keyword candidates for narrowing down the candidates for the plan information with reference to FIG. 12, and "2 people are good" and "2 people are good" and " Generate a dialogue information display that displays input examples such as "4 people are good" and "No".

As a result, on the display screen D22, on the 5th turn, the question "Are there any other conditions?" In the area D225, the first information "plan candidates: 470 cases" in the area outside and above the area D225, the area. The second information "input example: 2 people are good / 4 people are good / no" is displayed in the outer and lower area of D225 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "Nothing in particular" by the user and displays it as it is in the area D226 as the sixth turn on the display screen D22. When the information processing device 100 understands the user's answer, it corresponds to the denial shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, referring to FIG. 5, the information processing apparatus 100 determines that the dialogue state has transitioned from the state 2 to the state 3 because a negative input is input when the open question is output in the state 2. Update the interactive state.

<Turn 7 and 8>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4. Up to this point, information has been narrowed down based on open questions, but from here on, information will be narrowed down based on closed questions. When the plan information is narrowed down to 4 or less, the dialogue ends.

First, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, it is assumed that the information processing apparatus 100 extracts the keyword "to Europe" for narrowing down the candidates for the plan information with reference to FIG. In this case, as shown in FIG. 12, the relation between "to Europe" and the travel plan is "HasProperty", so the information processing apparatus 100 refers to FIG. 13 and makes the plan "X"? Use the template "" to generate a system utterance display with the closed question "Do you want to plan for Europe?". Here, the assumption of the next input shown in the closed question of the state 3 of FIG. 5 is affirmative or negative. In addition, since the number of keywords has not increased in the sixth turn, which is the previous turn, the number of recipe information candidates remains 470.

Next, the information processing device 100 generates a dialogue information display that displays that there are 470 candidates for plan information, and a dialogue information display that displays input examples of "yes" and "no."

As a result, on the display screen D32, on the 7th turn, the question "Do you want to make a plan for Europe?" The second information "input example: yes / no" is displayed in a mode suitable for each of the external and lower areas of.

Next, the information processing apparatus 100 acquires the input "where is it specifically?" By the user and displays it as it is in the area D322 as the eighth turn on the display screen D32. When the information processing device 100 understands the user's answer, it corresponds to the user question (WHERE question) shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, referring to FIG. 5, the information processing apparatus 100 determines that the dialogue state has transitioned from the state 3 to the state 4 because the user question is input when the closed question is output in the state 3. , Update the dialogue state.

<Turn 9 and 10>
Next, the information processing device 100 performs a question or the like display process in step S3. Specifically, the information processing apparatus 100 refers to FIG. 5, and since it is in the state 4, it decides to give an answer. Here, the information processing apparatus 100 refers to the area R12 and FIG. 13 of FIG. 12, and uses the template "For example, list_up (" Y ")" to answer "For example, to Italy, to Germany." Generates a system speech display. Here, the assumption of the next input shown in the answer of the state 4 of FIG. 5 is “Φ”. Therefore, the information processing apparatus 100 determines that the dialogue state has transitioned to the state 3 as it is, and updates the dialogue state.

Next, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, the keyword for narrowing down the candidates for the plan information is still "to Europe", so the information processing device 100 displays the same closed question as the 7th turn, "Do you want to make a plan for Europe?" To generate. Here, as described above, the assumption of the next input shown in the open question of state 3 in FIG. 5 is affirmative or negative, and the keywords have not increased even in the previous turn, so there are 470 candidates for plan information. There is up to. Therefore, the information processing apparatus 100 generates a dialogue information display indicating that there are 470 candidates for plan information and a dialogue information display displaying input examples of "Yes" and "No".

As a result, on the display screen D32, on turn 9, the answer and question "For example, to Italy, to Germany. Would you like to plan for Europe?", The first information in the area outside and above area D323. "Plan candidates: 470 cases" and the second information "input example: yes / no" are displayed in the outer and lower areas of the area D323 in a manner suitable for each.

Next, the information processing apparatus 100 acquires the input "OK" by the user and displays it as it is in the area D324 as the tenth turn on the display screen D32. When the information processing device 100 understands the user's answer, it corresponds to the affirmation shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5 and determines that the dialogue state remains in the state 3 because affirmation is input when the closed question is output in the state 3.

<Turn 11 and 12>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, it is assumed that the information processing apparatus 100 extracts the keyword "150,000 or less" for narrowing down the candidates for the plan information. In this case, the information processing apparatus 100 refers to FIG. 13 and uses the template "Do you want to make a plan of" X "?" To display the system utterance of the closed question "Do you want to make a plan of 150,000 or less?". Generate. Here, the assumption of the next input shown in the closed question of the state 3 of FIG. 5 is affirmative or negative. In addition, it is assumed that the candidates for plan information are narrowed down to 21 from the keywords "2 people", "5 nights", and "to Europe".

Next, the information processing device 100 generates a dialogue information display that displays that there are 21 candidates for plan information, and a dialogue information display that displays input examples of "yes" and "no."

As a result, on the display screen D32, on the 11th turn, the question "Do you want to make a plan of 150,000 or less?" The second information "input example: yes / no" is displayed in the outer and lower area of the D325 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "yes" by the user and displays it as it is in the area D326 as the twelfth turn on the display screen D32. When the information processing device 100 understands the user's answer, it corresponds to the affirmation shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5 and determines that the dialogue state remains in the state 3 because affirmation is input when the closed question is output in the state 3.

<Turn 12>
Subsequently, the information processing device 100 performs the process of step S3.

Here, if the candidates for plan information are narrowed down from the keywords "2 people", "5 nights", "to Europe", and "150,000 or less", there are two cases.

In this case, the information processing apparatus 100 decides to present the conclusion because it is in the state 3 and the number of candidates for the plan information is 4 or less, referring to FIG. Although the template for presenting the conclusion is not particularly shown, it may be stored in the sentence generation data of FIG.

The information processing apparatus 100 refers to FIG. 13 and takes in the number of candidates for plan information to generate, for example, a system utterance display of "two plans have been found".

As a result, on the display screen D42, on the 12th turn, the conclusion was presented in the area D421 "Two plans were found", and the two plan information candidates narrowed down to the outer and lower areas of the area D421, that is, found. A list of travel plans is displayed.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5, outputs a conclusion presentation in the state 3, and determines that the dialogue state has transitioned from the state 3 to the state 5 when some input is obtained from the user. And update the dialogue state.

(Example 3)
In the third embodiment, an example will be described in which a task of narrowing down real estate properties satisfying a plurality of conditions is performed as a task of proceeding with the narrowing down of information requested by the user. That is, a case where the search target is a property of a plurality of real estates and the question indicates the content of asking a request for the property of a plurality of real estates will be described.

[Content data 132]
FIG. 15 is a diagram showing an example of content data 132 according to the third embodiment of the present embodiment.

FIG. 15 shows, as an example of the content data 132, a knowledge base expressing the relationship between real estate properties and conditions in the form of a directed graph. Note that FIG. 15 corresponds to an example of a part of a huge knowledge base.

In the example shown in FIG. 15, the property ABC and the property DEF correspond to a plurality of candidates to be searched, and the conditions such as condominium, new construction, and near the station correspond to the element information (keyword). Further, in FIG. 15, for example, the relationship that the "property ABC" of the node 1 satisfies the conditions of the no d 2 such as the condominium, the new construction, the vicinity of the station, the second floor and above, and the floor is warm is shown. Shows the relationship that there are other real estate properties of node1 near the station such as property DEF. In order to show the above relationship, "HasProperty" is applied as the relative.

Therefore, information satisfying a plurality of conditions can be narrowed down by increasing the number of AND search keywords of a plurality of keywords by using the node2 of [node1 Hasproperty mode2] as a search keyword and the node1 associated with the search keyword. In this way, real estate properties can be narrowed down from a plurality of conditions.

Further, the elliptical region R13 shown by the dotted line shows the relationship that the condition of new construction is a superordinate concept of the conditions of new construction, 1 year construction, 2 years construction, and 3 years construction. The area R13 is an example of knowledge used for answering a user question when it is assumed that the user has no knowledge about "new construction". Area R13 is set as a separate frame because it is not directly referred to for the search of real estate properties.

[Sentence generation data 133]
FIG. 16 is a diagram showing an example of sentence generation data 133 according to the third embodiment of the present embodiment. FIG. 16 illustrates an open question, a closed question, a knowledge acquisition question, an answer template, and their application conditions referred to by the information processing apparatus 100.

That is, when the information processing device 100 outputs an open question regarding the search target, at the start of the dialogue with the user, the first open question shown in FIG. 16 "What kind of property are you looking for?" Is asked. After that, the second and subsequent open questions shown in FIG. 16 are (repeatedly) generated asking "Are there any other conditions?".

Further, when the information processing apparatus 100 outputs a closed question regarding the search target, the closed question "Is" X "better?" Is output using the template shown in FIG.

Further, when the information processing device 100 outputs a knowledge acquisition question, it always outputs a question sentence "What kind of property is it?" And asks the user to exemplify desired information.

Further, when the information processing device 100 outputs an answer to the user question, the user question is a WHAT question related to "X", and the relationship "Y" IsA "X" is a knowledge base from the content data 132. When found, the answer is output using the answer template shown in FIG. 16, "For example, list_up (" Y ")."

Further, when the information processing apparatus 100 outputs an answer to the user question, the user question is a WHAT question related to "X", and the relationship "X" IsA "Z" is a knowledge base from the content data 132. When found, the answer is output using another answer template, "An example of" Z "." That is, the information processing device 100 generates an answer such as "for example, Y1 and Y2" or "an example of Z" depending on the relationship existing in the content data 132 which is the knowledge base.

Although not shown, a template may be prepared in the sentence generation data 133 in order to generate an input example assumed as user input. For example, as the information providing template, "(keyword) is good", the affirmative answer template "yes", the negative answer template "no", and the like may be stored in the sentence generation data 133.

[Processing of information processing device 100]
FIG. 17 is a diagram showing an example of a display screen according to the third embodiment of the present embodiment.

The display screen D23 shows an example of a dialogue when the information processing device 100 asks an open question, and the display screen D33 shows an example of a dialogue when the information processing device 100 asks a closed question. The display screen D43 shows an example of a screen shown when the information processing apparatus 100 presents a conclusion.

Hereinafter, the processing described with reference to FIGS. 4 and 5 will be described.

In the initial state, that is, the state 1 of FIG. 5, it is assumed that the number of registered real estate property candidates, that is, property candidates to be searched is 198,700.

<Turn 1 and 2>
First, the information processing device 100 performs a question or the like display process in step S3. Specifically, first, the information processing apparatus 100 refers to the dialogue model 131 of FIG. 5, decides to ask an open question because it is in the initial state, that is, the state 1, and refers to the sentence generation data 133 of FIG. Generate the first open question "What kind of property are you looking for?" System speech display. Here, the assumption of the next input shown in the state 1 of FIG. 5 is information provision. Next, the information processing device 100 refers to the content data 132 of FIG. 15 and extracts "Tokyo" and "Osaka" as keyword candidates for narrowing down the property candidates, and "Tokyo is good" and "Osaka is good". Generate an interactive information display that displays an input example that says "OK".

As a result, on the display screen D23, in the first turn, the question "What kind of property are you looking for?" The second information "Input example: Tokyo is good / Osaka is good" is displayed in a mode suitable for each of the external and lower areas of. In this way, while the question "What kind of property are you looking for?" Is expressed in spoken language, the first information "property candidates: 198700" and the second information "input example: Tokyo is good / Osaka is good". "Is simply expressed using symbols.

Next, the information processing apparatus 100 performs an answer or the like acquisition process in step S4. Specifically, first, the information processing apparatus 100 acquires the input "Kadoma city is good" by the user, and displays it as it is in the area D232 as the second turn on the display screen D23. When the information processing device 100 understands the user's answer, it corresponds to the information provision shown in FIG. 5, and further extracts the keyword "Kadoma city" (not shown).

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to the dialogue model 131 shown in FIG. 5, and when the open question is output in the state 1, the information provision is input, so that the dialogue state transitions from the state 1 to the state 2. It is determined that it is, and the dialogue state is updated.

<Turn 3 and 4>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 decides to ask an open question because it is in the state 2 with reference to FIG. 5, and the second open question “Are there any other conditions?” With reference to FIG. Generates a system utterance display. Here, the assumption of the next input shown in the open question of the state 2 of FIG. 5 is information provision or denial.

Next, the information processing device 100 generates a dialogue information display that displays 350 cases in which the candidate properties are narrowed down only by the keyword "Kadoma city". Further, the information processing device 100 extracts "apartment" and "near the station" as keyword candidates for narrowing down the candidates for the property with reference to FIG. 15, and "a condominium is good" and "near the station is good". Generates an interactive information display that displays an input example of "No".

As a result, on the display screen D23, on the third turn, the question "Are there any other conditions?" In the area D233, the first information "property candidates: 350" in the area outside and above the area D233, the area. The second information "input example: apartment is good / near the station is good / no" is displayed in the outer and lower area of D233 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "it is good to be close to the station" by the user, and displays it as it is in the area D234 as the fourth turn on the display screen D23. When the information processing device 100 understands the user's answer, it corresponds to the information provision shown in FIG. 5, and further extracts the keyword "near the station".

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 determines that the dialogue state remains in the state 2 because the information provision is input when the open question is output in the state 2 with reference to FIG.

<Turn 5 and 6>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 decides to ask an open question because it is in the state 2 with reference to FIG. 5, and a third open question is asked with reference to FIG. Generate a system utterance display that says "Are you there?" Here, the assumption of the next input shown in the open question of the state 2 of FIG. 5 is information provision or denial.

Next, the information processing device 100 generates a dialogue information display that displays 80 properties narrowed down by the keywords "Kadoma city" and "near the station". Further, the information processing apparatus 100 extracts "new construction" and "second floor and above" as keyword candidates for narrowing down the property candidates with reference to FIG. 15, and "new construction is good" and "new construction is good". Generate a dialogue information display that displays input examples such as "I like the second floor and above" and "No".

As a result, on the display screen D23, on the 5th turn, the question "Are there any other conditions?" To the area D235, the first information "property candidates: 80 cases" in the area outside and above the area D235, the area. The second information "input example: new construction is good / second floor or higher is good / no" is displayed in the outer and lower area of D235 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "Nothing in particular" by the user and displays it as it is in the area D236 as the sixth turn on the display screen D23. When the information processing device 100 understands the user's answer, it corresponds to the denial shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, referring to FIG. 5, the information processing apparatus 100 determines that the dialogue state has transitioned from the state 2 to the state 3 because a negative input is input when the open question is output in the state 2. Update the interactive state.

<Turn 7 and 8>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4. Up to this point, information has been narrowed down based on open questions, but from here on, information will be narrowed down based on closed questions. When the number of properties is narrowed down to 4 or less, the dialogue ends.

First, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, it is assumed that the information processing apparatus 100 extracts the keyword "new construction" for narrowing down the candidate properties with reference to FIG. In this case, as shown in FIG. 15, the relation between "new construction" and the property is "HasProperty", so the information processing device 100 refers to FIG. 16 and should "X" be better? Use the template "" to generate a system utterance display with the closed question "Is it better to be new?". Here, the assumption of the next input shown in the closed question of the state 3 of FIG. 5 is affirmative or negative. In addition, since the number of keywords has not increased in the 6th turn, which is the previous turn, the number of property candidates remains 80.

Next, the information processing device 100 generates a dialogue information display indicating that there are 80 property candidates, and a dialogue information display displaying input examples of "Yes" and "No".

As a result, on the display screen D33, on the 7th turn, the question "Is it better to be new?" The second information "input example: yes / no" is displayed in the outer and lower area of the D331 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "what is it?" By the user and displays it as it is in the area D332 as the eighth turn on the display screen D33. When the information processing device 100 understands the user's answer, it corresponds to the user question (WHAT question) shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, referring to FIG. 5, the information processing apparatus 100 determines that the dialogue state has transitioned from the state 3 to the state 4 because the user question is input when the closed question is output in the state 3. , Update the dialogue state.

<Turn 9 and 10>
Next, the information processing device 100 performs a question or the like display process in step S3. Specifically, the information processing apparatus 100 refers to FIG. 5, and since it is in the state 4, it decides to give an answer. Here, the information processing apparatus 100 refers to the area R13 and FIG. 16 of FIG. 15, and uses the template “for example, list_up (“ Y ”)” to answer “for example, new construction, 1 year construction, construction”. It is 2 years old and 3 years old. " Here, the assumption of the next input shown in the answer of the state 4 of FIG. 5 is “Φ”. Therefore, the information processing apparatus 100 determines that the dialogue state has transitioned to the state 3 as it is, and updates the dialogue state.

Next, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, since the keyword for narrowing down the property candidates is still "new construction", the information processing device 100 displays the same closed question as the 7th turn, "Is it better to be new construction?" To generate. Here, as described above, the assumption of the next input shown in the open question of state 3 in FIG. 5 is affirmative or negative, and the keywords have not increased even in the previous turn, so the number of property candidates remains 80. Is. Therefore, the information processing device 100 generates a dialogue information display indicating that there are 80 property candidates and a dialogue information display displaying input examples of "Yes" and "No".

As a result, on the display screen D33, the 9th turn answers and asks the area D333 "For example, new construction, 1 year old, 2 years old, 3 years old. Is it better to be new?", Area D333 The first information "property candidates: 80 cases" is displayed in the external and upper area, and the second information "input example: yes / no" is displayed in the external and lower area of the area D333 in a manner suitable for each.

Next, the information processing apparatus 100 acquires the input "Oh, that is better" by the user, and displays it as it is in the area D334 as the tenth turn on the display screen D33. When the information processing device 100 understands the user's answer, it corresponds to the affirmation shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5 and determines that the dialogue state remains in the state 3 because affirmation is input when the closed question is output in the state 3.

<Turn 11 and 12>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, it is assumed that the information processing device 100 extracts the keyword "with floor warming" for narrowing down the candidate properties. In this case, the information processing apparatus 100 refers to FIG. 16 and uses the template "Is" X "better?" To utter a closed question "Is it better to have a warm floor?" Generate a display. Here, the assumption of the next input shown in the closed question of the state 3 of FIG. 5 is affirmative or negative. In addition, it is assumed that the number of property candidates is narrowed down to 21 from the keywords "Kadoma city", "near the station", and "new construction".

Next, the information processing device 100 generates a dialogue information display indicating that there are 21 property candidates and a dialogue information display displaying input examples of "Yes" and "No".

As a result, on the display screen D33, on the 11th turn, the question "Is it better to have a warm floor?" The second information "input example: yes / no" is displayed in a mode suitable for each of the outer and lower regions of the region D335.

Next, the information processing apparatus 100 acquires the input "it is not necessary" by the user and displays it as it is in the area D336 as the twelfth turn on the display screen D33. When the information processing device 100 understands the user's answer, it corresponds to the denial shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, referring to FIG. 5, the information processing apparatus 100 determines that the dialogue state remains in the state 3 because a negative is input when the closed question is output in the state 3.

<Turn 12>
Subsequently, the information processing device 100 performs the process of step S3.

Here, if the candidate properties are narrowed down from the keywords "Kadoma city", "near the station", "new construction", and "no floor warming", there are two cases.

In this case, the information processing apparatus 100 decides to present the conclusion because it is in the state 3 and the number of candidate properties is 4 or less with reference to FIG. Although the template for presenting the conclusion is not particularly shown, it may be stored in the sentence generation data of FIG.

The information processing device 100 refers to FIG. 16 and takes in the number of candidate properties, and generates, for example, a system utterance display such as "two properties have been found".

As a result, on the display screen D43, on the 12th turn, the conclusion was presented in the area D431 "Two properties were found", and two property candidates narrowed down to the outer and lower areas of the area D431, that is, the found real estate. A list of properties is displayed.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5, outputs a conclusion presentation in the state 3, and determines that the dialogue state has transitioned from the state 3 to the state 5 when some input is obtained from the user. And update the dialogue state.

(Example 4)
In the fourth embodiment, an example will be described in which a task of narrowing down video contents satisfying a plurality of conditions is performed as a task of proceeding with narrowing down of information requested by the user. That is, a case where the search target is a plurality of video contents and the question indicates a content asking for a request for the plurality of video contents will be described.

[Content data 132]
FIG. 18 is a diagram showing an example of content data 132 according to the fourth embodiment of the present embodiment.

FIG. 18 shows, as an example of the content data 132, a knowledge base expressing the relationship between the video content and the condition in the form of a directed graph. Note that FIG. 18 corresponds to an example of a part of a huge knowledge base.

In the example shown in FIG. 18, Z school and detective X correspond to a plurality of candidates to be searched, and conditions such as drama, late-night program, and TV Kanto correspond to element information (keywords). Further, in FIG. 18, for example, the relationship that "Z Gakuen" of node1 satisfies the conditions of node2 such as actor A appearance, drama, late-night program, TV Kanto, and school thing is shown, and the "drama" of node2 has a relationship. , Detective X and other late-night programs such as node1 have other video contents. In order to show the above relationship, "HasProperty" is applied as the relative.

Therefore, information satisfying a plurality of conditions can be narrowed down by increasing the number of AND search keywords of a plurality of keywords by using the node2 of [node1 Hasproperty mode2] as a search keyword and the node1 associated with the search keyword. In this way, video content candidates can be narrowed down from a plurality of conditions.

Further, the elliptical region R14 shown by the dotted line shows the relationship that the drama, which is a condition, is a superordinate concept of the school, the detective, the family, and the romance, which are the conditions. The area R14 is an example of knowledge used for answering a user question when it is assumed that the user has no knowledge about "drama". The area R14 is set as a separate frame because it is not directly referred to for searching for a candidate for video content.

[Sentence generation data 133]
FIG. 19 is a diagram showing an example of sentence generation data 133 according to the fourth embodiment of the present embodiment. FIG. 19 illustrates an open question, a closed question, a knowledge acquisition question, an answer template, and their application conditions referred to by the information processing apparatus 100.

That is, when the information processing device 100 outputs an open question regarding the search target, at the start of the dialogue with the user, the first question sentence of the open question shown in FIG. 19 "What kind of program are you looking for?" Is displayed. After that, the second and subsequent open questions shown in FIG. 19 are repeatedly generated asking "Are there any other conditions?".

Further, when the information processing apparatus 100 outputs a closed question regarding the search target, the closed question "Is it okay with" X "?" Is output using the template shown in FIG.

Further, when the information processing apparatus 100 outputs a knowledge acquisition question, it always outputs a question sentence "What kind of program is it?" And asks the user to exemplify desired information.

Further, when the information processing device 100 outputs an answer to the user question, the user question is a WHAT question related to "X", and the relationship "Y" IsA "X" is a knowledge base from the content data 132. When found, the answer is output using the answer template shown in FIG. 16, "For example, list_up (" Y ")."

Further, when the information processing apparatus 100 outputs an answer to the user question, the user question is a WHAT question related to "X", and the relationship "X" IsA "Z" is a knowledge base from the content data 132. When found, the answer is output using another answer template, "An example of" Z "." That is, the information processing device 100 generates an answer such as "for example, Y1 and Y2" or "an example of Z" depending on the relationship existing in the content data 132 which is the knowledge base.

Although not shown, a template may be prepared in the sentence generation data 133 in order to generate an input example assumed as user input. For example, as the information providing template, "(keyword) is good", the affirmative answer template "yes", the negative answer template "no", and the like may be stored in the sentence generation data 133.

[Processing of information processing device 100]
FIG. 20 is a diagram showing an example of a display screen according to the fourth embodiment of the present embodiment.

The display screen D24 shows an example of a dialogue when the information processing device 100 asks an open question, and the display screen D34 shows an example of a dialogue when the information processing device 100 asks a closed question. The display screen D44 shows an example of a screen shown when the information processing apparatus 100 presents a conclusion.

Hereinafter, the processing described with reference to FIGS. 4 and 5 will be described.

In the initial state, that is, the state 1 of FIG. 5, it is assumed that the number of registered candidates for a plurality of video contents to be searched, that is, candidates for program information is 987.

<Turn 1 and 2>
First, the information processing device 100 performs a question or the like display process in step S3. Specifically, first, the information processing apparatus 100 refers to the dialogue model 131 of FIG. 5, decides to ask an open question because it is in the initial state, that is, the state 1, and refers to the sentence generation data 133 of FIG. Generates the first open question "What program are you looking for?" System speech display. Here, the assumption of the next input shown in the state 1 of FIG. 5 is information provision. Next, the information processing apparatus 100 refers to the content data 132 of FIG. 18, extracts "drama" and "movie" (not shown) as keyword candidates for narrowing down the candidates for program information, and "drama is Generate an interactive information display that displays input examples such as "good" and "good movie".

As a result, on the display screen D24, the first turn asks the area D241 the question "What kind of program are you looking for?" The second information "input example: drama is good / movie is good" is displayed in a mode suitable for each of the external and lower areas of. In this way, while the question "What kind of program are you looking for?" Is expressed in spoken language, the first information "program candidates: 987" and the second information "input example: drama is good / movie is good". "Is simply expressed using symbols.

Next, the information processing apparatus 100 performs an answer or the like acquisition process in step S4. Specifically, first, the information processing apparatus 100 acquires the input "midnight program is good" by the user, and displays it as it is in the area D242 as the second turn on the display screen D23. When the information processing device 100 understands the user's answer, it corresponds to the information provision shown in FIG. 5, and further extracts the keyword "late night program".

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to the dialogue model 131 shown in FIG. 5, and when the open question is output in the state 1, the information provision is input, so that the dialogue state transitions from the state 1 to the state 2. It is determined that it is, and the dialogue state is updated.

<Turn 3 and 4>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 decides to ask an open question because it is in the state 2 with reference to FIG. 5, and the second open question “Are there any other conditions?” With reference to FIG. Generates a system utterance display. Here, the assumption of the next input shown in the open question of the state 2 of FIG. 5 is information provision or denial.

Next, the information processing apparatus 100 generates a dialogue information display that displays 350 programs in which the candidates for the program information are narrowed down only by the keyword "midnight program". Further, the information processing apparatus 100 extracts "drama" and "movie" as keyword candidates for narrowing down the candidates for program information with reference to FIG. 18, and "drama is good" and "movie is good". Generate an interactive information display that displays an example of the input "No".

As a result, on the display screen D24, in the third turn, the question "Are there any other conditions?" In the area D243, the first information "program candidates: 350" in the area outside and above the area D243, the area. The second information "input example: drama is good / movie is good / no" is displayed in the outer and lower area of D243 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "TV Kanto is good" by the user and displays it as it is in the area D244 as the fourth turn on the display screen D24. When the information processing device 100 understands the user's answer, it corresponds to the information provision shown in FIG. 5, and further extracts the keyword "television Kanto".

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 determines that the dialogue state remains in the state 2 because the information provision is input when the open question is output in the state 2 with reference to FIG.

<Turn 5 and 6>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 decides to ask an open question because it is in the state 2 with reference to FIG. 5, and a third open question is asked with reference to FIG. Generate a system utterance display that says "Are you there?" Here, the assumption of the next input shown in the open question of the state 2 of FIG. 5 is information provision or denial.

Next, the information processing device 100 generates a dialogue information display that displays 80 programs narrowed down by the keywords "midnight program" and "television Kanto". Further, the information processing apparatus 100 extracts "drama" and "movie" as keyword candidates for narrowing down the candidates for program information with reference to FIG. 18, and "drama is good" and "movie is good". , And generate an interactive information display that displays input examples such as "No".

As a result, on the display screen D24, in the fifth turn, the question "Are there any other conditions?" In the area D245, the first information "program candidates: 80 cases" in the area outside and above the area D245, the area. The second information "input example: drama is good / movie is good / no" is displayed in the outer and lower area of D235 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "Nothing in particular" by the user and displays it as it is in the area D246 as the sixth turn on the display screen D24. When the information processing device 100 understands the user's answer, it corresponds to the denial shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, referring to FIG. 5, the information processing apparatus 100 determines that the dialogue state has transitioned from the state 2 to the state 3 because a negative input is input when the open question is output in the state 2. Update the interactive state.

<Turn 7 and 8>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4. Up to this point, information has been narrowed down based on open questions, but from here on, information will be narrowed down based on closed questions. When the program information is narrowed down to 4 or less, the dialogue ends.

First, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, it is assumed that the information processing apparatus 100 extracts the keyword "drama" for narrowing down the candidates for program information with reference to FIG. In this case, as shown in FIG. 18, since the relative word between "drama" and the video content is "Hasproperty", the information processing apparatus 100 refers to FIG. 19 and asks "Is it okay to use" X "?" Use the template to generate a system utterance display with the closed question "Are you okay with the drama?" Here, the assumption of the next input shown in the closed question of the state 3 of FIG. 5 is affirmative or negative. In addition, since the number of keywords has not increased in the sixth turn, which is the previous turn, the number of program information candidates remains 80.

Next, the information processing device 100 generates a dialogue information display that displays that there are 80 program information candidates, and a dialogue information display that displays input examples of "yes" and "no."

As a result, on the display screen D34, on the 7th turn, the question "Is it okay with the drama?" Moreover, the second information "input example: yes / no" is displayed in the lower area in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "what is it?" By the user and displays it as it is in the area D342 as the eighth turn on the display screen D34. When the information processing device 100 understands the user's answer, it corresponds to the user question (WHAT question) shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, referring to FIG. 5, the information processing apparatus 100 determines that the dialogue state has transitioned from the state 3 to the state 4 because the user question is input when the closed question is output in the state 3. , Update the dialogue state.

<Turn 9 and 10>
Next, the information processing device 100 performs a question or the like display process in step S3. Specifically, the information processing apparatus 100 refers to FIG. 5, and since it is in the state 4, it decides to give an answer. Here, the information processing apparatus 100 refers to the area R14 and FIG. 19 of FIG. 18, and uses the template “for example, list_up (“ Y ”)” to answer “for example, school thing, detective thing, family”. Generate a system utterance display that says "things, love things." Here, the assumption of the next input shown in the answer of the state 4 of FIG. 5 is “Φ”. Therefore, the information processing apparatus 100 determines that the dialogue state has transitioned to the state 3 as it is, and updates the dialogue state.

Next, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, since the keyword for narrowing down the candidates for the program information remains "drama", the information processing device 100 generates the same closed question "Is it okay with the drama?" System utterance display as in the seventh turn. .. Here, as described above, the assumption of the next input shown in the open question of the state 3 of FIG. 5 is affirmative or negative, and the keywords have not increased even in the previous turn, so there are 80 candidates for program information. There is up to. Therefore, the information processing device 100 generates a dialogue information display that displays that there are 80 program information candidates, and a dialogue information display that displays input examples of "yes" and "no."

As a result, on the display screen D34, turn 9 answers and asks the area D343 "For example, school, detective, family, love. Is it a drama?", The area outside and above area D343. The first information "program candidates: 80 cases" and the second information "input example: yes / no" are displayed in the outer and lower areas of the area D343 in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "looks good" by the user and displays it as it is in the area D344 as the tenth turn on the display screen D34. When the information processing device 100 understands the user's answer, it corresponds to the affirmation shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5 and determines that the dialogue state remains in the state 3 because affirmation is input when the closed question is output in the state 3.

<Turn 11 and 12>
Subsequently, the information processing apparatus 100 repeats the processes of steps S3 and S4.

First, the information processing apparatus 100 determines to ask a closed question because it is in the state 3 with reference to FIG. Here, it is assumed that the information processing apparatus 100 extracts the keyword "actor A appearance" for narrowing down the candidates for program information. In this case, the information processing apparatus 100 refers to FIG. 19 and uses the template "Is it okay with" X "?" To generate a system utterance display of the closed question "Is it okay to appear in actor A?". .. Here, the assumption of the next input shown in the closed question of the state 3 of FIG. 5 is affirmative or negative. In addition, it is assumed that the candidates for program information are narrowed down to 21 from the keywords "midnight program", "TV Kanto", and "drama".

Next, the information processing device 100 generates a dialogue information display that displays that there are 21 program information candidates, and a dialogue information display that displays input examples of "yes" and "no."

As a result, on the display screen D34, on the 11th turn, the question "Is it okay to appear in actor A?" The second information "input example: yes / no" is displayed in the outer and lower area of the screen in a mode suitable for each.

Next, the information processing apparatus 100 acquires the input "good" by the user and displays it as it is in the area D346 as the twelfth turn on the display screen D33. When the information processing device 100 understands the user's answer, it corresponds to the affirmation shown in FIG.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5 and determines that the dialogue state remains in the state 3 because affirmation is input when the closed question is output in the state 3.

<Turn 12>
Subsequently, the information processing device 100 performs the process of step S3.

Here, if the candidates for program information are narrowed down from the keywords "midnight program", "TV Kanto", "drama", and "actor A appearance", there are four cases.

In this case, the information processing apparatus 100 decides to present the conclusion because it is in the state 3 and the number of candidates for the program information is 4 or less with reference to FIG. Although the template for presenting the conclusion is not particularly shown, it may be stored in the sentence generation data of FIG.

The information processing apparatus 100 refers to FIG. 19 and takes in the number of candidate program information, and generates, for example, a system utterance display such as "4 program items have been found".

As a result, on the display screen D44, on the 12th turn, the conclusion was presented in the area D441, "4 programs were found", and 4 program information candidates narrowed down to the outer and lower areas of the area D441, that is, the found images. A list of contents is displayed.

Next, the information processing apparatus 100 updates the dialogue state in step S2. Specifically, the information processing apparatus 100 refers to FIG. 5, outputs a conclusion presentation in the state 3, and determines that the dialogue state has transitioned from the state 3 to the state 5 when some input is obtained from the user. And update the dialogue state.

Although the information processing method, the information processing apparatus, and the like according to the embodiment have been described above, the present disclosure is not limited to this embodiment.

Further, each processing unit included in the information processing apparatus according to the above embodiment is typically realized as an LSI which is an integrated circuit. These may be individually integrated into one chip, or may be integrated into one chip so as to include a part or all of them.

Further, the integrated circuit is not limited to the LSI, and may be realized by a dedicated circuit or a general-purpose processor. An FPGA (Field Programmable Gate Array) that can be programmed after the LSI is manufactured, or a reconfigurable processor that can reconfigure the connection and settings of circuit cells inside the LSI may be used.

Further, the present disclosure may be realized as an information processing method executed by an information processing apparatus.

Further, in each of the above-described embodiments, each component may be configured by dedicated hardware or may be realized by executing a software program suitable for each component. Each component may be realized by a program execution unit such as a CPU or a processor reading and executing a software program recorded on a recording medium such as a hard disk or a semiconductor memory.

Further, the division of the functional block in the block diagram is an example, and a plurality of functional blocks can be realized as one functional block, one functional block can be divided into a plurality of functional blocks, and some functions can be transferred to other functional blocks. You may. In addition, the functions of a plurality of functional blocks having similar functions may be processed by a single hardware or software in parallel or in a time division manner.

Further, the order in which each step in the flowchart is executed is for exemplifying the present disclosure in detail, and may be an order other than the above. Further, a part of the above steps may be executed at the same time (parallel) as other steps.

Although the information processing apparatus and the like according to one or more aspects have been described above based on the embodiment, the present disclosure is not limited to this embodiment. As long as the gist of the present disclosure is not deviated, various modifications that can be conceived by those skilled in the art are applied to the present embodiment, and a form constructed by combining components in different embodiments is also within the scope of one or more embodiments. May be included within.

The present disclosure can be applied to information processing methods, information processing devices and programs having a function of interactively narrowing down and providing information using a database. For example, interview information, recipe information, travel information, real estate information, video content information. , Tourist information, job information, marriage partner information, etc. can be used for information processing methods, information processing devices and programs that provide information through dialogue with users.

11 Input processing unit 12 Dialogue control unit 13 Database 121 Dialogue management unit 122 System utterance display generation unit 123 Dialogue information display generation unit 124 Display integration unit 131 Dialogue model 132 Content data 133 Sentence generation data 100 Information processing device 200, 300 Terminal 400 Network D11, D12, D21, D22, D23, D24, D31, D32, D33, D34, D41, D42, D43, D44, D91 Display screen

Claims (21)

It is an information processing method that causes a computer to execute narrowing down of information requested by the user by repeating the display of a question and the acquisition of an answer from a user to the question.
A step of displaying a question regarding the search target in the first aspect on a display provided in the terminal, and
A step of obtaining an answer to the question entered by the user through the terminal, and
In the step of displaying the question, the first information numerically representing the progress of the narrowing down is displayed in the second aspect at the upper part of the question, and the answer guide to the question is displayed at the lower part of the question. 2 Let the computer execute to display the information in the third aspect.
The first information includes the remaining number of times when the question is displayed out of the total number of times the question can be asked in advance, or the cumulative number of times when the question is displayed with respect to the total number of times the question can be asked in advance. Including the degree of progress that indicates
Information processing method. The information processing method further
When the progress of the narrowing down satisfies a predetermined condition, the computer is made to perform a step of displaying the information indicating the candidate of the search target narrowed down by the progress of the narrowing down on the display.
The information processing method according to claim 1. The predetermined condition is that the number of candidates to be searched is equal to or less than a predetermined number.
The information processing method according to claim 2. In the step of displaying the question, in the frame area which is a balloon-shaped, rectangular or rounded rectangular area surrounded by a frame, characters indicating the contents of the question are displayed in a predetermined font name, font size, text color and background. By displaying the question in the specifications specified in color or typeface, the computer is made to display the question in the first aspect.
The information processing method according to any one of claims 1 to 3. In the step of displaying the question, the characters indicating the first information are displayed in the area outside and above the frame area with the specifications specified by a predetermined font name, font size, text color, background color, or typeface. By causing the computer to display the first information in the second aspect.
The information processing method according to claim 4. The character specifications indicating the content of the question in the first aspect and the character specifications indicating the first information in the second aspect are the same.
The information processing method according to claim 5. In the step of displaying the question, the characters indicating the second information are displayed in the area outside and below the frame area with the specifications specified by the predetermined font name, font size, text color, background color, or typeface. By causing the computer to display the second information in the third aspect.
The information processing method according to claim 5 or 6. The character specifications indicating the content of the question in the first aspect and the character specifications indicating the second information in the third aspect are the same.
The information processing method according to claim 7. The character specifications indicating the first information in the second aspect and the character specifications indicating the second information in the third aspect are the same.
The information processing method according to claim 7 or 8. The second information indicates a response method based on an example response including a character indicating affirmation and a character indicating denial.
The information processing method according to any one of claims 1 to 9. The second information indicates an answer method based on an example answer including words and phrases that can narrow down the search target.
The information processing method according to any one of claims 1 to 10. The target of the search is a recipe or menu of a plurality of dishes.
The question indicates the content of asking for a request for the recipe or menu of the plurality of dishes.
The information processing method according to any one of claims 1 to 11. The target of the search is a plurality of travel plans.
The question indicates the content of asking for a request for the plurality of travel plans.
The information processing method according to any one of claims 1 to 11. The target of the search is a plurality of real estate properties,
The question indicates the content of asking for a request for the plurality of real estate properties.
The information processing method according to any one of claims 1 to 11. The target of the search is a plurality of video contents.
The question indicates the content of asking for a request for the plurality of video contents.
The information processing method according to any one of claims 1 to 11. In the step of displaying the question,
The first question, which is a content for asking a request for the recipes or menus of the plurality of dishes, is displayed in the first aspect, and the total number of recipes or menus of the plurality of dishes included in the database is displayed at the top of the first question. Is displayed as the first information, and at the bottom of the first question, an example of an answer including a phrase indicating one ingredient or one classification included in the recipes or menus of the plurality of dishes based on the database. The first step of displaying as the second information and
After obtaining the first answer to the first question in the step of obtaining the answer, the second question, which is the content of further asking the request for the recipes or menus of the plurality of dishes, is displayed in the first aspect, and the second question is displayed. At the top of the question, the number of recipes or menus of the plurality of dishes included in the database, which is narrowed down by the first answer, is displayed as the first information, and at the bottom of the second question, the first The computer executes the second step of displaying as the second information an answer example including one ingredient or a phrase indicating one classification included in the recipes or menus of the plurality of dishes included in the database narrowed down by the answers. Let me
In the database, recipes of a plurality of dishes, classification information indicating the classification of each of the recipes of the plurality of dishes, and ingredient information indicating ingredients used for each of the recipes of the plurality of dishes are associated with each other.
The information processing method according to claim 12. In the step of displaying the question,
When a predetermined phrase is included in the second answer to the second question acquired in the step of acquiring the answer, further, one classification included in the classification information or one ingredient included in the ingredient information is included. The third question, which is the content asking whether to request or not, is displayed in the first aspect, and is narrowed down to the upper part of the third question by the second answer to the second question acquired in the step of acquiring the answer. The number of recipes or menus of the plurality of dishes included in the database is displayed as the first information, and at the bottom of the third question, a phrase affirming the one classification or the one ingredient and Have the computer execute to display the answer example including the negative phrase as the second information.
The information processing method according to claim 16. Moreover,
The number of recipes or menus of the plurality of dishes included in the database, which includes the affirmative phrase in the third answer to the third question acquired in the acquisition step for the answer and is narrowed down by the third answer. When is less than or equal to the threshold value, the computer is made to perform a step of displaying the information of the recipes or menus of the plurality of dishes contained in the database, which is narrowed down by the third answer on the display.
In the step of displaying the question,
If the number of recipes or menus of the plurality of dishes included in the database does not fall below the threshold due to the third answer to the third question acquired in the step of acquiring the answer, one classification different from the one classification is made. Alternatively, the fourth question, which is a content asking whether to request one ingredient different from the one ingredient, is displayed in the first aspect, and is narrowed down to the upper part of the fourth question by the third answer. The number of recipes or menus of the plurality of dishes included in the database is displayed as the first information, and the second information is an example of an answer including the affirmative phrase and the negative phrase at the bottom of the third question. Let the computer do what you want to see as
The information processing method according to claim 17. An information processing device that advances the narrowing down of information requested by the user to search by repeating the display of the question and the acquisition of the answer from the user to the question.
A dialogue control unit that displays a question about the search target in the first aspect on a display provided in the terminal.
An input processing unit for acquiring an answer to the question input by the user through the terminal is provided.
The dialogue control unit displays the first information numerically representing the progress of the narrowing down in the upper part of the question in the second aspect, and the lower part of the question regarding the answer guidance to the question. 2 Display the information in the third aspect ,
The first information includes the remaining number of times when the question is displayed out of the total number of times the question can be asked in advance, or the cumulative number of times when the question is displayed with respect to the total number of times the question can be asked in advance. Including the degree of progress that indicates
Information processing device. A program that causes a computer to execute an information processing method that advances the narrowing down of information requested by the user by repeating the display of a question and the acquisition of an answer from the user to the question.
A step of displaying a question regarding the search target in the first aspect on a display provided in the terminal, and
Through the terminal, the step of acquiring the answer to the question input by the user is executed.
In the step of displaying the question, the first information numerically representing the progress of the narrowing down is displayed in the second aspect at the upper part of the question, and the answer guide to the question is displayed at the lower part of the question. 2 Information is displayed in the third mode,
The first information includes the remaining number of times when the question is displayed out of the total number of times the question can be asked in advance, or the cumulative number of times when the question is displayed with respect to the total number of times the question can be asked in advance. Including the degree of progress that indicates
program. It is an information processing method that causes a computer to execute narrowing down of information requested by the user by repeating the display of a question and the acquisition of an answer from a user to the question.
A step of displaying a question regarding the search target in the first aspect on a display provided in the terminal, and
A step of obtaining an answer to the question entered by the user through the terminal, and
In the step of displaying the question, the first information regarding the progress of the narrowing down is displayed in the second aspect at the upper part of the question, and the second information regarding the answer guidance to the question is displayed at the lower part of the question in the third aspect. Let the computer do what you want to see in
The first information includes (1) the remaining number of times when the question is displayed out of the total number of times the question can be asked in advance, and (2) when the question is displayed for the total number of times the question can be asked in advance. Including at least one of the degree of progress, which indicates the cumulative number of times,
Information processing method.

Images