JP6166889B2 - Dialog support apparatus, dialog system, dialog support method and program - Google Patents

Description

  The present invention relates to a so-called voice agent service.

  A technique is known in which a user searches for information by inputting a question or the like by voice, and outputs a search result as characters or images (see, for example, Patent Document 1). Also, in mobile terminals such as smartphones, a voice interaction function is also put into practical use, in which information is searched by voice or the terminal is operated, and the terminal can respond by voice. With such a voice interaction function, the user can operate the portable terminal as if speaking to a human.

JP 2006-209022 A

By the way, in a dialogue with a real person, there may be a case where a certain response is expected, such as a search for information or an operation of a portable terminal, and a case where it is not. For example, the user does not necessarily expect a constant response in a casual conversation such as chat. Moreover, if only a fixed response is made to the chat, the response will be exactly mechanical and it can be said that human feeling is not felt. Therefore, as in the technique described in Patent Document 1, simply searching for an answer from an existing database cannot provide an appropriate response to a chat or the like, even if an appropriate answer to the information search is obtained. .
Therefore, an object of the present invention is to provide a technique capable of changing a response method according to a user's intention.

  The present invention provides an acquisition unit that acquires input information generated based on voice uttered by a user, an interpretation unit that interprets the intention of the user's utterance based on the input information acquired by the acquisition unit, When the intention interpreted by the interpreter is a search for information, response information corresponding to the search result of the search executed according to the intention is output, and the intention interpreted by the interpreter is not a search for information In some cases, a dialogue support device is provided that includes an output unit that selectively outputs any one of a plurality of response information prepared corresponding to the intention.

Further, the output unit may be configured such that the information source to be referred to is different depending on whether the intention interpreted by the interpretation unit is a search for information or not.
The output unit may be configured to determine whether the intention interpreted by the interpretation unit is a search for information depending on whether a predetermined character string is included.
The output unit may output response information corresponding to a character selected by the user among the plurality of response information.
The output unit may output the response information and additional information added to the response information and corresponding to the character selected by the user.
Further, the output unit is configured to selectively output any of a plurality of response information prepared in advance when the user who is speaking, the position of the user, or the date and time of speaking satisfies a predetermined condition. Also good.

  The present invention also provides a dialog system comprising the dialog support device and a user terminal that collects the user's voice and reproduces at least the voice according to the response information output by the output unit. .

  The present invention also includes a step of acquiring input information generated based on speech uttered by a user, a step of interpreting the user's utterance intention based on the acquired input information, and the interpretation When the intention is a search of information, response information corresponding to the search result of the search executed according to the intention is output, and when the interpreted intention is not a search of information, And a step of selectively outputting any of the plurality of prepared response information.

  According to another aspect of the present invention, a computer acquires input information generated based on speech uttered by a user, interprets the user's utterance intention based on the acquired input information, When the interpreted intention is a search of information, response information is output according to the search result of the search executed according to the intention, and when the interpreted intention is not a search of information, the intention is A program for executing a step of selectively outputting any of a plurality of response information prepared correspondingly is provided.

  According to the present invention, it is possible to change the response method according to the user's intention.

Block diagram showing the overall configuration of the dialogue system Block diagram showing the hardware configuration of the user terminal Block diagram showing the hardware configuration of a speech recognition server, intention interpretation server, and chat management server Block diagram showing functional configuration of intent interpretation server Diagram illustrating data structure of chat data Sequence chart showing operation until response to voice input is obtained Flow chart showing intention interpretation processing The figure which illustrates the output result in a user terminal The figure which illustrates the output result in a user terminal The figure which illustrates response information and additional information

[Example]
FIG. 1 is a block diagram showing the overall configuration of a dialog system 10 according to an embodiment of the present invention. The dialogue system 10 is a system for providing a so-called voice agent service, and includes a user terminal 100, a voice recognition server 200A, an intention interpretation server 200B, a chat management server 200C, and a voice synthesis server 200D. It has the structure connected to. In addition, although illustration is abbreviate | omitted here, there exist as many user terminals 1000 as the number of users who use this service. The network 300 is a complex network that combines a mobile communication network and the Internet.

  Here, the voice agent service refers to a service that mediates various user operations on the user terminal 100 by voice. According to this service, when searching for information, the user can obtain a desired result by simply speaking to the user terminal 100 in a natural sentence without manually inputting characters into the user terminal 100. Is possible.

  In addition, the voice agent service of the present embodiment has a first feature in that a service is provided through a character. The character referred to here is a character imitating a cartoon character, a celebrity, an animal, or the like, and is animated by, for example, three-dimensional CG (Computer Graphics). Various characters are provided by content providers, and each user can select and use a desired character for a fee or free of charge. These characters have different voices (voice colors) and tone for each character. In the service of this embodiment, the user uses his / her favorite character and speaks to it (as if interacting with a real person), so that he / she is attached to a specific character and promotes the use of this service. Is expected to be achieved.

  Furthermore, the voice agent service of the present embodiment has a second feature in that a user can perform a dialog other than information retrieval. Here, the dialogue other than information retrieval refers to something like so-called chat or social talk. In other words, the dialog here is not what the user asks for a specific result as in the so-called web search, but is what is performed as a conversation or killing time in communication with a real person.

  The user terminal 100 is a communication terminal used by a user to use this service. Here, it is assumed that the user terminal 100 is a smartphone. The speech recognition server 200A is a server device for converting speech data generated based on speech uttered by a user into text data. The intention interpretation server 200B is a server device that interprets a user's intention based on text data and transmits response information corresponding to the intention to the user terminal 100. The chat management server 200C is a server device that manages chat data for each character, and transmits the chat data to the intention interpretation server 200B at a predetermined timing. The speech synthesis server 200D is a server device that stores speech data for each character (hereinafter referred to as “synthesized speech data”), and transmits this to the user terminal 100 as necessary. The synthesized voice data is data provided by the content provider, and is associated with a character ID and a response sentence, which will be described later.

  FIG. 2 is a block diagram illustrating a hardware configuration of the user terminal 100. The user terminal 100 includes at least a control unit 110, a storage unit 120, a communication unit 130, a display unit 140, a voice input unit 150, and a voice output unit 160. The user terminal 100 may include a keypad such as a numeric keypad or a vibrator.

The control unit 110 is means for controlling the operation of each unit of the user terminal 100. The control unit 110 includes an arithmetic processing device such as a CPU (Central Processing Unit) and a memory, and performs control by executing a predetermined program.
The storage unit 120 is means for storing data. The storage unit 120 includes a recording medium such as a flash memory, and stores data used by the control unit 110 for control.

The communication unit 130 is means for transmitting and receiving data via the network 300. The communication unit 300 includes an antenna and a modem corresponding to the communication method of the network 300, and executes processing necessary for data communication such as data modulation / demodulation.
The display unit 140 is a means for displaying an image. The display unit 140 includes a display panel (that is, a display area) configured by a liquid crystal element or an organic EL (electroluminescence) element and a drive circuit that drives the display panel, and displays an image according to image data.

  The voice input unit 150 is a means for inputting voice. The audio input unit 150 includes a microphone or an input terminal for connecting the microphone, and supplies audio data to the control unit 110. On the other hand, the audio output unit 160 is means for outputting audio. The audio output unit 160 includes a speaker or an output terminal for connecting the speaker, and outputs audio corresponding to the audio information supplied from the control unit 110. Note that the audio input unit 150 and the audio output unit 160 may be configured to wirelessly connect a headset or the like.

  FIG. 3 is a block diagram showing the hardware configuration of the speech recognition server 200A, the intention interpretation server 200B, the chat management server 200C, and the speech synthesis server 200D. Although these server devices have different data to be stored and processes to be executed, the main hardware configurations are common. Therefore, here, the speech recognition server 200A, the intention interpretation server 200B, the chat management server 200C, and the speech synthesis server 200D are collectively referred to as “server device 200”, and the main hardware configuration common to these server devices is described. explain.

  The server device 200 includes a control unit 210, a storage unit 220, and a communication unit 230. The control unit 210 is means for controlling the operation of each unit of the server device 200. The control unit 210 includes an arithmetic processing unit and a memory, and performs control by executing a predetermined program. The storage unit 220 is a means for storing data, and includes a recording medium such as a hard disk. The communication unit 130 is means for transmitting and receiving data via the network 300.

  FIG. 4 is a block diagram showing a functional configuration of the intention interpretation server 200B. The control unit 210 of the intention interpretation server 200B implements functions corresponding to the respective units of the acquisition unit 211, the interpretation unit 212, and the output unit 213 by executing a predetermined program. These functions realized by the control unit 210 are functions corresponding to the dialogue support apparatus according to the present invention.

  The acquisition unit 211 is means for acquiring input information. Here, the input information is a general term for information input by the user by voice. The input information includes voice data generated in the user terminal 100 and text data obtained by converting the voice data. In the present embodiment, the acquisition unit 211 acquires input information in a state where the voice recognition server 200A converts voice data into text data.

  The interpretation unit 212 is a means for interpreting the user's intention to speak based on the input information acquired by the acquisition unit 211. The interpretation unit 212 performs morphological analysis on the text data to identify words included in the text data, and further interprets the syntax and the like to interpret the intention of the utterance by the user. In this embodiment, the interpretation unit 212 determines whether the user's intention is “information search” or “chat”. This determination is made using chat data.

  The output unit 213 is means for outputting response information corresponding to the interpretation result by the interpretation unit 212. The output unit 213 outputs response information corresponding to the case where the user's intention interpreted by the interpretation unit 212 is a search for information and the case where the intention is a chat. When the user's intention interpreted by the interpretation unit 212 is a search for information, the output unit 213 performs a search using an external search engine or database, receives the search result, and responds according to the search result Output information. On the other hand, when the user's intention interpreted by the interpretation unit 212 is chat, the output unit 213 selectively extracts appropriate data from the plurality of chat data stored in the storage unit 220, and responds accordingly. Response information is output. That is, the output unit 213 changes the information source to be referred to according to the user's intention interpreted by the interpretation unit 212.

  FIG. 5 is a diagram illustrating a data structure of chat data. The chat data of this embodiment includes items such as “registered character string”, “time”, “prohibited term flag”, “character ID”, “response text”, and “action ID”. Among these items, the content provider can register “prohibited term flag”, “character ID”, “response text”, and “action ID”. Other items are determined in advance by a service provider (for example, a communication provider) of this service.

  The “registered character string” is a character string for determining whether or not the intention of the user's utterance is a chat. The intention interpretation server 200B can determine that the intention of the user's utterance is a chat when the registered character string is included in the text data. Registration string, for example, "Good morning", a greeting such as "Hello".

  “Time” is data indicating the time when the user speaks. This item is provided so that the chat response can be changed according to the time zone spoken by the user by describing the time zone such as “morning”, “daytime”, and “night”, for example. In this way, for example, the specific content of the response can be made different when the user utters “good morning” in the morning and when utters “good morning” in the daytime.

  The “prohibited term flag” is data for uniformly designating responses to text data including prohibited terms. When this item is used, it is determined whether or not prohibited terms are included in the text data separately from the determination of the intention of the user's utterance. The intention interpretation server 200B extracts chat data in which the prohibited term flag is “1” when the prohibited term is included in the text data. Here, the prohibited term is, for example, an obscene expression or a violent expression, but can be arbitrarily set by a content provider or a service provider. In addition, as long as it is possible to identify the user's age, the prohibited terms may differ depending on the age.

  The “character ID” is an ID for identifying a character used by the user on the user terminal 100. “Response text” is text data representing a response by the character. The “action ID” is an ID for identifying an action (such as body movement or facial expression) of the character in response. For each character, animation data for animation display corresponding to each action ID is prepared in advance by the content provider.

  The chat management server 200C accepts such chat data registration from the content provider. The chat management server 200C transmits the chat data to the intention interpretation server 200B at a predetermined timing (every hour, once a day, etc.). When the intention interpretation server 200B receives new chat data from the chat management server 200C, the intention interpretation server 200B updates the chat data stored in its own device.

  The configuration of the dialogue system 10 is as described above. Under this configuration, the user terminal 100 displays a character on a standby screen and waits for a user input by voice. The user speaks to the user terminal 100 on which the character is displayed so as to speak to the character, and inputs a voice. When voice is input by the user, the user terminal 100, the voice recognition server 200A, and the intention interpretation server 200B cooperate to execute the following processing.

  FIG. 6 is a sequence chart showing an operation until a response to the user's voice input is obtained. This operation is started when the user inputs voice to the user terminal 100. The user terminal 100 collects the user's voice to generate voice data (step S1), and transmits it to the voice recognition server 200A (step S2). The speech recognition server 200A receives the speech data and executes speech recognition to convert the speech data into text data (data that can be described with a character code) (step S3). The voice recognition server 200A returns (transmits) the text data to the user terminal 100 (step S4).

  When receiving the text data, the user terminal 100 transfers (transmits) the text data to the intention interpretation server 200B (step S5). At this time, the user terminal 100 transmits the character ID of the character displayed on the display unit 140, that is, the character ID of the character selected by the user. The intention interpretation server 200B receives text data as input information, and executes a process of obtaining necessary response information by interpreting the intention of the voice spoken by the user (step S6). This process is hereinafter referred to as “intention interpretation process”.

  The intention interpretation server 200B can obtain response information by executing intention interpretation processing. The intention interpretation server 200B transmits response information obtained by the intention interpretation process to the user terminal 100 (step S7). Note that the response information includes at least a response sentence and an action ID.

  When receiving the response information, the user terminal 100 transmits the response text included in the response information to the speech synthesis server 200D (step S8). At this time, the user terminal 100 transmits the same character ID as the character ID transmitted in step S5. The speech synthesis server 200D can uniquely identify the synthesized speech data based on the received character ID and response text. The speech synthesis server 200D transmits the identified synthesized speech data to the user terminal 100 (step S9). And the user terminal 100 responds with respect to a user's audio | voice input by reproducing | regenerating the audio | voice and image according to response information (step S10).

  FIG. 7 is a flowchart showing in more detail the intention interpretation process in step S6. The intention interpretation server 200B (the control unit 210) first analyzes the text data by morphological analysis or the like, and interprets the intention of the user's utterance (step S61). In addition, what is necessary is just to use the well-known technique regarding a language analysis suitably for the process of step S61. In addition, the intention interpretation server 200B may cause a dedicated device for language analysis to execute part or all of this processing on its own device. In short, the intention interpretation server 200B only needs to acquire the interpretation result of the text data, and its specific method is not limited.

  Next, the intention interpretation server 200B determines whether or not the user's intention indicated by the text data is information retrieval (step S62). Here, the intention interpretation server 200B refers to the chat data stored in its own device, and determines whether or not the user's intention is chat by determining whether or not the registered character string is included in the text data. Can be judged. In addition, the intention interpretation server 200B may determine whether or not it is a chat using a combination other than the registered character string (such as syntax). For example, the intention interpretation server 200B determines that the user's intention is a search for information when the text data includes a predetermined character string meaning a request or a request such as “Tell me” or “Check it”. You may judge. That is, the determination as to whether or not the registered character string is included in the text data is only one specific example of the process of step S62.

  The intention interpretation server 200B changes the subsequent processing according to the determination result of step S62. First, when the intention of the user indicated by the text data is a search for information, the intention interpretation server 200B executes a search using a character string included in the text data (step S63), and receives the search result (step S63). S64). At this time, the intention interpretation server 200B may obtain a search result from an unspecified number of web pages like a so-called search engine, or obtain a search result using a database specialized for information in a specific field. May be. For example, if the user utters the phrase “Tell me an Italian restaurant in Tokyo”, the intention interpretation server 200B may perform a web search using the search terms “Tokyo” and “Italian”. A search condition such as “Tokyo” as a place and “Italian” as a genre may be set, and a service provider database that provides information on restaurants in various parts of the country may be accessed.

  On the other hand, when the intention of the user indicated by the text data is not a search of information but a chat, the intention interpretation server 200B refers to the chat data and chat data corresponding to the text data and the character ID transmitted from the user terminal 100. (In particular, “response text” and “action ID”) are extracted (step S65).

  When these processes are completed, the intention interpretation server 200B generates response information (step S66). In step S <b> 66, the intention interpretation server 200 </ b> B generates, as response information, data for reproducing audio and images on the user terminal 100 based on the chat data and the search result. At this time, the intention interpretation server 200B may process chat data and search results.

  8 and 9 are diagrams illustrating output results in the user terminal 100. FIG. FIG. 8 shows a case where the user speaks the word “Tell me an Italian restaurant in Tokyo” as a question (request). On the other hand, FIG. 9 shows a case where the user utters the word “How are you?” As a chat. If the intention of the user is a search of information, the intention interpretation server 200B notifies the fact that the search has been executed and the search condition (search word) at that time by voice, and displays a response that displays information corresponding to the search result Generate information. On the other hand, when the user's intention is a chat, the intention interpretation server 200B does not perform a web search by using the search term “genki” but reproduces a voice and an image corresponding to the registered character string “genki”. To generate response information.

  Thus, according to the dialogue system 10 of the present embodiment, the response method can be changed according to the intention of the utterance interpreted by the intention interpretation server 200B. This makes it possible for the user to use voice input for purposes other than information retrieval, and to make it feel as if he / she is interacting with a real person without making the response unnatural. be able to.

  Also, according to the present embodiment, when searching for information, the user can obtain a constant response regardless of the character selected by the user, whereas when performing a chat, the user obtains a different response for each character. Can do. By doing so, it is possible to give interest to the selection of the character, evoke the user's interest, and it can be expected that the user will be motivated to use this service. Also, by making the response different for each character, the user can selectively use his / her favorite character, and it can be expected that attachment to a specific character will increase.

[Modification]
The present invention is not limited to the embodiment described above, but can be implemented in other embodiments. The following is an example of another embodiment of the present invention. Note that these modifications can be appropriately combined as necessary.

(1) The response of the present invention may be composed of response information and other information. Here, the other information is information added to the response information, for example, information representing a word that changes the tone (speech tone) such as a ending change. Such information is hereinafter referred to as “additional information”.

  FIG. 10 is a diagram illustrating response information (response text) and additional information. Here, the response information is composed of only the noun “Genki”. In addition, the additional information expresses the phrase following the noun “Genki”. Here, “I'm fine”, “I ’m fine”, “I ’m fine. Thank you.” And a plurality of additional information. At this time, it is assumed that the additional information is different for each character. Note that the response information may or may not differ for each character.

  In this way, the response can be changed by the additional information. Also, by making the response information and the additional information different for each character, for example, while changing the meaning of the response depending on the response information, subtle nuances such as tone are changed by the additional information. A mode of use is also conceivable.

(2) The speech synthesis server 200D may transmit music data to the user terminal 100 instead of the synthesized speech data when the user issues a specific term or asks a specific question. For example, when the user utters “sing”, the speech synthesis server 200D transmits music data of a predetermined music as a response corresponding thereto. The music data may be different depending on the date and time.

  Further, the characters that can be sung in this way may be limited to some characters. In this case, the speech synthesis server 200D cannot properly sing (in other words, for characters that are not set to sing), the synthesized speech data (for example, “ I do n’t know ”,“ I ca n’t sing ”, etc.). In this way, the individuality of each character can be made more prominent.

(3) The intention interpretation server 200B may output predetermined response information without searching for information when a predetermined condition is satisfied. For example, in the intention interpretation server 200B, when the speaking user is not a specific user (for example, the owner of the user terminal 100), the position of the user terminal 100 is not a predetermined position (for example, a position set in advance by the user). When the utterance date / time is a predetermined date / time (for example, midnight), response information determined in advance may be output regardless of the user's intention. In this case, the intention interpretation server 200B is configured to be able to acquire a user recognition result, position information indicating the position of the user terminal 100, time information indicating the utterance date and time, and the like. In this way, it is possible to prevent an unexpected situation where information is retrieved by an unintended third party.

  In addition, what is necessary is just to utilize a well-known speaker recognition technique or register a user's (owner's) voice | voice previously for a user's discrimination | determination. Further, the position of the user terminal 100 may be specified by using a known technique that can acquire position information such as positioning by GPS (Global Positioning System). Further, the response of the character when the predetermined condition is satisfied is, for example, “I don't know well” or “I can't answer”, but it may be different for each character.

  Further, the intention interpretation server 200B may determine whether or not a predetermined condition is satisfied before or after the intention interpretation process. When this determination is executed before the intention interpretation process, the intention interpretation server 200B may omit the intention interpretation process without performing the intention interpretation process when determining that the predetermined condition is satisfied. When the intention interpretation process is executed first, the intention interpretation server 200B only needs to execute this determination when it is determined that the user's intention is information retrieval.

  Furthermore, this modification can also be applied to the case where the speech synthesis server 200D transmits music data in response to a user's inquiry as in the case of the modification 2 described above. For example, if the position of the user terminal 100 is not a predetermined position, the intention interpretation server 200B may transmit synthetic voice data that says “I can't sing here” instead of music data, In the case of midnight, it is also possible to transmit synthetic voice data that says “I can't sing now” instead of the music data.

(4) The present invention does not require that both the input from the user and the output to the user are audio. That is, the present invention may be in a mode in which the output to the user (that is, the response to the input) is only the reproduction (display) of the image and is not accompanied by the reproduction of the sound. In this case, a configuration corresponding to the speech synthesis server 200D is not necessary. Alternatively, the output to the user may be accompanied by tactile feedback such as vibrating the vibrator.

(5) In the present invention, voice input may be used for responses other than information retrieval and chat. For example, the present invention may use voice input for operation of the user terminal 100 in addition to information search and chat. The operation here is, for example, switching on / off of a specific function provided in the user terminal 100, execution of a specific program, or the like.

(6) The specific configuration for carrying out the present invention is not limited to the embodiment described above. For example, voice recognition may be executed in the user terminal 100 or the intention interpretation server 200B instead of the voice recognition server 200A. Further, the synthesized speech data may be transmitted from the intention interpretation server 200B (in step S7) instead of being transmitted from the speech synthesis server 200D. Furthermore, some or all of the functions of the dialogue support apparatus according to the present invention may be provided in the user terminal 100, or may be provided in a plurality of server apparatuses in a distributed manner.

(7) Various terminals other than a smartphone may also correspond to the user terminal of the present invention. For example, the present invention can be applied to a so-called tablet PC (Personal Computer), a game machine, a music player, and the like. In addition, the user terminal of the present invention is not limited to a wireless communication terminal, but may be a desktop PC connected by wire via a modem or the like.

  The present invention is not limited to a dialog support apparatus or a server apparatus (or user terminal) including the dialog support apparatus, but includes a dialog system including one or more server apparatuses and a user terminal, a dialog support method using the dialog support apparatus, It may be specified as a program for causing a computer to function as the dialogue support apparatus of the present invention. Further, the program according to the present invention can be provided in a form recorded on a recording medium such as an optical disk, or a form that is downloaded to a computer via a network such as the Internet and can be installed and used. .

DESCRIPTION OF SYMBOLS 10 ... Dialog system, 100 ... User terminal, 110 ... Control part, 120 ... Memory | storage part, 130 ... Communication part, 140 ... Display part, 150 ... Voice input part, 160 ... Voice output part, 200A ... Voice recognition server, 200B ... Intent interpretation server, 200C ... chat management server, 200D ... speech synthesis server, 210 ... control unit, 211 ... acquisition unit, 212 ... interpretation unit, 213 ... output unit, 220 ... storage unit, 230 ... communication unit, 300 ... network

Claims (8)

An acquisition unit for acquiring input information generated based on voice spoken by the user;
An interpretation unit that interprets the intention of the user's utterance based on the input information acquired by the acquisition unit;
When the intention interpreted by the interpretation unit is a search for information, response information is output according to the search result of the search executed according to the intention, and the intention interpreted by the interpretation unit is a search for information. An output unit that selectively outputs any one of a plurality of response information prepared in accordance with the intention,
The output unit, when intended interpreted by the interpretation unit is a request of singing by characters, when position of the user satisfies a predetermined condition, while outputting music data, position of the user If the predetermined condition is not satisfied, response information indicating a message for rejecting singing is output. The output unit is
The dialogue support apparatus according to claim 1, wherein the information source to be referred to is different depending on whether the intention interpreted by the interpretation unit is a search for information or not. The output unit is
The dialogue support apparatus according to claim 1 or 2, wherein whether or not the intention interpreted by the interpretation unit is a search for information is determined based on whether or not a predetermined character string is included. The output unit is
The dialogue support device according to any one of claims 1 to 3, wherein response information corresponding to a character selected by the user among the plurality of response information is output. The output unit is
The dialogue support device according to any one of claims 1 to 4, wherein the response information and additional information added to the response information and corresponding to the character selected by the user are output. The dialogue support apparatus according to any one of claims 1 to 5,
A dialogue system comprising: a user terminal that picks up the user's voice and reproduces at least the voice according to the response information output by the output unit. Obtaining input information generated based on voice spoken by the user;
Interpreting the user's intention to speak based on the acquired input information;
When the interpreted intention is a search for information, response information corresponding to the search result of the search executed according to the intention is output, and when the interpreted intention is not a search for information, the intention is Selectively outputting one of a plurality of response information prepared corresponding to
In the step of outputting one of the plurality of response information selectively, if the interpreted intended is a request of singing by characters, when position of the user satisfies a predetermined condition, it outputs the music data while, when the position of the user does not satisfy the predetermined condition, interactive support method and outputting the response information indicating the refuse singing message. On the computer,
Obtaining input information generated based on voice spoken by the user;
Interpreting the user's intention to speak based on the acquired input information;
When the interpreted intention is a search for information, response information corresponding to the search result of the search executed according to the intention is output, and when the interpreted intention is not a search for information, the intention is A step of selectively outputting any one of a plurality of response information prepared in correspondence with
In the step of outputting one of the plurality of response information selectively, if the interpreted intended is a request of singing by characters, when position of the user satisfies a predetermined condition, it outputs the music data while, when the position of the user does not satisfy the predetermined condition, a program and outputting the response information indicating the refuse singing message.

Images