JP2003108581A - Interactive information retrieving device and interactive information retrieving method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an interactive information retrieving device and a method for allowing a user to always smoothly continue the interaction without waiting for a specific time in an unexecutable state of smooth input by the user, and capable of retrieving required information on the user even if the device does not freely stand proxy for the user. SOLUTION: An input interval measuring part 7 measures and outputs an input interval up to inputting voice to the device in the next place from the time when the device finishes presenting a response, and since an interaction initiative switching part 9 outputs an operation switching command so as to operate a system leading interaction control part 3 when the input interval exceeds a threshold value, when the unexecutable state of smooth input by the user is caused, a system takes the initiative in the interaction on an after that time, the user can always smoothly continue the interaction without waiting for the specific time, and the user can retrieve the required information even if the device does not freely stand proxy for the user.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an interactive information retrieval method and an interactive information retrieval method for a user to obtain necessary information by interacting with a device.

[0002]

2. Description of the Related Art In recent years, an interactive man-machine interface in which a user interacts with a system and retrieves necessary information from a database is becoming more important. The dialogue progresses when the user and the system speak alternately, but when the user cannot decide what to input next, or when the user does not know the state of the system and does not know what can be input, There is a problem that the next input cannot be made to the system response and the dialogue stops.

Conventionally, in order to deal with the situation where the user cannot react to the response of the system and the dialogue does not proceed smoothly,
A method of presenting a guide for operation to the user or continuing the conversation on behalf of the user has been proposed.

For example, FIG. 46 shows Japanese Patent Laid-Open No. 2000-4803.
It is a block diagram of the natural language dialogue system which is the conventional apparatus of this kind shown by the 8th publication. In the figure, 111 is an input unit, 112 is an output unit, 113 is a dialogue scenario, and 114.
Is current state data, 115 is an input waiting timer, 116 is a proxy input list, 117 is a proxy dialogue unit, and 118 is a dialogue control unit.

In such a natural language dialogue system, the input unit 111 receives a natural language input from the user, and the output unit 112 outputs the output data on the system side. The dialogue scenario 113 includes a state ID, output data,
The current state data 114 is written by describing the scenario of the entire dialogue with the inputtable data and the state ID corresponding to the inputtable data.
The current position in the dialogue scenario is stored in. The input waiting timer 115 waits for the user's input for a certain period of time after presenting the information to the user, and outputs a timeout signal to the proxy dialogue unit 117 if there is no input within that time. When the timeout signal is input, the proxy dialogue unit 117 issues an input sentence for controlling the dialogue on behalf of the user and inputs the sentence to the dialogue control unit 118. In the proxy input list 116,
Data to be entered on behalf of the user when the user is unresponsive is stored in association with each state of the dialogue scenario.

In the natural language dialogue system configured as described above, when the user cannot respond to the response from the system and becomes unresponsive, the system acts on behalf of the user and acts as a substitute for the dialogue. , It becomes possible to continue the dialogue smoothly.

[0007]

However, in the conventional system as described above, the input waiting timer always waits for a certain period of time even if the user cannot smoothly input data for several times. Since the section does not work, there is a problem that the user is constantly kept waiting for a certain period of time and the convenience is low.

Further, in a situation in which the user cannot smoothly input, the proxy dialogue section pretends to be the user by the input specified in advance in the proxy input list, and the dialogue continues without permission, so that the dialogue continues. However, there is a problem that the information that the user wants is not always obtained.

The present invention has been made in order to solve the above problems, and in a situation where the user cannot smoothly input, the user can always continue the conversation smoothly without waiting for a certain period of time. An object of the present invention is to obtain an interactive information search device and an interactive information search method that can search for information required by a user without the device arbitrarily acting on behalf of the user.

[0010]

In view of the above-mentioned object, the present invention recognizes the voice of the user input to the apparatus and outputs the recognition result, and the voice recognition section inputs the voice. A user-initiated dialogue management unit that generates and outputs an information search request from the voice recognition result, and generates and outputs a response content to the user from the search result input from the information search unit described later, and the voice recognition unit. An information search request is generated and output from the input voice recognition result, a response content to the user is generated and output from the search result input from the information search unit described later, and a question content generation knowledge storage unit described later is generated. And a system-initiated dialogue management unit for generating and outputting a question content to the user by referring to a dialogue context storage unit, which will be described later, and a question content for generating the question content to the user in the system-initiated dialogue management unit. Generate Knowledge storage unit that stores knowledge, the dialogue context storage unit that stores the voice recognition result, the search result, the response content, and the question content, the user-initiated dialogue management unit, or the system initiative. The information search unit that performs information search in response to the information search request input from the dialog management unit and outputs the search result, and the response content and question input from the user-initiated dialog management unit or the system-initiated dialog management unit A response presentation unit that presents the contents to the user, an input interval measurement unit that measures and outputs an input interval from the time when the response presentation unit presents the response to the time when the voice is input to the device, and the input A dialog initiative switching unit that outputs an operation switching command for the user-initiated dialog management unit and the system-initiated dialog management unit based on an interval. In the interactive information retrieval apparatus that.

The input interval measuring unit measures and outputs the input interval from the time when the response presenting unit finishes presenting the response to the time when the voice is input to the device, and the dialogue initiative switching. The section outputs an operation switching command to operate the system-initiated dialogue management section when the input interval exceeds a threshold value.

Further, the dialogue initiative switching section outputs an operation switching command to temporarily operate the system initiative dialogue management section when the input interval exceeds the threshold, and decreases the threshold. It is characterized by

Further, an input interval storage unit for storing the input interval is further provided, and the dialogue initiative switching unit refers to the input interval storage unit, and when an input interval exceeding a threshold value occurs a prescribed number of times or more. The operation switching command is output so as to operate the system-initiated dialogue management unit.

Further, the input interval measuring unit may notify the device from the time when the response presenting unit starts presenting a response when voice is input to the device before the response presenting unit finishes presenting the response. The interrupt input interval up to the time when the voice is input is measured and output, and the dialog initiative switching unit operates the operation switching signal to operate the user-initiated dialog management unit when the interrupt input interval is shorter than the threshold value. Is output.

Further, the input interval storage unit stores the interrupt input interval, and the dialogue initiative switching unit refers to the input interval storage unit, and when an interrupt input interval shorter than a threshold value occurs a prescribed number of times or more. In addition, an operation switching command is output to operate the user-initiated dialogue management unit.

Further, the user-initiated dialogue management section further comprises a user-initiated speech dictionary in which utterance contents for switching the operation are registered, and the dialogue initiative switching section inputs the voice when the interrupt input interval is input. When the voice recognition result input from the recognition unit matches the utterance content registered in the user-initiated utterance dictionary, an operation switching command is output to operate the user-initiated dialogue management unit.
It is characterized by

Further, the input interval measuring unit outputs a response stop command to the response presenting unit when the voice is input to the device before the response presenting unit finishes presenting the response, and the response presenting unit Is characterized in that when the response stop command is input, the response presentation process being output is stopped.

Further, with reference to the meaningless utterance dictionary in which meaningless utterance contents relating to the operation of the device and the meaningless utterance dictionary are referred to, whether the voice recognition result input from the voice recognition unit is meaningless utterance or not. A nonsense utterance detection unit that detects whether or not
Further, the dialog initiative switching unit outputs an operation switching command to operate the system-initiated dialog management unit when the meaningless utterance detection unit detects a meaningless utterance. .

Further, a meaningless utterance number storage unit for storing the number of the unconscious utterances is further provided, and the dialogue initiative switching unit refers to the meaningless utterance number storage unit, and the meaningless utterance count exceeds a threshold value. In this case, the operation switching command is output so as to operate the system-initiated dialogue management unit.

Further, referring to the dialogue context storage unit, it is judged whether or not the voice recognition result input from the voice recognition unit is a repeat of the response content or the question content presented by the device immediately before. Further comprising a unit, the dialogue initiative switching unit outputs an operation switching command to temporarily operate the user-initiated dialogue management unit when the replay utterance detection unit detects a recitation utterance. And

Further, a predictive utterance knowledge storage unit defining the utterance content of the user predicted for the question contents generated by the system-initiated dialogue management unit, the dialogue context storage unit, and the predictive utterance knowledge storage unit are provided. With reference to, when the device presents a question to the user, a predictive utterance determination unit that determines whether or not the voice recognition result input from the voice recognition unit is a predictive utterance for the question content of the device. Furthermore, the dialogue initiative switching unit, when the voice recognition result is not determined to be a predicted utterance in the predicted utterance determination unit,
It is characterized in that an operation switching command is output so as to operate the user-initiated dialogue management unit.

Further, the predictive utterance determining unit further includes a non-predictive utterance number storage unit that stores the number of times the voice recognition result is determined not to be the predictive utterance, and the dialogue initiative switching unit includes the non-predictive utterance. When the number of times stored in the number-of-times storage unit exceeds a threshold value, an operation switching command is output to operate the user-initiated dialogue management unit.

Further, a voice recognition step of recognizing a user's voice input to the apparatus and outputting a recognition result, and an information retrieval request generated and output from the voice recognition result input from the voice recognition step, A user-initiated dialogue management step of generating and outputting a response content to a user from a search result input from an information search step described later, and an information search request from the voice recognition result input from the voice recognition step. Output,
The content of the response to the user is generated and output from the search result input from the information search step described later, and the question content to the user is referred by referring to the question content generation knowledge storage step described later and the dialogue context storage step described later. A system-initiated dialogue management step for generating and outputting a question content generation knowledge storage step for storing question content generation knowledge for generating question content for a user in the system-initiated dialogue management step, and the voice recognition An information search is performed in response to an information search request input from the dialog context storing step of storing the result, the search result, the response content, and the question content, the user-initiated dialog management step or the system-initiated dialog management step. The information search step of performing and outputting the search result, and the response input from the user-initiated dialogue management step or the system-initiated dialogue management step A response presenting step of presenting the content and question content to the user, and an input interval measuring step of measuring and outputting an input interval from the time when the response is presented by the response to the time when the voice is input to the device. An interactive information retrieval method comprising: a user-initiated dialogue management step and a dialogue initiative control step of outputting an operation switching command for the system-initiated dialogue management step based on the input interval.

In the input interval measuring step, the input interval from the time when the response presenting step has finished presenting the response to the time when the voice is input to the device is measured and output, and the dialogue initiative switching is performed. The step is characterized in that when the input interval exceeds a threshold value, an operation switching command is output to operate the system-initiated dialogue management step.

Further, in the dialog initiative switching step, when the input interval exceeds a threshold value, an operation switching command is temporarily output to operate the system initiative dialog management step, and the threshold value is decreased. It is characterized by

Further, the method further comprises an input interval storing step of storing the input interval, and the dialogue initiative switching step,
With reference to the input interval storing step, when the input interval exceeding the threshold value occurs a prescribed number of times or more, an operation switching command is output to operate the system-initiated dialogue management step,
It is characterized by

In the input interval measuring step, when a voice is input to the device before the response presenting step finishes presenting the response, the device starts from the time when the response presenting step starts presenting the response. The interrupt input interval up to the time when the voice is input is measured and output, and the dialogue initiative switching step causes the user-initiated dialogue management step to operate when the interrupt input interval is shorter than a threshold value. Is output.

Further, the input interval storing step stores the interrupt input interval, and the dialogue initiative switching step,
With reference to the input interval storing step, when an interrupt input interval shorter than a threshold value occurs a prescribed number of times or more, an operation switching command is output so as to operate the user-initiated dialogue management step.

Further, the user-initiated dialogue management step further comprises a user-initiated utterance dictionary in which the utterance content for switching the operation is registered, and in the dialogue initiative switching step, when the interrupt input interval is input, the voice When the voice recognition result input from the recognition step matches the utterance content registered in the user-initiated utterance dictionary, an operation switching command is output to operate the user-initiated dialogue management step. .

The input interval measuring step outputs a response stop command to the response presenting step when a voice is input to the device before the response presenting step finishes presenting the response, and the response presenting step. Is characterized in that when the response stop command is input, the response presentation process being output is stopped.

Further, with reference to the meaningless utterance dictionary in which meaningless utterance contents regarding the operation of the apparatus and the meaningless utterance dictionary are referred, whether the voice recognition result input in the voice recognition step is meaningless utterance or not. Further comprising a meaningless utterance detection step of detecting whether or not, the dialogue initiative switching step, when the meaningless utterance detection step detects a meaningless utterance,
It is characterized in that an operation switching command is output so as to operate the system-initiated dialogue management process.

Further, the method further comprises a meaningless utterance number storing step of storing the number of times of the insignificant utterance, and the dialogue initiative switching step refers to the meaningless utterance number storing step, and the meaningless utterance number exceeds a threshold value. In this case, an operation switching command is output to operate the system-driven dialogue management process.

Further, referring to the dialog context storing step,
The voice recognition result input from the voice recognition step further comprises a repeat utterance detection step of determining whether or not the response content or question content presented by the device immediately before is repeated. When the utterance detection step detects a repeat utterance, the operation switching command is output so as to temporarily operate the user-initiated dialogue management step.

Further, a predictive utterance knowledge storing step that defines the user's utterance content predicted for the question content generated by the system-initiated dialog managing step, the dialog context storing step, and the predictive utterance knowledge storing step. With reference to, when the device presents a question to the user, a predictive utterance determination process of determining whether or not the voice recognition result input from the voice recognition process is a predictive utterance for the question content of the device. Furthermore, the dialog initiative switching step outputs an operation switching command to operate the user-initiated dialog management step when the speech recognition result is not determined to be a predicted utterance in the predicted utterance determination step, Is characterized by.

Further, the method further comprises a non-predicted utterance number storing step of storing the number of times that the speech recognition result is determined not to be the predictive utterance in the predictive utterance determination step, and the dialogue initiative switching step includes the non-predicted utterance. When the number of times stored in the number of times storage step exceeds a threshold value, an operation switching command is output to operate the user-initiated dialogue management step,
It is characterized by

[0036]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention will be described below according to each embodiment. Embodiment 1. 1 is a block diagram of an interactive information search device according to a first embodiment of the present invention. In FIG. 1, reference numeral 1 is a voice recognition unit that recognizes a user's voice input to the device and outputs a recognition result, and 2 is an information retrieval request generated and output from the voice recognition result input from the voice recognition unit 1. Then, the user-initiated dialogue management unit 3 that generates and outputs the content of the response to the user from the search result input from the information search unit 6 to be described later is an information search request from the voice recognition result input from the voice recognition unit 1. Is generated and output, and the content of the response to the user is generated and output from the search result input from the information search unit 6 described later, and the content of the question to the user is referred to by referring to the dialogue context storage unit 5 described later. It is a system-driven dialogue management unit that generates and outputs.

Reference numeral 4 denotes a question content generation knowledge storage unit that stores question content generation knowledge for generating question content to the user in the system-initiated dialogue management unit 3, and reference numeral 5 denotes the voice recognition result, the search result, and the response. A dialogue context storage unit for storing the contents and the question contents, and 6 is information for performing an information search with respect to an information retrieval request input from the user-initiated dialogue management unit 2 or the system-initiated dialogue management unit 3 and outputting a retrieval result. It is a search unit.

Reference numeral 8 is a response presentation unit for presenting the response content and question content input from the user-initiated dialogue management unit 2 or the system-initiated dialogue management unit 3 to the user, and 7 is from the time when the response presentation unit 8 presents the response. , An input interval measuring unit that measures and outputs an input interval until the time when voice is input to the device, 9
Is a dialog initiative switching unit that outputs an operation switching command for the user-initiated dialog management unit 2 and the system-initiated dialog management unit 3 based on the input interval.

The specific operation of the present invention when used in an accommodation facility information retrieval device by telephone will be described below.
The accommodation facility information retrieval device by telephone is a device in which a user makes a telephone call to a retrieval service and inputs conditions such as a place and an amount of money to retrieve a desired accommodation facility. For the example of the interaction between the device and the user shown in FIG. 2, the threshold of the input interval is 5 seconds,
A case will be described in which the user-initiated dialogue management unit 2 is operating at the beginning of the dialogue.

The present embodiment describes the case where the present invention is applied to the accommodation facility information retrieval device by telephone. However, the present invention is not limited to the accommodation facility information retrieval device by telephone, and the railway time. It can be used for other interactive information retrieval devices such as a table retrieval device, a flight schedule retrieval device, a house information retrieval device, and a stock price information retrieval device.

When the user makes a phone call, the device says "Yes,
This is an accommodation facility search service. "
Output 11 and start the dialogue. The input interval measuring unit 7 measures the elapsed time from the time 10:34:25 when the output of the utterance SYS11 is completed, and outputs it to the dialogue initiative switching unit 9. When the user utters USR12 saying "I'm looking for it near Yokohama Station" at 10:34:28,
The utterance of the user is input to the voice recognition unit 1 and the input interval measurement unit 7. The input interval measuring unit 7 completes the measurement of the elapsed time when the utterance USR 12 is input, and the elapsed time output from the input interval measuring unit 7 stops at 3 seconds. Since this input interval does not exceed the threshold of 5 seconds, the dialog initiative switching unit 9 does not perform the switching operation, and the user-initiated dialog management unit 2 continues to operate.

When the user's utterance USR12 is input to the voice recognition unit 1, voice recognition is performed, and the recognition result shown in FIG. 3 is input to the user-initiated dialogue management unit 2. In FIG.
“Search request” represents the meaning of the input utterance, and “{place = near Yokohama station}” represents a parameter necessary for the search operation.

When the voice recognition result is input, the user-initiated dialogue management unit 2 generates a search expression in SQL, for example, and outputs it to the information search unit 6. The information retrieval unit 6 retrieves information from a database (not particularly shown) using the input retrieval formula, and, for example, 10 accommodation facility names as shown in FIG. Output.

When the search result shown in FIG. 4 is input, the user-initiated dialogue management unit 2 creates the response content shown in FIG. The response presenting unit 8 presents the input response contents to the user as a utterance SYS13 by "speech synthesis""10 items such as Yokohama Bay Sheraton and Yokohama Tokyu Hotel meet the condition."

The input interval measuring unit 7 measures the elapsed time from the time 10:34:39 when the output of the utterance SYS 13 is completed, and outputs it to the dialogue initiative switching unit 9. Time 10
If the user does not input the next time even after 34 hours and 44 minutes, the elapsed time output from the input interval measuring unit 7 becomes 5 seconds or more, which exceeds the threshold of the input interval. Reference numeral 9 switches the operation from the user-initiated dialogue management unit 2 to the system-initiated dialogue management unit 3.

FIG. 6 shows an example of the question content generation knowledge stored in the question content generation knowledge storage unit 4. A list of conditions that can be specified by the information search unit 6 is held as the question content generation knowledge. The system-initiated dialogue management unit 3 refers to the question content generation knowledge of FIG. 6 and the dialogue context shown in FIG. 7 stored in the dialogue context storage unit 5, and refers to, for example, a search condition “that the user has not yet entered”. 8 is generated and output to the response presenting unit 8. The response presenting unit 8 presents the question content to the user as a utterance SYS14 "Do you have a request for a budget?" By voice synthesis.

The user utters "about 10,000 yen" US
When R15 is input, the voice recognition unit 1 performs voice recognition, and the recognition result shown in FIG.
Entered in. The system-initiated dialogue management unit 3 generates a search formula based on the voice recognition result and outputs it to the information retrieval unit 6, and the information retrieval unit 6 outputs the retrieval result as shown in FIG. Is entered.

When the search result shown in FIG. 10 is input, the system-initiated dialogue management section 3 generates the response content shown in FIG. 12 and outputs it to the response presentation section 8. Furthermore, the system-initiated dialogue management unit 3 refers to the dialogue context shown in FIG. 11 stored in the dialogue context storage unit 5 and, for example, as a question content regarding the search condition “room type” that the user has not yet input, The question content shown in 13 is generated and output to the response presentation unit 8. The response presenting section 8 synthesizes the input response content and question content by voice synthesis, and “There are 5 cases that meet the conditions, such as Yokohama Beside Hotel and Hotel Sunport.”
Utterance "Do you have a request for a room type?"
16 and SYS 17 are presented to the user.

By the above operation, the input interval measuring unit 7 measures and outputs the input interval from the time when the device finishes providing the response to the next voice input to the device, and the input interval becomes the threshold value. When it exceeds the limit, the dialogue initiative switching unit 9 outputs an operation switching command to operate the system initiative dialogue management unit 3. Therefore, if a situation occurs in which the user cannot input smoothly, the device will be changed thereafter. Taking the initiative in dialogue, the dialogue continues smoothly without the user having to wait for a certain period of time, and the user can search for the necessary information without the device acting on its behalf. Become.

In the present embodiment, an example in which the present invention is realized as an interactive information retrieval device has been described. However, it is configured by software having the same function as each functional block,
The computer-readable software recording medium may be executed by any computer. The same applies below.

Embodiment 2. The second embodiment is different from the above-described first embodiment in the dialog initiative switching unit 9 in FIG. 1, and is otherwise the same as the first embodiment. Therefore,
Since it basically has the configuration of FIG. 1, reference is made to FIG. Hereinafter, the dialogue initiative switching unit 9 in the present embodiment will be described.

In the example of the dialog between the device and the user shown in FIG. 14, the initial value of the threshold value of the input interval is 4 seconds, the dialog initiative switching unit 9 reduces the threshold value to half, and the user-initiated dialog management at the start of the dialog. A case where the unit 2 is operating will be described.

When the user makes a phone call, the device says "Yes,
This is an accommodation facility search service. "And outputs a response. Thereafter, up to the utterance SYS23, the same operation as in the first embodiment is performed and the dialogue proceeds.

After the utterance SYS23 is output, the input interval measuring unit 7 measures the elapsed time from time 10:34:42 and outputs it to the dialogue initiative switching unit 9. If the user does not input the next input even after 10:34:46, the elapsed time output from the input interval measurement unit 7 becomes 4 seconds or more, which exceeds the threshold of the input interval. The switching unit 9 temporarily switches the operation from the user-initiated dialogue management unit 2 to the system-initiated dialogue management unit 3. Further, the threshold value of the input interval is reset to 2 seconds which is half of 4 seconds.

The system-initiated dialogue management unit 3 stores the question content generation knowledge of FIG. 6 and the dialogue context storage unit 5 of FIG.
For example, the content of the question regarding the search condition “place” that the user has not yet input is generated and output to the response presenting unit 8 with reference to the dialogue context shown in FIG. The response presenting unit 8 presents the utterance SYS 24 to the user by voice synthesis of the question content.
After that, up to the utterance SYS26, the same operation as that in the first embodiment is performed and the dialogue proceeds. When the utterance SYS 26 is output, the dialog initiative switching unit 9 switches the operation from the temporary operation of the system initiated dialog management unit 3 to the user initiated dialog management unit 2.

After the utterance SYS 26 is output, the input interval measuring unit 7 measures the elapsed time from time 10:34:59 and outputs it to the dialogue initiative switching unit 9. If the user does not input the next time even after 10:35:01, the elapsed time output from the input interval measuring unit 7 becomes 2 seconds or more, which exceeds the threshold of the input interval. The switching unit 9 temporarily switches the operation from the user-initiated dialogue management unit 2 to the system-initiated dialogue management unit 3. Further, the threshold value of the input interval is reset to 1 second which is half of 2 seconds.

The system-initiated dialogue management unit 3 stores the question content generation knowledge of FIG. 6 and the dialogue context storage unit 5 of FIG.
With reference to the dialogue context shown in (1), for example, the content of the question regarding the search condition “budget” that the user has not yet input is generated and output to the response presenting unit 8. The response presentation unit 8 presents the utterance SYS 27 to the user by voice synthesis of the question content.

By the above operation, the input interval measuring unit 7 measures and outputs the input interval from the time when the device finishes presenting the response to the next voice input to the device, and the input interval becomes the threshold value. When exceeding the threshold, the dialogue initiative switching section 9 outputs an operation switching command to temporarily operate the system initiative dialogue management section 3, and further decreases the threshold value.
Basically, the user has the initiative to proceed with the dialogue, but if a situation occurs where the user cannot input smoothly, the device temporarily takes the initiative in the dialogue and continues the dialogue smoothly. Further, when such a situation occurs repeatedly, it becomes possible for the device to more quickly take the initiative and smoothly continue the dialogue.

Third Embodiment FIG. 17 is a block diagram of an interactive information search device according to the third embodiment of the present invention. The third embodiment is different from the above-described first embodiment in an input interval storage unit 10 and a dialogue initiative switching unit 9, and is otherwise the same as the first embodiment. Therefore, the operation of the input interval storage unit 10 and the dialog initiative switching unit 9 in the present embodiment is performed with the input interval threshold value of 3 seconds and the input interval exceeding the threshold value in the interaction example of the apparatus and the user shown in FIG. A case will be described in which the prescribed number of times is two, and the user-initiated dialogue management unit 2 is operating at the beginning of the dialogue.

When the user makes a call, the device says "Yes,
This is an accommodation facility search service. "And outputs a response. The input interval measuring unit 7 measures the elapsed time from the time when the output of the response is completed at 10:34:25 and outputs it to the dialogue initiative switching unit 9. When the user's utterance USR32 is input at 10:34:30,
The input interval measurement unit 7 completes the measurement of the elapsed time, the input interval of 5 seconds is output to the input interval storage unit 10, and is stored in the input interval storage unit 10 as shown in FIG.

The number of input intervals exceeding the threshold of 3 seconds is one, and the number of input intervals exceeding the threshold is the specified number of times 2.
Since the number of times is not exceeded, the dialog initiative switching unit 9 does not perform the switching operation, and the user-initiated dialog management unit 2 continues to operate. After that, the same operation is performed up to the utterance SYS 35 of the apparatus to proceed with the dialogue. The input interval stored in the input interval storage unit 10 at the time when the device utterance SYS 35 is output is as shown in FIG.

After the utterance SYS 35 is output, the input interval measuring unit 7 measures and outputs the elapsed time from time 10:34:51. If the user does not make the next input after the time of 10:34:54, the elapsed time output from the input interval measurement unit 7 becomes 3 seconds or more, and the input interval storage unit 10 shown in FIG. Since the input interval exceeding the threshold value of 3 seconds exceeds the stored input interval twice, the dialog initiative switching unit 9 switches the operation from the user-initiated dialog management unit 2 to the system-initiated dialog management unit 3.

The system-initiated dialogue management unit 3 stores the question content generation knowledge of FIG. 6 and the dialogue context storage unit 5 of FIG.
For example, the content of the question regarding the search condition “place” that the user has not yet input is generated and output to the response presenting unit 8 with reference to the dialogue context shown in FIG. The response presenting unit 8 presents the utterance SYS 36 to the user by voice synthesis of the question content.

By the above operation, the input interval measuring unit 7 measures and outputs the input interval from the time when the device finishes presenting the response until the next voice is input to the device. The input interval storage unit 10 stores the input interval measured by, and when the input interval exceeding the threshold value occurs a prescribed number of times or more, the dialog initiative switching unit 9 causes the system-initiated dialog management unit 3 to operate. To output the message, the user basically takes the initiative to proceed with the dialogue.
When the situation in which the user cannot input smoothly occurs more than the specified number of times, the device can take the initiative in the dialogue and smoothly continue the dialogue.

Fourth Embodiment 22 is a block diagram of an interactive information retrieval apparatus according to Embodiment 4 of the present invention. The fourth embodiment is different from the first embodiment in an input interval measuring unit 7 and a dialogue initiative switching unit 9, and is otherwise the same as the first embodiment. Therefore, regarding the operation of the input interval measuring unit 7 and the dialog initiative switching unit 9 in the present embodiment, the threshold value of the interrupt input interval is 2 seconds for the dialog example between the device and the user shown in FIG. A case where the system-initiated dialogue management unit 3 is operating will be described as an example.

When the user makes a phone call, the device says "Yes,
This is an accommodation facility search service. "And outputs a response. Since nothing is stored in the dialogue context storage unit 5, the system-initiated dialogue management unit 3 refers to the question content generation knowledge in FIG. 6 to generate a question content for the search condition “place” and respond to it. Output to the presentation unit 8.
The response presentation unit 8 utters the question contents by voice synthesis SYS4.
Present 2 to the user. Then, the device utters SYS45
Up to the point of time when the output of (1) is started, the same operation as in the first embodiment is carried out and the dialogue proceeds.

When the response presenting unit 8 starts outputting the utterance SYS 45 at the time 10:34:38, the input interval measuring unit 7
Starts measurement of the elapsed time from the output start time until the user's voice is input. The response presenting unit 8 makes the utterance SYS4.
The output of all 5 is finished at 10:34:41 (Fig. 23).
It is assumed that the user performs the next utterance USR46 at the time 10:34:39 when the utterance SYS45 of the apparatus does not complete the output. In this case, the input interval measuring unit 7 outputs the utterance SYS45.
The elapsed time of 1 second from the start of the utterance USR46 to the input of the utterance USR46 is output to the dialogue initiative switching unit 9 as an interrupt input interval, and a response stop instruction is output to the response presenting unit 8. The dialog initiative switching unit 9 outputs an operation switching signal to operate the user-initiated dialog management unit 2 because the input interrupt input interval is 1 second and is shorter than the threshold value. When the response presentation unit 8 receives the response stop command at 10:34:39, it stops the presentation process of the utterance SYS45. Therefore, the response in the parentheses of the utterance SYS45 is not presented to the user.

When the utterance USR46 of the user is input,
The user-initiated dialogue management unit 2 operates similarly to the first embodiment, outputs the utterance SYS 47, and waits for the user's utterance to be input. The user inputs the utterance USR 48, and the device operates in the same manner as in the first embodiment to continue the dialogue.

With the above operation, when the voice is input to the device before the response presenting unit 8 finishes presenting the response, the input interval measuring unit 7 starts from the time when the response presenting unit 8 starts presenting the response. The interrupt input interval until the time when the voice is input to the device is measured and output, and a response stop command is further output to the response presentation unit 8. When the response stop command is input, the response presentation unit 8 outputs the response. When the response presentation process is stopped and the interrupt input interval is shorter than the threshold value, the dialog initiative switching unit 9 outputs an operation switching signal to operate the user-initiated dialog management unit 2. In a situation where the user has a clear understanding, the user's request can be entered by blocking questions and guidance from the device, and after that, the user can take the initiative and continue the dialogue smoothly. Yes To become.

Embodiment 5. FIG. 24 shows a block diagram of an interactive information retrieval apparatus according to Embodiment 5 of the present invention. The fifth embodiment is the same as the fourth embodiment and the input interval storage unit 1.
0 is different from the dialogue initiative switching unit 9, and the others are the same as those in the fourth embodiment. Therefore, regarding the operations of the input interval storage unit 10 and the dialogue initiative switching unit 9 in the present embodiment, the interrupt input interval threshold value is 2 seconds and the interrupt input is shorter than the threshold value in the interaction example between the device and the user shown in FIG. An example will be described in which the specified number of intervals is once, and the system-initiated dialogue management unit 3 is operating at the beginning of the dialogue.

When the user makes a call, the device speaks SY.
Output S51 to start the dialogue. The input interval measuring unit 7 starts measuring the elapsed time from the output start time of the utterance SYS 51 until the user's voice is input. Response presenting section 8
It finishes outputting all utterance SYS51 at 10:34:25 (time in parentheses of utterance SYS51 in FIG. 25). Here, the user outputs utterance USR52 and utterance S
The time when the output of YS51 is not completed at 10:34:23
Suppose you went in seconds. In this case, the input interval measuring unit 7 outputs the elapsed time of 1 second from the start of the utterance SYS51 to the input of the utterance USR52 as the interrupt input interval to the dialogue initiative switching unit 9 and the input interval storage unit 10, and stops the response. The command is output to the response presentation unit 8. The interrupt input interval is stored in the input interval storage unit 10 as shown in FIG.

The response presentation unit 8 issues a response stop command at 10:34.
When it is received in 23 minutes, the presentation process of the utterance SYS 51 is stopped. Therefore, the response in the parentheses of the utterance SYS51 is not presented to the user.

With respect to the user's utterance USR52, the response presenting unit 8 operates in the same manner as in the first embodiment and the utterance SYS53.
Is presented to the user. Further, the system-initiated dialogue management unit 3 refers to the dialogue context of FIG. 7 and the question content generation knowledge of FIG. 6 to generate the question content regarding the search condition “budget” and outputs it to the response presentation unit 8. The response presenting unit 8 outputs the utterance SYS5.
Present 4 to the user.

When the response presenting unit 8 starts outputting the utterance SYS 54 at time 10:34:34, the input interval measuring unit 7
Starts measurement of the elapsed time from the output start time until the user's voice is input. The response presenting unit 8 makes the utterance SYS5.
The output of all 4 is finished at 10:34:37 (Fig. 25).
It is assumed that the user performs the next utterance USR55 at a time 10:34:35 when the output of the utterance SYS54 of the apparatus is not completed. In this case, 1 second is output to the dialogue initiative switching unit 9 as the interrupt input interval, and a response stop instruction is output to the response presenting unit 8. The interrupt input interval of 1 second is shorter than the threshold value, and since the interrupt input interval shorter than the threshold value is stored once in the input interval storage unit 10, the interrupt input interval shorter than the threshold value exceeds the specified number of times. For this reason,
The dialog initiative switching unit 9 outputs an operation switching signal to operate the user-initiated dialog management unit 2. Also,
When the response presentation unit 8 receives the response stop command at 10:34:35, it stops the presentation process of the utterance SYS 54. Therefore, the response in the parentheses of the utterance SYS 54 is not presented to the user.

When the utterance USR55 of the user is input,
The user-initiated dialogue management unit 2 operates similarly to the first embodiment, outputs the utterance SYS 56, and waits for the user's utterance to be input. The user inputs the utterance USR57, and the device operates similarly to the first embodiment to continue the dialogue.

By the above operation, when the voice is input to the device before the response presenting unit 8 finishes presenting the response, the input interval measuring unit 7 starts from the time when the response presenting unit 8 starts presenting the response. The interrupt input interval until the time when the voice is input to the device is measured and output, and further, a response stop command is output to the response presentation unit 8, and the interrupt input interval measured by the input interval measurement unit 7 is input interval storage unit 10. When the response stop command is input, the response presenting unit 8 stops the response presenting process being output, and when the interrupt input interval shorter than the threshold occurs more than the specified number of times, the dialogue initiative switching unit 9 Since the operation switching signal is output so as to operate the user-initiated dialogue management unit 2, in a situation where the user clearly has a search condition to be input to the device, questions and guidance from the device can be blocked. It is possible to input the request of the person, even after the user makes it possible to continue the smooth dialogue to take the initiative.

Sixth Embodiment 27 is a block diagram of an interactive information retrieval apparatus according to Embodiment 6 of the present invention. The sixth embodiment is the same as the fifth embodiment except that the function of the dialogue initiative switching unit 9 and a new user-led vocabulary dictionary 11 are newly provided. Therefore, regarding the operation of the dialogue initiative switching unit 9 in the present embodiment, for the dialogue example between the device and the user shown in FIG. 28, the system led dialogue management unit 3 is operating at the beginning of the dialogue, and the user led utterance dictionary. A case where the content of 11 is shown in FIG. 29 will be described as an example.

When the user makes a call, the device speaks SY.
S61 is output to start the dialogue. User is SYS61
Uttered at 10:34:23, the time before the output of
When SR62 is input, the input interval measuring unit 7 operates in the same manner as in the fifth embodiment, and 1 second is set as the interrupt input interval.
The dialogue initiative switching unit 9 outputs the response stop instruction to the response presenting unit 8. The response presentation unit 8 issues a response stop command to 1
When received at 0:34:23, the presentation process of the utterance SYS 61 is stopped, and the response in the parentheses of the utterance SYS 61 is not presented to the user.

The voice recognition unit 1 performs voice recognition when the user's utterance USR 62 is input, and outputs a "utterance stop request" as a recognition result. When the interrupt input interval is input from the input interval measuring unit 7, the dialogue initiative switching unit 9 determines whether or not the voice recognition result is registered in the user-directed utterance dictionary 11. The recognition result "speak stop request" is shown in Fig. 2.
Since it is registered in the user-initiated speech dictionary 11 shown in FIG. 9, the dialog initiative switching unit 9 outputs an operation switching command to temporarily activate the user-initiated dialog management unit 2.

When the user's utterance USR63 is input,
The system-initiated dialogue management unit 3, the information retrieval unit 6, and the response presentation unit 8 operate in the same manner as in the first embodiment, and the device utterance SYS
64 is presented to the user.

With the above operation, when a voice is input to the device before the response presenting unit 8 finishes presenting the response, the input interval measuring unit 7 starts from the time when the response presenting unit 8 starts presenting the response. To measure and output the interrupt input interval until the time when the voice is input to the device, further output the response stop command to the response presentation unit 8, and switch the operation to the user-initiated dialogue management unit 2 in the user-initiated speech dictionary 11. When the interrupt input interval is input, the dialogue initiative switching unit 9 registers the speech recognition result input from the speech recognition unit 1 as the user-initiated speech registered in the user-initiated speech dictionary 11. In the situation where the user clearly has the search condition to input to the device, since the operation switching command is output so as to temporarily operate the user-initiated dialogue management unit 2 when And put these questions and guidance intercepts explicitly, the user can take the initiative of dialogue, it is possible to continue the smooth dialogue.

Seventh Embodiment FIG. 30 shows a block diagram of an interactive information search device according to a seventh embodiment of the present invention. The seventh embodiment is the same as the first embodiment except that it has a function of the dialogue initiative switching unit 9 and that a meaningless utterance dictionary 13 and a meaningless utterance detection unit 12 are newly provided. is there. Therefore, the operations of the dialogue initiative switching unit 9 and the meaningless utterance detecting unit 12 in the present embodiment are
An example of a dialogue between the device shown in FIG. 31 and the user will be described by taking the case where the contents of the meaningless utterance dictionary 13 are as shown in FIG. 32 and the user-initiated dialogue management unit 2 is operating at the start of the dialogue.

When the user makes a call, the device speaks SY.
Output S71 to start the dialogue. User speaks USR
When 72 is input, the voice recognition unit 1 performs voice recognition and outputs "unnecessary word" as a recognition result.

When the meaningless utterance detection unit 12 receives the speech recognition result "unnecessary word", it refers to the meaningless speech dictionary 13 shown in FIG. 32, and the speech recognition result "unnecessary word" is meaningless speech. To determine.

When the meaningless utterance detection unit 12 detects a meaningless utterance, the dialog initiative switching unit 9 outputs an operation switching command to operate the system initiative dialog management unit 3. After that, the same operation as in the first embodiment is carried out to continue the dialogue.

By the above operation, utterances that are meaningless with respect to the operation of the device are registered in the meaningless utterance dictionary 13, and the meaningless utterance detection unit 12 refers to the meaningless utterance dictionary 13 to obtain no speech recognition result. It is detected whether it is a semantic utterance, and when a meaningless utterance is detected, the dialog initiative switching unit 9 outputs an operation switching command to operate the system-initiated dialog management unit 3, so that the user can smoothly When the input cannot be made and the operation of the device is meaningless, the device can take the initiative in the dialog and smoothly continue the dialog.

Eighth Embodiment 33 is a block diagram of an interactive information search device according to an eighth embodiment of the present invention. The eighth embodiment is the same as the seventh embodiment except that the function of the dialogue initiative switching unit 9 and the meaningless utterance number storage unit 14 are newly provided. Therefore, the contents of the meaningless utterance dictionary 13 are shown in FIG. 32 for the operations of the dialogue initiative switching unit 9 and the meaningless utterance number storage unit 14 in the present embodiment, and for the example of the dialogue between the device and the user shown in FIG. An example will be described in which the threshold value of the number of meaningless utterances is 2, and the user-initiated dialogue management unit 2 is operating at the start of dialogue.

When the user makes a call, the device speaks SY.
Output S81 to start the dialogue. User speaks USR
When 82 is input, the voice recognition unit 1 performs voice recognition and outputs "unnecessary word" as a recognition result. Nonsense utterance detection unit 1
When the voice recognition result “unnecessary word” is input, 2 refers to the meaningless utterance dictionary 13 shown in FIG. 32, determines that the voice recognition result “unnecessary word” is meaningless utterance, and determines the number of meaningless utterances. 1 is stored in the storage unit 14. For the user's utterance USR82, the response presenting unit 8 operates in the same manner as in the third embodiment.
Presents the device utterance SYS83.

When the user inputs the utterance USR84, the voice recognition unit 1 outputs "search request {NUL
L} ”is output. The meaningless utterance detection unit 12 refers to the meaningless utterance dictionary 13 shown in FIG. 32, determines that the voice recognition result “search request {NULL}” is a meaningless utterance, and stores the meaningless utterance number storage unit 14 in the meaningless utterance count storage unit 14. The number of held meaningless utterances is updated to 2. With respect to the user's utterance USR84, the response presenting unit 8 operates in the same manner as in the second embodiment to present the device's utterance SYS85.

Next, when the user inputs the utterance USR86, the apparatus operates in the same manner as in the case of the utterance USR82 described above, and the number of meaningless utterances stored in the meaningless utterance count storage unit 14 is updated to 3. It The response presenting unit 8 uses the utterance S of the device.
Present YS87.

When the number of times held in the meaningless utterance count storage unit 14 is updated to 3, the threshold value 2 is exceeded, so the dialog initiative switching unit 9 issues an operation switching command to operate the system-initiated dialog management unit 3. Is output. After that, the device operates similarly to the third embodiment and presents the utterance SYS88 to the user.

With the above operation, utterances that are meaningless with respect to the operation of the apparatus are registered in the meaningless utterance dictionary 13, and the meaningless utterance detection unit 12 refers to the meaningless utterance dictionary 13 to obtain no speech recognition result. When the meaning utterance is detected, the meaningless utterance count storage unit 14 stores the number of meaningless utterances, and when the number of meaningless utterances exceeds a threshold value, the dialogue initiative switching unit 9 causes the system-initiated dialogue management unit. Since the operation switching command is output so as to operate No. 3, when the user cannot smoothly input and repeats meaningless input to the operation of the apparatus, the apparatus smoothly takes the initiative of the dialogue. It is possible to continue the dialogue.

Ninth Embodiment 35 is a block diagram of an interactive information search device according to a ninth embodiment of the present invention. The ninth embodiment is different from the fourth embodiment in that the dialogue initiative switching unit 9 has a new function and a repeat utterance detection unit 15 is newly added.
Other points are the same as those of the fourth embodiment.
Therefore, the dialogue initiative switching unit 9 in the present embodiment
The operation of the repetitive utterance detection unit 15 will be described with respect to an example of the dialog between the device and the user shown in FIG.

When the user makes a call, the device speaks SY.
Output S91 to start the dialogue. After that, the device operates similarly to the fourth embodiment and presents the utterance SYS92 to the user. The dialogue context stored in the dialogue context storage unit 5 at the time when the output of the utterance SYS 92 is finished is shown in FIG.

Next, when the user's utterance USR93 is input, the voice recognition unit 1 outputs "condition name declaration {condition = place}" as the voice recognition result. When the voice recognition result is input, the repeat utterance detection unit 15 refers to the dialogue context shown in FIG. 37, and the parameter portion of the input voice recognition result is
Since it matches the parameter portion of the question utterance “condition question {condition = place}” presented by the device immediately before, it is determined that the user's utterance USR93 is a repeat of the device utterance SYS92.

When the repeat utterance detection unit 15 detects a repeat utterance, the dialog initiative switching unit 9 outputs an operation switching command to temporarily operate the user-initiated dialog management unit 2. The device operates similarly to the fourth embodiment, outputs the utterance SYS94 and waits for the user's input. When the user inputs the utterance USR95, the user operates in the same manner as in the fourth embodiment and presents the utterances SYS96 and SYS97 to the user.

By the above operation, the repeat utterance detection unit 15
Refers to the dialogue context storage unit 5 to determine whether or not the voice recognition result is a repeat of the utterance content presented by the device, and when the repeat utterance detection unit 15 detects the repeat utterance, the dialog initiative switching unit 9 outputs an operation switching command to temporarily operate the user-initiated dialogue management unit 2, so that the user temporarily takes the initiative in the dialogue in a situation where the user is thinking about the response of the device. By passing and waiting for input, it becomes possible to continue the dialogue smoothly.

Embodiment 10. 38 is a block diagram of an interactive information retrieval apparatus according to Embodiment 10 of the present invention. The tenth embodiment is different from the fourth embodiment in that
The difference is that the function of the dialogue initiative switching unit 9 is newly provided with a predicted utterance knowledge storage unit 16 and a predicted utterance determination unit 17,
Others are the same as those in the fourth embodiment. Therefore, the dialogue initiative switching unit 9 and the predicted utterance determination unit 17 according to the present embodiment.
39 will be described with respect to an example of the dialog between the apparatus and the user shown in FIG. 39, taking as an example the case where the system-initiated dialog management unit 3 is operating at the start of the dialog.

FIG. 40 shows an example of the knowledge stipulated in the predictive utterance knowledge storage unit 16, in which the utterances of the user predicted with respect to the question contents of the apparatus are defined as a table. For example, it is specified that the user's utterance predicted for the device question “conditional question {condition = location}” is “condition specification {location = *}” or “search request {location = *}”. ing. “*” In the table means that it matches an arbitrary value.

When the user makes a call, the device speaks SY.
Output S101 to start the dialogue. The device operates similarly to the fourth embodiment and presents the utterance SYS102. At this point, the dialogue context stored in the dialogue context storage unit 5 is as shown in FIG.

When the user inputs the utterance USR103,
The voice recognition unit 1 displays the result of voice recognition as “condition designation {location =
Near Yokohama station} "is output. When the voice recognition result is input, the predictive utterance determination unit 17 refers to the predictive utterance knowledge storage unit 16 illustrated in FIG. 40 and the dialogue context illustrated in FIG. 41, and inputs the input voice recognition result “condition designation {place = Yokohama. It is determined that “near the station}” is a predicted utterance for the device question utterance “condition question {condition = place}”. Therefore, the dialogue initiative switching unit 9 does not operate, and the operation of the system initiative dialogue management unit 3 continues. The system-initiated dialogue management unit 3 operates the user's utterance USR 103 in the same manner as in the fourth embodiment, and presents the utterances SYS 104 and SYS 105 to the user. The dialogue context held in the dialogue context storage unit 5 at this point is as shown in FIG.

Next, when the user's utterance USR 106 is input, the voice recognition unit 1 outputs "condition designation {room type = single}" as the voice recognition result. When the voice recognition result is input, the predictive utterance determination unit 17 refers to the predictive utterance knowledge storage unit 16 illustrated in FIG. 40 and the dialogue context illustrated in FIG. 42 to input the input voice recognition result “condition specification {room type =
It is determined that “single}” is not the expected utterance for the device question utterance “condition question {condition = budget}”. Therefore, the dialog initiative switching unit 9 outputs an operation switching command to operate the user-initiated dialog management unit 2.

When the user's utterance USR 106 is input, the user-initiated dialogue management unit 2, information retrieval unit 6, and response presentation unit 8 operate in the same manner as in the fourth embodiment, and the device utterance SY
It outputs S107 and waits for the user's utterance. When the user's utterance USR 108 is input, the device operates similarly to the fourth embodiment and presents the utterance SYS 109.

With the above operation, the predicted utterance knowledge and the dialogue context that define the utterance content of the user predicted for the question content generated by the system-initiated dialogue management unit 3 are referred to,
The predictive utterance determination unit 17 determines whether or not the voice recognition result is a predictive utterance for the question content of the device, and when the voice recognition result is not determined to be the predictive utterance, the dialog initiative switching unit 9 takes the initiative of the user. Since the operation switching command is output to operate the dialogue management unit 2, the user takes the initiative in a situation where the user clearly has a condition other than the search condition inquired from the device, and the user has a smooth dialogue. It is possible to continue.

Eleventh Embodiment 43 is a block diagram of an interactive information retrieval apparatus according to Embodiment 11 of the present invention. The eleventh embodiment differs from the tenth embodiment in that the function of the dialogue initiative switching unit 9 and a new unpredicted utterance count storage unit 18 are newly provided, and the other embodiments are the same as the tenth embodiment. Therefore, the operations of the dialogue initiative switching unit 9 and the unpredicted utterance count storage unit 18 according to the present embodiment will be described with reference to FIG.
An example will be described in which the predictive utterance knowledge storage unit 0 of 0 is used and the threshold of the number of unpredicted utterances is 1, and the system-initiated dialogue management unit 3 is operating at the time of the dialogue start.

When the user makes a call, the device speaks SY.
Output S111 to start the dialogue. Utterance SYS115
Until then, the apparatus operates in the same manner as in the tenth embodiment. The dialogue context held in the dialogue context storage unit 5 at the time when the utterance SYS 115 is presented is as shown in FIG.

When the user's utterance USR 116 is input, the predictive utterance determination unit 17 operates in the same manner as in the tenth embodiment, and the input voice recognition result “condition designation {room type = single}” is set in the device. It is determined that the question utterance is not the utterance predicted for the “condition question {condition = budget}”, and the number stored in the unpredicted utterance count storage unit 18 is updated to 1. The dialog initiative switching unit 9 does not operate because the threshold value of the number of unexpected utterances does not exceed 1.

For the user's utterance USR116, the system-initiated dialogue management unit 3, the information retrieval unit 6, and the response presentation unit 8 operate in the same manner as in the fourth embodiment, and output the utterances SYS117 and SYS118. The dialogue context at this point is as shown in FIG.

When the user's utterance USR 119 is input, the predictive utterance determination unit 17 operates in the same manner as in the tenth embodiment, and the input voice recognition result “condition designation {room type = single}” is stored in the device. It is determined that the question utterance is not the utterance predicted for the “condition question {condition = budget}”, and the number stored in the unpredicted utterance count storage unit 18 is updated to 2. Since the number of unpredicted utterances exceeds the threshold value 1, the dialog initiative switching unit 9 outputs an operation switching command to operate the user-initiated dialog management unit 2.

In response to the user's utterance USR 119, the user-initiated dialogue management unit 2, the information retrieval unit 6, and the response presenting unit 8 operate in the same manner as in the first embodiment to output the utterance SYS 120 and to output the user's utterance. wait.

By the above operation, the number of unpredicted utterances storage unit 18 stores the number of times the voice recognition result is determined not to be the predicted utterance, and when the number of unpredicted utterances exceeds the threshold, the dialogue initiative switching unit Since 9 outputs the operation switching command to operate the user-initiated dialogue management unit 2, the user clearly has a condition other than the search condition inquired from the device, and the user inputs the question from the device. The condition different from the desired condition is repeated, and in a situation where the dialogue does not proceed smoothly, the user can take the initiative and continue the dialogue smoothly.

[0112]

As described above, according to the present invention, a voice recognition unit for recognizing a voice of a user input to a device and outputting a recognition result, and a voice recognition result input from the voice recognition unit are used. A user-initiated dialogue management unit that generates and outputs an information search request, and generates and outputs a response content to a user from a search result that is input from an information search unit described later, and a voice input from the voice recognition unit. An information search request is generated and output from the recognition result, a response content to the user is generated and output from the search result input from the information search unit described later, and a question content generation knowledge storage unit described later and a dialogue described later. A system-initiated dialogue management unit that refers to the context storage unit to generate and output question contents to the user, and stores question-content generation knowledge for generating question contents to the user in the system-initiated dialogue management unit Do From the question content generation knowledge storage unit, the dialogue context storage unit that stores the voice recognition result, the search result, the response content, and the question content, and the user-initiated dialogue management unit or the system-initiated dialogue management unit. The information retrieval unit that performs information retrieval in response to an input information retrieval request and outputs the retrieval result, and the response content and question content input from the user-initiated dialogue management unit or the system-initiated dialogue management unit Based on the input interval, the response presenting section to present to, the input interval measuring section that measures and outputs the input interval from the time when the response presenting section presents the response to the time when the voice is input to the device, Interactive information, comprising: a user-initiated dialog management unit and a dialog initiative switching unit that outputs an operation switching command of the system-initiated dialog management unit. Since the search device and the interactive information search method based on the search device are used, the operation switching command of the user-initiated dialogue management unit and the system-initiated dialogue management unit is output based on the input interval, so that the user can input smoothly. If a situation does not occur, the device will take the initiative in the subsequent dialogues and the dialogue will continue smoothly without the user having to wait for a certain period of time, and even if the device does not act on his behalf It becomes possible to search for necessary information.

The input interval measuring unit measures and outputs the input interval from the time when the response presenting unit finishes presenting the response to the time when the voice is input to the device, and the dialogue initiative switching is performed. When the input interval exceeds the threshold, the unit outputs an operation switching command to operate the system-initiated dialog management unit. Therefore, when the input interval exceeds the threshold, the system-initiated dialog management unit operates. Since the operation switching command is output so that the user can operate, if the user cannot input smoothly, the device takes the initiative in dialogue after that, and the user does not always wait for a certain period of time. Can be smoothly continued, and the user can search for necessary information without the device arbitrarily acting for the user.

Further, the dialog initiative switching unit outputs an operation switching command to temporarily operate the system initiative dialog management unit when the input interval exceeds the threshold, and decreases the threshold. Because it was characterized
When the input interval measuring unit measures and outputs the input interval from the time when the device finishes providing the response to the next voice input to the device, when the input interval exceeds the threshold, the dialogue initiative The switching unit outputs an operation switching command to temporarily operate the system-initiated dialogue management unit and further reduces the threshold, so basically the user takes the initiative to proceed with the dialogue, but The device temporarily takes control of the dialogue in the event that no input can be made to the device and the dialogue continues smoothly, and when such a situation occurs repeatedly, the device takes the initiative more quickly. It is possible to continue the dialogue smoothly.

Further, an input interval storage unit for storing the input interval is further provided, and the dialogue initiative switching unit refers to the input interval storage unit to detect an input interval exceeding a threshold value more than a prescribed number of times. Since the operation switching command is output so as to operate the system-initiated dialogue management unit, the input interval from the time when the device finishes presenting the response until the next voice is input to the device. The input interval measuring unit measures and outputs the input interval, and the input interval measuring unit stores the input interval measured by the input interval storage unit. When the input interval exceeding the threshold value occurs more than the specified number of times, the dialogue initiative switching unit sets the system. Since the operation switching command is output to operate the initiative dialogue management unit, basically the user has the initiative to proceed with the dialogue, but the situation in which the user cannot input smoothly occurs more than the specified number of times. If, it is possible that the device continues smoothly dialogue taking initiative interaction.

Further, the input interval measuring unit may notify the device from the time when the response presenting unit starts presenting the response when the voice is input to the device before the response presenting unit finishes presenting the response. The interrupt input interval up to the time when the voice is input is measured and output, and the dialog initiative switching unit operates the operation switching signal to operate the user-initiated dialog management unit when the interrupt input interval is shorter than the threshold value. Is output, so when the voice is input to the device before the response presentation unit finishes presenting the response, the input interval measurement unit, from the time when the response presentation unit starts presenting the response, Measures and outputs the interrupt input interval until the time when the voice is input to the device, further outputs the response stop command to the response presentation unit, and when the response stop command is input, the response presentation unit presents the response being output. Processing When the interrupt input interval is shorter than the threshold, the dialog initiative switching unit outputs an operation switching signal to operate the user-initiated dialog management unit, so the user clearly has the search condition to input to the device. In such a situation, the user's request can be input by interrupting questions and guidance from the device, and the user can take the initiative and continue the conversation smoothly thereafter.

Further, when the input interval storage unit stores the interrupt input interval, and the dialogue initiative switching unit refers to the input interval storage unit, and when the interrupt input interval shorter than the threshold value occurs a prescribed number of times or more. In addition, since the operation switching command is output to operate the user-initiated dialogue management unit, when the voice is input to the device before the response presentation unit finishes presenting the response, the input interval measurement is performed. Section measures and outputs the interrupt input interval from the time when the response presenting section starts presenting the response to the time when the voice is input to the device, and further outputs the response stop command to the response presenting section, and measures the input interval. The input interval storage unit stores the interrupt input interval measured by the unit, and when the response stop command is input, the response presenting unit stops the response presenting process during output, and the interrupt input interval shorter than the threshold is regulated. If the number of occurrences exceeds the number of times, the dialog initiative switching unit outputs an operation switching signal so as to operate the user-initiated dialog management unit. Therefore, in a situation where the user clearly has the search condition to be input to the device, The user's request can be entered by interrupting questions and guidance from the user, and after that, the user can take the initiative and continue the conversation smoothly.

Further, the dialog control unit further comprises a user-initiated utterance dictionary in which utterance contents for switching the operation are registered in the user-initiated dialog management unit, and the dialog initiative switching unit, when the interrupt input interval is input, When the voice recognition result input from the recognition unit matches the utterance content registered in the user-initiated utterance dictionary, an operation switching command is output to operate the user-initiated dialogue management unit.
Therefore, when the voice is input to the device before the response presenting unit finishes presenting the response, the input interval measuring unit outputs the voice to the device from the time when the response presenting unit starts presenting the response. The interrupt input interval up to the input time is measured and output, and the utterance for switching the operation is registered in the user-initiated utterance dictionary in the user-initiated utterance dictionary. Is input, if the voice recognition result input from the voice recognition unit matches the user-initiated utterance registered in the user-initiated utterance dictionary, the operation is switched to temporarily operate the user-initiated dialogue management unit. In order to output the command, in the situation where the user clearly has the search condition to input to the device,
The user can take initiative in the dialogue by explicitly blocking questions and guidance from the device, and the dialogue can be smoothly continued.

Further, the input interval measuring unit outputs a response stop command to the response presenting unit when the voice is input to the device before the response presenting unit finishes presenting the response, and the response presenting unit Is characterized in that the response presenting process being output is stopped when a response stop command is input, so when the voice is input to the device before the response presenting unit finishes presenting the response, the input is performed. The interval measurement unit measures and outputs the interrupt input interval from the time when the response presentation unit starts presenting the response to the time when the voice is input to the device, and further outputs the response stop command to the response presentation unit, and the user The utterance for switching the operation is registered in the user-directed dialogue management unit in the initiative vocabulary dictionary, and the dialogue initiative switching unit, when the interrupt input interval is input, the voice recognition result input from the voice recognition unit is User driven When it matches the user-initiated utterance registered in the speech dictionary, the operation switching command is output to temporarily operate the user-initiated dialogue management unit, so the user clearly has the search condition to input to the device. In such a situation, the user can take initiative in the dialogue by explicitly blocking questions and guidance from the device, and the dialogue can be continued smoothly.

Further, with reference to the meaningless utterance dictionary in which meaningless utterance contents regarding the operation of the device and the meaningless utterance dictionary are referred, whether the voice recognition result input from the voice recognition unit is meaningless utterance or not. A nonsense utterance detection unit that detects whether or not
Further, the dialog initiative switching unit outputs an operation switching command to operate the system-initiated dialog management unit when the meaningless utterance detection unit detects a meaningless utterance. Therefore, the meaningless utterance regarding the operation of the device is registered in the meaningless utterance dictionary, and the meaningless utterance detection unit refers to the meaningless utterance dictionary to detect whether the voice recognition result is meaningless utterance, When a meaningless utterance is detected, the dialog initiative switching unit outputs an operation switching command to operate the system-initiated dialog management unit, so that the user cannot input smoothly and the device operation cannot be performed. When a meaningful input is made, the device can take the initiative in the dialogue and smoothly continue the dialogue.

Further, the non-meaningful speech number storage unit for storing the number of times of the nonsense speech is further provided, and the dialogue initiative switching unit refers to the meaningless speech number storage unit, and the number of meaningless speech exceeds a threshold value. In this case, the operation switching command is output to operate the system-initiated dialogue management unit, so meaningless utterances related to the operation of the device are registered in the meaningless utterance dictionary, and meaningless utterances are registered. The detection unit
Detecting whether or not the speech recognition result is meaningless utterance by referring to the meaningless utterance dictionary, the meaningless utterance number storage unit stores the number of meaningless utterances, and when the number of meaningless utterances exceeds a threshold value, Since the dialog initiative switching unit outputs the operation switching command to operate the system-initiated dialog management unit, if the user cannot input smoothly and repeats meaningless input to the operation of the device, Can take the initiative in the dialogue and continue the dialogue smoothly.

Further, by referring to the dialogue context storage unit, it is judged whether or not the voice recognition result input from the voice recognition unit is a repeat of the response content or the question content presented by the apparatus immediately before. Further comprising a unit, the dialogue initiative switching unit outputs an operation switching command to temporarily operate the user-initiated dialogue management unit when the replay utterance detection unit detects a recitation utterance. Therefore, the repeat utterance detection unit refers to the dialogue context storage unit,
It is determined whether or not the voice recognition result is a repeat of the utterance content presented by the device, and when the repeat utterance detection unit detects the repeat utterance,
Since the dialog initiative switching unit outputs the operation switching command to temporarily operate the user-initiated dialog management unit,
In the situation where the user is thinking about the response of the device, by temporarily giving the initiative of the dialogue to the user and waiting for the input, the dialogue can be smoothly continued.

Further, a predictive utterance knowledge storage unit defining the utterance content of the user predicted for the question content generated by the system-initiated dialogue management unit, the dialogue context storage unit, and the predicted utterance knowledge storage unit are provided. With reference to, when the device presents a question to the user, a predictive utterance determination unit that determines whether or not the voice recognition result input from the voice recognition unit is a predictive utterance for the question content of the device. Furthermore, the dialogue initiative switching unit, when the voice recognition result is not determined to be a predicted utterance in the predicted utterance determination unit,
Since the operation switching command is output so as to operate the user-initiated dialogue management unit, the predictive utterance that defines the expected utterance content of the user for the question content generated by the system-initiated dialogue management unit With reference to the knowledge and the dialogue context, the predictive utterance determination unit determines whether or not the voice recognition result is the predictive utterance for the question content of the device, and if the voice recognition result is not determined to be the predictive utterance, the dialogue is performed. Since the initiative switching unit outputs the operation switching command to operate the user-initiated dialogue management unit, the user takes the initiative in a situation where the user clearly has a condition other than the search condition inquired from the device. Therefore, it becomes possible to continue the dialogue smoothly.

Further, the predictive utterance determining unit further includes a non-predictive utterance number storage unit that stores the number of times the voice recognition result is determined not to be the predictive utterance, and the dialogue initiative switching unit includes the non-predictive utterance. When the number of times stored in the number-of-times storage unit exceeds a threshold value, an operation switching command is output so as to operate the user-initiated dialogue management unit. To store the number of times that the result is determined not to be predictive utterance, and when the number of unpredicted utterances exceeds a threshold value, the dialogue initiative switching unit outputs a movement switching command to operate the user-initiated dialogue management unit. , The user clearly has a condition other than the search condition inquired from the device, and the question from the device is repeatedly different from the condition that the user wants to input, and the dialogue does not proceed smoothly. In the situation, it becomes possible for the user to continue smoothly dialogue to take the initiative.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an interactive information search device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing an example of a dialogue between a user and an interactive information search device according to the first embodiment of the present invention.

FIG. 3 is a diagram for explaining the operation of the interactive information search device according to the first embodiment of the present invention.

FIG. 4 is a diagram for explaining the operation of the interactive information search device according to the first embodiment of the present invention.

FIG. 5 is a diagram for explaining the operation of the interactive information search device according to the first embodiment of the present invention.

FIG. 6 is a diagram for explaining the operation of the interactive information search device according to the first embodiment of the present invention.

FIG. 7 is a diagram for explaining the operation of the interactive information search device according to the first embodiment of the present invention.

FIG. 8 is a diagram for explaining the operation of the interactive information search device according to the first embodiment of the present invention.

FIG. 9 is a diagram for explaining the operation of the interactive information search device according to the first embodiment of the present invention.

FIG. 10 is a diagram for explaining the operation of the interactive information search device according to the first embodiment of the present invention.

FIG. 11 is a diagram for explaining the operation of the interactive information search device according to the first embodiment of the present invention.

FIG. 12 is a diagram for explaining the operation of the interactive information search device according to the first embodiment of the present invention.

FIG. 13 is a diagram for explaining the operation of the interactive information search device according to the first embodiment of the present invention.

FIG. 14 is a diagram showing an example of interaction between a user and an interactive information search device according to a second embodiment of the present invention.

FIG. 15 is a diagram for explaining the operation of the interactive information search device according to the second embodiment of the present invention.

FIG. 16 is a diagram for explaining the operation of the interactive information search device according to the second embodiment of the present invention.

FIG. 17 is a configuration diagram of an interactive information search device according to a third embodiment of the present invention.

FIG. 18 is a diagram showing an example of interaction between a user and an interactive information search device according to a third embodiment of the present invention.

FIG. 19 is a diagram for explaining the operation of the interactive information search device according to the third embodiment of the present invention.

FIG. 20 is a diagram for explaining the operation of the interactive information search device according to the third embodiment of the present invention.

FIG. 21 is a diagram for explaining the operation of the interactive information search device according to the third embodiment of the present invention.

FIG. 22 is a configuration diagram of an interactive information search device according to a fourth embodiment of the present invention.

FIG. 23 is a diagram showing an example of interaction between a user and an interactive information search device according to a fourth embodiment of the present invention.

FIG. 24 is a configuration diagram of an interactive information search device according to a fifth embodiment of the present invention.

FIG. 25 is a diagram showing an example of interaction between a user and an interactive information search device according to a fifth embodiment of the present invention.

FIG. 26 is a diagram for explaining the operation of the interactive information search device according to the fifth embodiment of the present invention.

FIG. 27 is a configuration diagram of an interactive information search device according to a sixth embodiment of the present invention.

FIG. 28 is a diagram showing an example of a dialogue between the interactive information search device and the user according to the sixth embodiment of the present invention.

FIG. 29 is a diagram for explaining the operation of the interactive information search device according to the sixth embodiment of the present invention.

FIG. 30 is a configuration diagram of an interactive information search device according to a seventh embodiment of the present invention.

FIG. 31 is a diagram showing an example of interaction between a user and an interactive information search device according to a seventh embodiment of the present invention.

FIG. 32 is a diagram for explaining the operation of the interactive information search device according to the seventh embodiment of the present invention.

FIG. 33 is a configuration diagram of an interactive information search device according to an eighth embodiment of the present invention.

FIG. 34 is a diagram showing an example of interaction between the interactive information search device and the user according to the eighth embodiment of the present invention.

FIG. 35 is a configuration diagram of an interactive information search device according to a ninth embodiment of the present invention.

FIG. 36 is a diagram showing an example of a dialogue between the interactive information search device and the user according to the ninth embodiment of the present invention.

FIG. 37 is a diagram for explaining the operation of the interactive information search device according to the ninth embodiment of the present invention.

FIG. 38 is a configuration diagram of an interactive information search device according to a tenth embodiment of the present invention.

FIG. 39 is a diagram showing an example of a dialogue between the interactive information search device and the user according to the tenth embodiment of the present invention.

FIG. 40 is a diagram for explaining the operation of the interactive information search device in accordance with the tenth embodiment of the present invention.

FIG. 41 is a diagram for explaining the operation of the interactive information search device in accordance with the tenth embodiment of the present invention.

FIG. 42 is a diagram for explaining the operation of the interactive information search device in accordance with the tenth embodiment of the present invention.

FIG. 43 is a configuration diagram of an interactive information search device according to an eleventh embodiment of the present invention.

FIG. 44 is a diagram showing an example of a dialogue between the interactive information search device and the user according to the eleventh embodiment of the present invention.

FIG. 45 is a diagram for explaining the operation of the interactive information search device according to the eleventh embodiment of the present invention.

[Fig. 46] Fig. 46 is a configuration diagram of a natural language dialogue system which is a conventional device of this type.

[Explanation of symbols]

1 voice recognition unit, 2 user-initiated dialogue management unit, 3 system-initiated dialogue management unit, 4 question content generation knowledge storage unit, 5
Dialog context storage unit, 6 Information retrieval unit, 7 Input interval measurement unit, 8 Response presentation unit, 9 Dialog initiative switching unit, 10
Input interval storage unit, 11 User-initiated speech dictionary, 12
Meaningless utterance detection unit, 13 Meaningless utterance dictionary, 14 Meaningless utterance count storage unit, 15 Repeat utterance detection unit, 16 Predicted utterance knowledge storage unit, 17 Predicted utterance determination unit, 18 Unpredicted utterance count storage unit.

Claims (26)

[Claims] 1. A voice recognition unit for recognizing a voice of a user input to a device and outputting a recognition result, and an information search request generated and output from the voice recognition result input from the voice recognition unit, A user-initiated dialogue management unit that generates and outputs a response content to a user from a search result input from an information search unit described later, and an information search request from the voice recognition result input from the voice recognition unit. It outputs and outputs the response content to the user from the search result input from the information search unit described later and outputs it to the user by referring to the question content generation knowledge storage unit described later and the dialogue context storage unit described later. A system-initiated dialogue management unit for generating and outputting the question content, and a question-content generation knowledge storage unit for storing question-content generation knowledge for generating question content for the user in the system-initiated dialogue management unit, In response to an information search request input from the dialog context storage unit that stores the voice recognition result, the search result, the response content, and the question content, and the user-initiated dialog management unit or the system-initiated dialog management unit. An information search unit that performs an information search and outputs a search result; a response presentation unit that presents a response content and a question content input from the user-initiated dialogue management unit or the system-initiated dialogue management unit to a user; An input interval measuring unit that measures and outputs an input interval from a time when the response presenting unit presents a response to a time when a voice is input to the device; and based on the input interval, the user-initiated dialogue management unit and the An interactive information retrieval device comprising: a dialog initiative switching unit that outputs an operation switching command of a system-initiated dialog management unit. 2. The input interval measuring unit measures and outputs an input interval from the time when the response presenting unit finishes presenting a response to the time when a voice is input to the device, and the dialogue initiative switching. The interactive information retrieval device according to claim 1, wherein the unit outputs an operation switching command to operate the system-initiated dialog management unit when the input interval exceeds a threshold value. 3. The dialog initiative switching unit outputs an operation switching command to temporarily operate the system-initiated dialog management unit when the input interval exceeds a threshold, and decreases the threshold. The interactive information search device according to claim 2, wherein 4. An input interval storage unit for storing the input interval is further provided, wherein the dialogue initiative switching unit refers to the input interval storage unit, and when an input interval exceeding a threshold value occurs a prescribed number of times or more. The interactive information retrieval device according to claim 2, wherein an operation switching command is output to operate the system-initiated dialog management unit. 5. The input interval measurement unit, when a voice is input to the device before the response presentation unit finishes presenting the response, starts inputting the response to the device from the time when the response presentation unit starts presenting the response. The interrupt input interval up to the time when the voice is input is measured and output, and the dialog initiative switching unit operates the operation switching signal to operate the user-initiated dialog management unit when the interrupt input interval is shorter than a threshold value. The interactive information retrieval device according to any one of claims 1 to 4, wherein 6. The input interval storage unit stores the interrupt input interval, and the dialogue initiative switching unit refers to the input interval storage unit, and when an interrupt input interval shorter than a threshold value occurs a prescribed number of times or more. The interactive information search device according to claim 5, further comprising: outputting an operation switching command to operate the user-initiated dialog management unit. 7. The user-initiated dialogue management unit further comprises a user-initiated utterance dictionary in which utterance contents for switching the operation are registered, wherein the dialogue initiative switching unit receives the voice when the interrupt input interval is input. When the voice recognition result input from the recognition unit matches the utterance content registered in the user-initiated utterance dictionary, an operation switching command is output so as to operate the user-initiated dialogue management unit. The interactive information search device according to claim 5. 8. The input interval measuring unit outputs a response stop command to the response presenting unit when voice is input to the device before the response presenting unit finishes presenting the response, and the response presenting unit The interactive information search device according to any one of claims 5 to 7, wherein the response presentation process being output is stopped when a response stop command is input. 9. A meaningless utterance dictionary in which meaningless utterance contents are registered regarding the operation of the device, and whether the voice recognition result input from the voice recognition unit is a meaningless utterance by referring to the meaningless utterance dictionary. Further comprising a meaningless utterance detection unit for detecting whether or not the dialogue initiative switching unit operates the system led dialogue management unit when the meaningless utterance detection unit detects a meaningless utterance. 9. The interactive information search device according to claim 1, wherein an operation switching command is output. 10. The non-meaningful utterance number storage unit for storing the number of the nonsense utterances, the dialogue initiative switching unit refers to the nonsense utterance number storage unit, and the number of nonsense utterances exceeds a threshold value. The interactive information search device according to claim 9, wherein an operation switching command is output so as to operate the system-initiated dialog management unit. 11. A repeat speech detection for determining whether or not the voice recognition result input from the voice recognition unit is a repeat of the response content or the question content presented by the device immediately before with reference to the dialogue context storage unit. Further comprising a unit, the dialogue initiative switching unit outputs an operation switching command to temporarily operate the user-initiated dialogue management unit when the repetitive utterance detection unit detects a repetitive utterance, The interactive information retrieval device according to any one of claims 1 to 10. 12. A predictive utterance knowledge storage unit defining the utterance content of a user predicted for the question content generated by the system-initiated dialogue management unit, the dialogue context storage unit, and the predicted utterance knowledge storage unit. With reference to, when the device presents a question to the user, a predictive utterance determination unit that determines whether or not the voice recognition result input from the voice recognition unit is a predictive utterance for the question content of the device. Furthermore, the dialog initiative switching unit outputs an operation switching command to operate the user-initiated dialog management unit when the speech recognition result is not determined to be a predicted utterance by the predicted utterance determination unit, The interactive information search device according to any one of claims 1 to 11, characterized in that: 13. The non-predicted utterance number storage unit stores the number of times the predictive utterance determination unit determines that the voice recognition result is not the predictive utterance, and the dialogue initiative switching unit includes the non-predicted utterance. 13. The interactive information retrieval device according to claim 12, wherein when the number of times stored in the number of times storage unit exceeds a threshold value, an operation switching command is output to operate the user-initiated dialog management unit. . 14. A voice recognition step of recognizing a voice of a user input to a device and outputting a recognition result, and an information search request is generated and output from the voice recognition result input from the voice recognition step, A user-initiated dialogue management step of generating and outputting a response content to a user from a search result input from an information search step described later, and an information search request from the voice recognition result input from the voice recognition step. Output and output the response contents to the user from the search result input from the information search process described below, and refer to the question content generation knowledge storage process described below and the dialog context storage process described below to the user System-initiated dialogue management step of generating and outputting the question content, and the question storing question content generation knowledge for generating question content to the user in the system-initiated dialogue management step The content generation knowledge storage step, the dialogue context storage step of storing the voice recognition result, the search result, the response content, and the question content, and the user-initiated dialogue management step or the system-initiated dialogue management step. The information search step of performing an information search in response to an information search request to output a search result, and presenting to the user the response content and question content input from the user-initiated dialogue management step or the system-initiated dialogue management step. A response presentation step to do, an input interval measurement step of measuring and outputting an input interval from the time when the response presentation step presents a response to the time when voice is input to the device, and based on the input interval, A user-initiated dialogue management step and a dialogue initiative switching step for outputting an operation switching command of the system-initiated dialogue management step. Interactive information retrieval method characterized. 15. The input interval measuring step measures and outputs an input interval from a time when the response presenting step finishes presenting a response to a time when a voice is inputted to the device, and the dialogue initiative switching. The interactive information retrieval method according to claim 14, wherein the step outputs an operation switching command to operate a system-initiated dialog management step when the input interval exceeds a threshold value. 16. The dialog initiative switching step outputs an operation switching command to temporarily operate the system initiative dialog management step when the input interval exceeds a threshold, and decreases the threshold. 16. The interactive information search method according to claim 15, wherein: 17. An input interval storing step of storing the input interval, the dialogue initiative switching step referring to the input interval storing step, when an input interval exceeding a threshold value occurs a prescribed number of times or more. 16. The interactive information retrieval method according to claim 15, wherein an operation switching command is output to operate the system-initiated dialog management process. 18. The input interval measuring step, in the case where a voice is input to the device before the response presenting step finishes presenting a response, from the time when the response presenting step starts presenting a response to the apparatus. The interrupt input interval up to the time when the voice is input is measured and output, and the dialog initiative switching step, when the interrupt input interval is shorter than a threshold, an operation switching signal to operate the user-initiated dialog management step. 18. The interactive information retrieval method according to any one of claims 14 to 17, wherein: 19. The input interval storing step stores the interrupt input interval, and the dialogue initiative switching step refers to the input interval storing step, when an interrupt input interval shorter than a threshold value occurs a prescribed number of times or more. 19. The interactive information search method according to claim 18, further comprising outputting an operation switching command to operate the user-initiated dialog management process. 20. Further comprising a user-initiated utterance dictionary in which utterance content for switching the operation is registered in the user-initiated dialogue management step, wherein the dialogue initiative switching step includes the step of inputting the voice when the interrupt input interval is input. When the voice recognition result input from the recognition step matches the utterance content registered in the user-initiated utterance dictionary, an operation switching command is output to operate the user-initiated dialogue management step. The interactive information search method according to claim 18. 21. The input interval measuring step outputs a response stop command to the response presenting step when voice is input to the device before the response presenting step finishes presenting the response, and the response presenting step. 21. The interactive information retrieval method according to claim 18, wherein the response presentation process being output is stopped when a response stop command is input. 22. A meaningless utterance dictionary in which meaningless utterance contents are registered regarding the operation of the apparatus, and whether the voice recognition result input from the voice recognition step is a meaningless utterance by referring to the meaningless utterance dictionary. Further comprising a meaningless utterance detection step of detecting whether or not the dialogue initiative switching step operates the system-initiated dialogue management step when the meaningless utterance detection step detects a meaningless utterance. 22. The interactive information retrieval method according to claim 14, wherein an operation switching command is output. 23. Further comprising a meaningless utterance number storing step of storing the number of times of the insignificant utterance, wherein the dialogue initiative switching step refers to the meaningless utterance number storing step, and the meaningless utterance number exceeds a threshold value. 23. The interactive information search method according to claim 22, further comprising outputting an operation switching command so as to operate the system-initiated dialog management process. 24. A repeat utterance detection for determining whether or not the voice recognition result input from the voice recognition step is a repeat of the response content or question content presented immediately before by referring to the dialog context storing step. Further comprising a step, the dialogue initiative switching step, when the repeat utterance detection step detects a repeat utterance, outputs an operation switching command to temporarily operate the user-initiated dialogue management step, The interactive information retrieval method according to any one of claims 14 to 23. 25. A predictive utterance knowledge storing step, which defines a user's utterance content predicted for the question content generated by the system-initiated dialogue managing step, the dialogue context storing step, and the predictive utterance knowledge storing step. With reference to, when the device presents a question to the user, a predictive utterance determination process of determining whether or not the voice recognition result input from the voice recognition process is a predictive utterance for the question content of the device. Furthermore, the dialog initiative switching step outputs an operation switching command to operate the user-initiated dialog management step when the voice recognition result is not determined to be a predicted utterance in the predicted utterance determination step, The interactive information retrieval method according to any one of claims 14 to 24, wherein: 26. A non-predicted utterance number storage step of storing the number of times that the speech recognition result is determined not to be a predictive utterance in the predictive utterance determination step, wherein the dialogue initiative switching step includes the non-predicted utterance. 26. The interactive information retrieval method according to claim 25, wherein when the number of times stored in the number of times storage step exceeds a threshold value, an operation switching command is output to operate the user-initiated dialog management step. .