JP2002073662A - Information presenting device and recording medium with information presenting program recorded thereon - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information presenting device capable of efficiently presenting effective information to a user by simplifying the inputting operation of request information by the user. SOLUTION: The information presenting device is provided with an information request inputting function by conversation in natural language and a function for unitarily presenting-controlling an answer sentence concerning requested information and additional information of video, voice, an image, etc. Thus, the user can input required information by conversation with the information presenting device by using words which are usually used by the user. Furthermore, the user can acquire a response presented in an easily understandable form using the video, characters and voice.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information presenting apparatus for presenting information including video using a computer.
In particular, the present invention relates to an operation and control technology of the information presentation device and a natural language processing technology for inputting operation information to the information presentation device.

[0002]

2. Description of the Related Art Conventionally, in a video search device using a computer, that is, an information presentation device such as a video database, a user inputs a search keyword from a keyboard, or displays a classification or information indicated by the information presentation device. Is selected by using an input device such as a keyboard or a mouse to input operation information to the information presentation device, and information corresponding to the input operation information is displayed on the screen of the information presentation device. Such an information presenting device that involves an input operation is difficult to use for a user unfamiliar with search or the like.

[0003] In addition, information presented to a user from a conventional information presenting apparatus is a file in which materials such as a plurality of images, audio for explanation, and characters to be displayed on a screen are combined as one image in advance. The video files stored in the database and retrieved from the database in response to a request from the user are presented to the user.

On the other hand, regarding a natural language processing system for the purpose of dialogue between a computer and a human, Stuart C. Shap
Many studies have been conducted on ELIZA and others, as described in iro et al., translated by Setsuo Osuka, "A Dictionary of Artificial Intelligence", Maruzen, (1991).

[0005] Many conventional natural language processing systems involve syntactic analysis and semantic analysis, so the processing is complex. Since the syntax and semantic analysis processing is based on a model centered on written words, non-grammatical utterances are used. Or, it is an issue to deal with fragmentary utterances. In this regard, Kumamoto,
“Generating Cooperative Responses in Natural Language Dialogue Systems”, Journal of the Japanese Society for Artificial Intelligence, Vol. 14, No. 1, pp. 3-10 (1999).

[0006]

In a conventional video database, when a user unfamiliar with the search system performs a search, it is difficult to input or select a search word, and it is difficult to obtain a desired video. There is a problem. In order to improve this problem, a video database in which a user inputs search conditions by selecting a classification or information displayed on a screen has been proposed. However, a large amount of information is required, and the operation becomes complicated. There is a problem that a considerable operation is required until reaching the information to be performed.

In a conventional general information presentation device,
Information is presented to the user in such a manner that individual images and characters are presented alone, and information that can be effectively rendered by combining images and characters cannot be presented. However, in order to obtain an effective effect using only images, it is necessary to produce a wide variety of images, and this is not practical because a huge cost is generated. Also, as the amount of information increases, the amount of work to produce the information to be presented increases, so it becomes difficult to provide information according to the user's wishes and preferences, and the information is fixed, and as a result, , Which would cause the user to lose interest.

Accordingly, an object of the present invention is to provide an information presenting apparatus which simplifies a task of inputting required information by a user and can efficiently present effective information to the user.

Further, the present invention provides a recording medium in which an information presenting program is recorded in which an operation of inputting request information by a user is simplified and effective information can be efficiently presented to the user. Aim.

[0010]

In order to achieve the above object, an information presenting apparatus of the present invention requests information by inputting in a language normally used by a user through dialogue with a computer, and requests the information by a user. The information is presented to the user in an easy-to-understand format such as video, text, and audio.

FIG. 1 is a diagram showing the principle of the present invention. As shown in FIG. 1, the invention according to claim 1 receives a question sentence in which request information is expressed in a natural language, analyzes the question sentence,
Request information input means 10 for acquiring response control information corresponding to the analyzed question sentence; and presentation for producing presentation information by combining material information according to the acquired response control information for responding to the question sentence. Information production means 20
An information presentation device 1 that provides presentation information according to request information, characterized by having:

FIG. 2 is a configuration diagram of the request information input means 10 of the information presentation device 1 of the present invention. According to the invention of claim 2, the request information input means 10 receives the question sentence, morphologically analyzes the question sentence, and creates a word pattern corresponding to the question sentence, And a conversation control script storing means for storing a template pattern corresponding to the request information and a conversation control script describing a process to be performed when the template pattern appears in the question sentence. The means 12 detects a template pattern corresponding to the created word pattern from the conversation control script storage means 11, and acquires the conversation control script stored corresponding to the detected template pattern as the response control information. It is characterized by doing.

According to a third aspect of the present invention, there is provided a computer-readable recording medium recording an information presentation program for providing presentation information according to request information, wherein the program is expressed in a natural language requesting the information. A morphological analysis of the question sentence, a process of creating a word pattern from the words of the question sentence obtained by the morphological analysis, and a template pattern that matches the word pattern from template patterns stored in advance. A process for detecting, a process for acquiring a conversation control script that is stored corresponding to the detected template pattern and describes a process to be performed when the template pattern appears in the question message; And a response to the question Recording, comprising: a process of dividing presentation information into presentation control instructions specifying the presentation information; and a process of producing the presentation information including the material information combined in accordance with the presentation control instructions and the reply sentence. Medium.

According to the first to third aspects of the present invention, the information request input function by the natural language dialogue, the response sentence relating to the requested information, and the additional information such as video, audio, and image are unified. A function of controlling presentation is obtained. Therefore, the user can input necessary information by using a dialogue with the information presenting device by using words that the user normally uses, and the response to the user is performed by video, characters, audio, or the like. Will be presented in an easy-to-understand form.

The features of the present invention will be more apparent from the following description.

FIG. 3 is a block diagram of the presentation information producing means 20 of the information presentation device 1 of the present invention. The presentation information producing unit 20 includes an information storage unit 22 that stores the material information,
According to the response control information from the request information input means 10, the material information is extracted from the information storage means 22, and the extracted material information is combined to produce the presentation information, which is output together with a response to the question sentence. And a presentation control means 21 for performing the operation.

FIG. 4 is an explanatory diagram of an information presentation method implemented by the information presentation device 1 of the present invention. The information presentation method for providing presentation information according to the request information includes a step 1 of inputting a question sentence expressed in a natural language requesting the information.
And a step 2 for morphologically analyzing the input question sentence
Step 3 of creating a word pattern from the words of the question sentence obtained by the morphological analysis, Step 4 of detecting a template pattern that matches the word pattern from template patterns stored in advance,
Acquiring a conversation control script corresponding to the detected template pattern and describing a process to be performed when the template pattern appears in the question message; and 5; Step 6 of dividing the information into presentation control instructions that specify presentation information to be presented in response to the question sentence;
Step 7 of producing the presentation information by combining the material information in accordance with the presentation control command and outputting the presentation information together with the reply sentence.

FIG. 5 is a configuration diagram of an information presentation system including the information presentation device 1 of the present invention. The information presentation system includes a microphone 81 for inputting voice, a voice recognition device 82 connected to the microphone 81 for recognizing the input voice and generating character string data, and an auxiliary input device 83 for directly inputting character string data. Input device 8 for inputting a question sentence
And a display device 91 for displaying display information such as video, images and characters, a speech synthesizer 92 for receiving and converting character string data into an audio signal, and the audio synthesizer, and outputting the audio signal as audio. An output device 9 having a speaker 93, and the input device 8 and the presentation device are connected to the output device 9 and produce presentation information corresponding to the request information represented by the character string data received from the input device 8; An information presentation device that supplies the presentation information to the output device 9, wherein the information presentation device 1 includes a template pattern corresponding to the request information and a template pattern in the question text. Conversation control script storage means 11 for storing a conversation control script describing processing to be performed when the message appears,
A question sentence in which the request information is expressed in a natural language is received from the input device 8, a morphological analysis is performed on the question sentence, a word pattern corresponding to the question sentence is created, and the word pattern is created from the conversation control script storage unit 11. Input sentence analyzing means 12 for detecting a template pattern corresponding to the detected word pattern and obtaining a conversation control script stored corresponding to the detected template pattern, and an information storing means for storing viewable material information 22 and connected to the input sentence analysis means 12 and the information storage means 22, and extracts the material information from the information storage means 22 according to the conversation control script from the input sentence analysis means 12. Combine the information to produce the presentation information and output the presentation information together with the response to the question And having a presentation control unit 21 supplies the location.

The information presentation system includes an input device capable of processing a natural language as a device for inputting a user's request to the system when inputting request information. Questions can be given to the system using any language. The system analyzes the question sentence, extracts keywords and control commands of information requested by the user, and can perform control for producing presentation information. As a result, the user can easily perform system operations such as information search and input of an information presentation request.

In addition, when producing presentation information, a function of separately accumulating materials such as video, audio, and characters, and presenting such information in combination according to request information is added. It is easier to understand than information from a single material, and it is possible to present effective information. Further, by performing the presentation control in detail based on the information obtained by analyzing the question text, it is possible to personalize the presentation information, for example.

[0021]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail.

FIG. 6 is a configuration diagram of the information presentation system according to the first embodiment of the present invention. The information presentation system includes an input device 8, an output device 9, and the information presentation device 1. The input device 9 and the output device 9 can be connected to the information presentation device 1 in various forms, respectively. As this connection form, for example, a bus connection, a connection by a serial line, a LAN connection, a connection via a communication network such as the Internet, and the like are considered.

The information presentation device 1 includes a conversation control script storage unit 11, an input sentence analysis unit 12, a presentation control unit 21,
And an information storage unit 22.

The input analyzer 12 is connected to the input device 8. The input device 8 includes a microphone 81, a voice recognition device 82,
And an auxiliary input device 83 such as a mouse / keyboard, which is used by a user to input a request into the system in the form of voice or text. The input analysis unit 12 is connected to the conversation control script storage unit 11 and can read out the control script stored in the conversation control script storage unit 11 as needed.

The information storage section 22 stores various information to be presented. The presentation control unit 22 effectively presents various types of accumulated information according to instructions described in the control script.

An output device 9 for actually presenting information to a user is connected to the presentation control unit 21.
Has a speech synthesizer 92, a speaker 93, and a display device 91.

The operation of the information presenting apparatus 1 will be described in more detail with reference to the operation flowchart of the input sentence analyzing unit according to the first embodiment of the present invention shown in FIG.

The user's voice input through the microphone 81 is converted by the voice recognition device 82 into character string data. For example, consider a case where a question sentence "Please search for a sunflower image (a)" is input.

The input sentence analysis unit 12 receives the character string data (step 11) and performs morphological analysis (step 1).
2) Generate a word pattern at the word level (step 13). When generating a word pattern, words obtained by morphological analysis are converted into basic forms. For example, the word pattern obtained for the character string data (a) is to search for a sunflower image (b). The word "search" has been converted to the basic form "search".

Next, the input sentence analyzing unit 12 stores the conversation control script storing unit 11 in which the word appearance pattern is described.
, And determines whether the word pattern matches the read template pattern (step 14).

Here, the control script stored in the conversation control script storage unit 11 will be described. The conversation control script storage unit 11 stores a template pattern used by the input sentence analysis unit 12 to determine a match with a word pattern, and a corresponding (conversation) control script.

The template pattern is, for example, a word string in the form of * sunflower * [search | show] * (c). The symbol * indicates that an arbitrary word applies, and [Search | Show] means "Search"
Or "show".

This template pattern represents a pattern in which "sunflower" and "search" or "sunflower" and "show" appear in this order.

On the other hand, the control script stored corresponding to the template pattern is an instruction that describes a process to be performed in the information presentation device 1 when a word pattern that matches the template pattern appears in a question sentence. is there. More specifically, the control script is composed of a response sentence presented to the question sentence, and a presentation control instruction including information to be presented simultaneously with the response sentence and a control instruction for the presentation control unit. You. For example, if the word pattern included in the question is a template pattern (c)
If it matches, the answer sentence contains only the sunflower video (d), and the presentation control instruction <Search Video sunflower; Explain sunflower> (e). This is an instruction to search for a video and a description about “sunflower”.

If the word pattern matches the template pattern in step 14, for example, in the above example, the input sentence analysis unit 12 converts the word pattern (b) “Please search for sunflower video” into the template pattern (c). If it matches “* sunflower * [search | show] *”, the input sentence analysis unit 12 outputs a control script corresponding to the matched template pattern,
For example, a control script including a reply sentence (d) and a presentation control command (e) is obtained (step 15), and the process is terminated.

If they do not match, the input unit analysis unit 12 determines whether there is a template pattern that has not been read yet in the conversation control script storage unit 11 (step 16). If there is a template pattern that has not been read yet, the process returns to step 14 to read another template pattern and determine whether or not it matches the word pattern.

When all the template patterns have been read, there is no template pattern that matches the word pattern, so the input unit analysis unit 12 determines that there is no template pattern corresponding to the question sentence, that is, Acquire a control script for presenting that there is no corresponding presentation information (step 1
7), the process ends.

In the description of the above embodiment, the template pattern that first matches the word pattern is selected. However, an attribute of importance is added to the template pattern.
It is possible to determine whether or not all template patterns match, and select the template panel with the highest importance among the matched template patterns.

The presentation control unit 21 acquires a control script from the input sentence analysis unit 12 and supplies presentation information to an output device. FIG. 8 shows a presentation control unit 2 according to the first embodiment of the present invention.
3 is an operation flowchart of FIG. Hereinafter, the operation of the presentation control unit 21 will be described with reference to FIG.

The presentation control unit 21 receives a response script and a control script including a presentation control command from the input sentence analysis unit 12 (step 21). Next, the presentation control unit 21 converts the control script into a response sentence and a presentation control instruction, for example, a response sentence (d) "This is the only image of the sunflower" and a presentation control instruction (e) "<Search Video sunflower; Expla
in sunflower>"(step 22).

The presentation control section 21 gives the reply message to the voice synthesizer 92 and the display device 91 of the output device 9 in order to present the reply message to the user in the form of a voice message and a screen display message (step 23).

The presentation control unit 21 analyzes the presentation control command, extracts the label of the data to be presented, and the timing and position at which the data is presented, and displays the control information such as the timing and the position on the output device 9. It is passed to the device 91 (step 24).

In the presentation command (e), "search" is a control command, "Video sunflower" means video data having an information label of sunflower, and "Explain".
"Sunflower" means an explanatory note (character data) having an information label of sunflower. In addition, a command for designating the data presentation timing and the presentation position on the screen may be added to the presentation command.

Next, in order to present the presentation data to the user via the display device 91, the presentation control unit 21 uses the label of the information to be presented described in the presentation control command to transmit the image data from the information storage unit 22. The data, image data, or character data is read and passed to the display device 91 (step 2).
5).

In the above description, step 23 and step 2
4 are processed sequentially, but steps 23 and 2
4 may be performed simultaneously or step 24 may be performed before step 23.

As described above, the presentation control unit 21 divides the control script into a reply sentence and a presentation control command, and presents the reply sentence to the user using the speech synthesizer and the display device in a screen and voice. The video, image, and character information are read from the information storage unit 22 using the label of the information to be presented described in the presentation control command, and presented to the user through the display device 91.

The information storage unit 22 retrieves, with a label, video, audio, image, text, etc. of a presumed inquiry field (for example, ecology of animals and plants) in order to answer an inquiry from a user. It is stored in an easy form.

For example, when the information storage unit 22 stores video data: sunflower video data: Asagao video data: Yuri image data: Sunflower image data: Asagao image data: Yuri Character data: Sunflower (explanatory text) Character data: Asagao (explanatory text) ) Character data: Consider the case where lilies (explanatory text) are stored. At this time, the presentation control unit 2
When 1 executes the presentation control instruction (e) “<Search Video sunflower; Explain sunflower>”, video data: sunflower character data: sunflower is acquired as presentation information.

Although the first embodiment of the present invention has been described using Japanese as an example of the natural language, the present invention is not limited by the language used. For example, a language processed by a speech recognition device and a speech synthesizer, a language handled in a natural language processing part such as a morphological analysis process of an input sentence analysis unit, and a language for presenting presentation information in a language other than Japanese such as English By changing to, the information presentation device and the information presentation system corresponding to various languages can be configured.

FIG. 9 is a configuration diagram of an information presentation system according to a second embodiment of the present invention. The second embodiment of the present invention differs from the first embodiment in that the conversation control script storage unit 11 of the first embodiment is replaced.

In the second embodiment of the present invention, a plurality of types of conversation control scripts are prepared according to the question answering field. For example, the conversation control script is divided into the fields of animals and plants, and the field of animals is further divided into dinosaurs, mammals, reptiles, and the like, and separately described. Then, by configuring the conversation control script storage unit 31 to switch and use the conversation control script according to the field of the question and answer, the search speed can be increased by reducing the size of each file of the conversation control script, Erroneous processing such as obtaining homonyms in other fields is avoided.

FIG. 10 is a configuration diagram of an information presentation system according to a third embodiment of the present invention. In the third embodiment of the present invention, an information presentation system that selects or restricts information presented to a user according to user information such as the user's desire, preference, or age is realized.

Therefore, when the user starts inputting information, the user inputs user information such as gender, age, or preference. After the user information is analyzed by the input sentence analysis unit 12, the user information is stored in the user information storage unit 31.

The conversation control script storage unit 11 adds, to the conversation control script, a control command for selecting and outputting a response sentence or presentation information according to user information. Then, when an instruction to make a selection according to the user information is added to the control script read from the conversation control script storage unit 11, the presentation control unit 21 reads the current user's current user information from the user information storage unit 31. It is configured to acquire user information and to limit the presentation information.

As a result, it is possible to provide presentation information that meets the user's wishes and to limit the presentation information to young people.

FIG. 11 is a block diagram of a video search system according to a fourth embodiment of the present invention. Hereinafter, an embodiment of the video search system will be described with reference to FIG. The video search system generally includes a natural language processing unit 100,
It comprises an information presentation unit 200 and a video database unit 300. The user sends a question to the web server 210 via the web browser 220 using the keyboard / mouse 230 and the voice recognition / synthesis unit 240,
The question is analyzed by the natural language processing program 110, and a response to the question and a video are presented using a CGI program or the like. The content group (video, still image) and the like presented by the video search system are stored in the video database unit 300 of the local web server 210 and presented in response to a request from a user. In addition, the use of an agent, for example, MS Agent, enables more effective presentation of information.

Note that the web browser 220 and the web browser 220
The server 210 can be connected in various forms such as Ethernet (registered trademark) and the Internet.
It is not limited by the connection form.

The natural language processing system of the present system employs pattern matching which is an extension of the above-described ELIZA pattern matching method. That is, the appearance pattern of a word in a conversation sentence is specified along with some metacharacters,
Conversation processing is advanced by matching with the input conversation sentence. Further, it is possible to control conditions using variables using a buffer memory, and a more advanced conversation pattern can be configured.

As a result, the natural language processing system of the present system can tolerate to some extent non-grammatical sentences that are features of spoken words, missing words, and conversations using fragmentary words. Particularly, it is characterized in that robustness and fast response to nongrammatical input when a question is input by voice recognition are realized.

Further, since the appearance pattern can be described using a simple template, conversation data can be easily described.

Hereinafter, the natural language processing program 111 will be described in detail. The natural language dialogue processing program of this embodiment is the agent TV (Audio-Visual Entertainment and Broadc
This is a Japanese language system of a natural language processing program for conversation (hereinafter abbreviated as NLI) used in asting: AVEB). Regarding agent TV, Murasaki et al., “Next-generation TV entertainment system using agents”, IEICE Technical Report, OFS2000-23,
(2000).

The NLI obtains a user's profile (name, preference, etc.) and controls the system through dialogue with the user on the agent television. In the present embodiment, video search targets and inquiry information are extracted by applying this NLI.

The language processing in NLI is different from a general interactive processing system and is limited to a morphological analysis level. Then, matching is performed between the extracted morpheme group and the set pattern, and when a match is found, predetermined processing (such as a response) is performed.

Approaches to extract search terms by surface analysis of words, for example, Takano et al., “Use of Dictionary Information in Interactive Document Search Using Natural Language”
IEICE Technical Report, NLC2000-7, pp49-54, (20
The approach described in (00) is considered to be suitable when the input from the user in the natural language does not accurately express the search condition or the request.

A template is defined by a character string including words divided into morpheme levels and metacharacters that abstractly represent their appearance patterns. Then, an inquiry sent from the user (question sentence) is compared with the defined template to determine whether they match, and if they match, the operation described in relation to the template is executed. A collection of a plurality of these templates is the template dictionary files 121 and 122. The template dictionary file corresponds to the knowledge information of the interactive processing. A plurality of template dictionary files may be provided so that the template dictionary files can be switched and used depending on the theme of the conversation. Further, the interactive processing is controlled using the synonym dictionary 124, the word dictionary 123, the variable definition file 125, and the like.

Next, each file used by the NLI for performing the interactive processing will be described in detail.

(1) Template dictionary file The rule elements used in the template include the following (i)
To (iv). FIG. 12 is a diagram illustrating the configuration of the template pattern. Elements (i) to (iv)
Can be used to realize the configuration of the template pattern shown in FIG.

(I) Pattern FIG. 13 is a table for explaining metacharacters. A template pattern is formed by combining the metacharacters described in this chart. The words used in this template are also morphologically analyzed and compared with the morphologically analyzed pattern of the query sentence.

(Ii) Variable This represents a variable that uses the buffer memory. Used in the conditions and actions described below.

(Iii) Condition The left and right sides of the conditional expression are compared, and if true, the action is executed. In the conditional expression, equal sign == and inequality sign! = And are used. A plurality of conditions can be described by combining logical operations and and or. (Example) c.abc == value (iv) Action Assign the value on the right side to the variable on the left side. (Example) a.abc = value Goto and control are special actions.

As shown in the explanatory diagram of template switching by goto in FIG. 14, goto moves between a plurality of template dictionary files to switch template dictionary files.

Control is an action for transmitting a character string assigned to control to a program when the NLI processing system is called from another program.

(V) Importance The importance of the template pattern is described numerically. The higher the numerical value, the higher the importance, and the same template is preferentially matched. If the importance is not described, it is treated as 0. If a plurality of patterns match, the one with the longer pattern length is prioritized. (Example) w.7 (iv) Response sentence (and control) This is the character string returned when the pattern matches, and is the sentence that the system answers. In addition, it can include a video presentation to an information presentation unit described later and an operation control command of a CG character. (Example) Answer sentence <play Video video1 100 200 400 200> (2) Word dictionary file In order to prevent writing many words in the template in parallel, the word group is written in the word dictionary file indicated by @WordFileName as follows. Can be described as follows. (Example) File name: fishName In the Ayu salmon trout template, the word group is represented as @W
Specify with ordFileName.

* @FishName * Can be eaten * (3) Synonym dictionary file This file is used to unify the notation of a plurality of words having the same meaning and has the following structure. (Example) Ayu ← Ayu: Ayu: Perfume Synonym processing is performed at the beginning of the analysis, and if there is the same thing as the second and subsequent words in the colon-separated word list, it is replaced with the first word in the list and the language Processing continues.

(4) Variable definition file This is a file for declaring variables used in the template.

Next, the search interface will be described.

The user interface system for video search learning of this embodiment is configured using the above-described natural language processing program. The basic function of this user interface system is to process text input by the user, and to write the actions described in the matching template,
Specifically, it is to perform video search and information presentation.

For example, when an inquiry request "What kind of bird is a gull?"
Search for video and information using as a search keyword, and present the results to users as Web content. The processing to be performed after the keyword is extracted can be defined in the response sentence of the natural language processing unit.

In this embodiment, as shown in FIG.
In the information presenting unit 200, moving images, still images,
A web browser 220 is used to display characters.

The basic display data (HTML, CGI) is stored in the program production system under study DTPP (Desk-Top Program Productio).
The program information (program structure, video captions, and announcement comments) used in n) are converted and used. The program production system DTPP
Sumiyoshi et al., "Program Information Structure in Hierarchical Program Production Method", IEICE General Conference, SMD-2,
(1998). With DTPP, programs can be designed and edited structurally. This program information is converted into XML and converted into HTML data for displaying a page. FIG.
FIG. 4 is a conceptual diagram of a user interface screen displayed using the HTML data generated in this manner. The caption information added to the video at the time of editing is displayed as a description of the video and is also used as search data.

The information presenting unit 200 sends the question sentence 223 input by the user to the natural language processing unit 100, displays the resulting answer sentence 224 on the screen, and communicates with the agent (for example, MS Agent) 222 by voice. Synthetic software 24
By controlling 0, a response sentence is uttered, and information is effectively presented by pointing video and image information.

The agent 222 incorporates an agent operation program of a specific pattern into the CGI script, and calls a program that calls the program into the natural language processing unit 10.
By describing them in the template dictionary files 121 and 122 of “0”, when the agent gives an explanation, it is controlled so as to simultaneously point to an image.

The video database unit 300 can be configured by various methods. In this embodiment, a display control function of a web page using JavaScript is used for simplicity.

The caption to the video displayed on the screen, embedded in the web page, is used as search index data, extracted as search index data by the natural language processing unit 100, and extracted as a search keyword by the natural language processing unit 100. The search is performed to determine whether or not the image includes the search term, and only the video corresponding to the caption including the search keyword is displayed on the screen as the search result.

[0085]

[Embodiment] Finally, an application example of a video search system according to a fourth embodiment of the present invention will be described. In this application example, the dictionary file (template group) was created by manual input assuming a conversation.

The assumed conversation is for retrieving an image of an object of interest and, if there is an image and information related thereto, interactively presenting information about the object. More specifically, it interactively searches for images related to river creatures and presents various information related to the images in an explanatory manner. If the subject is an animal, information such as where it lives and what it eats is presented.

For example, when searching for a black-headed gull video, information about the black-headed gull to be searched is presented interactively before and after the search. At this time, the supposed inquiry is as follows.・ What kind of bird is the black-headed gull?・ Where do black gulls live?・ I want to see an image of a black-headed gull.

The contents presented to the user from the system prepare the following information corresponding to the assumed inquiry.・ Name; Official name ・ Subject: What subject, What item ・ Feature: Shape, color, size ・ Habitat: Area, place, environment ・ Food: What These information is once about habitat and food A template was created to respond to inquiries in which the entered subject term "black gull" was omitted.

For example, [What kind of bird is a black-headed gull? ] Template: * Black-headed gull *
When the input sentence matches the bird *, the system replies to this question, and at the same time, stores “Black Gull” in the buffer memory subject, and holds the current topic, that is, the topic assumed by the matched template.

If the question consecutively followed by this question is “Where do you live?”, It is assumed that the target is the current topic, and the variable of the topic word is $$$.
Check the subject and match the following template.

* Where * live * c. $$ subject == Black-headed gull On the other hand, when the question "Where do you live?" Is input first, the variable question is set to "Where do you live?" Is returned to the user. As a response to this, the following template is prepared assuming "black-headed gulls" and made to answer.

* Black-headed gull * c. $$ question == Where you live> Black-headed gulls live in XX.

As described above, the topic is assumed to be the subject of the question, and the lack of the subject which is likely to occur in the conversation sentence is dealt with. Also, the replacement of the word order is dealt with by registering a plurality of templates. If the topic moves to another topic, the variable
Clear subject and question.

Although the template dictionary serving as conversation data can be easily constructed, a large amount of description is required to have a comprehensive conversation, and knowledge of the system is required.

Therefore, by classifying the information to be resigned and inputting it in a predetermined format, it becomes possible to generate a template group that can cope with an assumed conversation relating to this information.

In the case of information such as a video database relating to animals, the information and inquiry sentence of the object are both
Since they can be limited, a template group that can cope with the abbreviation of the subject or the change of the word order can be generated by simply giving necessary information. This technique is effective for interactive processing for supplementing information of a video database with limited use.

The information presenting system according to the embodiment of the present invention is not limited to the example described in the above embodiment, but may be an input / output device in the system, a communication device, or a hardware such as a CPU or a memory of a computer. The parts other than the hardware part are constructed with software (programs) that realize the same functions, recorded on a disk device or the like, and installed on a computer as needed, or downloaded to a computer via a network to obtain information. It is also possible to provide a presentation service. Further, the constructed program is stored in a portable recording medium such as a floppy (registered trademark) disk, a memory card, a CD-ROM, and the like.
Such a service can be used for general purposes in a scene where the service is used.

Although the representative embodiments of the present invention have been described above, the present invention is not limited to the above embodiments, and various modifications and applications are possible within the scope of the claims.

[0099]

As described above, according to the present invention, the input of the information presentation request in the natural language is analyzed, and the required information and the presentation method are extracted. Since this operation is performed, a simpler operation is realized for the user than in the related art, so that it is possible to improve the operability of a video presenting apparatus using a computer such as a video search.

Further, according to the present invention, the control of information presentation is described as a group of instructions, and is presented in combination with various types of material information, so that the degree of freedom of expression is enhanced and a more diverse presentation method can be adopted. And the amount of information to be stored is reduced.

Furthermore, the present invention has a remarkable effect when combined with digital broadcasting. This is because a digital broadcast data channel can broadcast text, text information describing the video, image, and video information together with the video. Also, by using a home server that collectively stores these broadcast information,
Information can be easily re-viewed. In addition, the huge amount of video stored in the broadcasting station is converted into a video database, so that not only information stored on the terminal side such as a home server but also information and video stored on the broadcasting station side through a network Can be searched from the terminal.

The information presentation system based on natural language dialogue according to the present invention can construct a new content service, for example, a learning application using a video database that supports spontaneous research and study, with such technology as the background. .

In particular, as a video database, video information centering on science and history is accumulated, and users investigate their own interests and problems with their own powers. Suppose you have an environment where you can learn through the work process of finishing and presenting. Considering a situation where a video database is used in such an environment, learning does not proceed very effectively only by providing a conventional search tool using a simple keyword index.
The user interface provided by the present invention is a user interface that can be applied to such a scene, in which a learner performs an operation such as an inquiry, a question, or a video search for an object of interest by using a natural language dialogue. The content and various accompanying knowledge can be acquired, and at the time of retrieval, learning is further enhanced by presenting not only the video but also various information related to the video.

[Brief description of the drawings]

FIG. 1 is a principle configuration diagram of the present invention.

FIG. 2 is a configuration diagram of request information input means of the information presentation device of the present invention.

FIG. 3 is a configuration diagram of presentation information creation means of the information presentation device of the present invention.

FIG. 4 is an explanatory diagram of an information presentation method performed by the information presentation device of the present invention.

FIG. 5 is a configuration diagram of an information presentation system including the information presentation device of the present invention.

FIG. 6 is a configuration diagram of the information presentation system according to the first embodiment of the present invention.

FIG. 7 is an operation flowchart of the input sentence analyzing unit according to the first embodiment of the present invention.

FIG. 8 is an operation flowchart of the presentation control unit according to the first embodiment of the present invention.

FIG. 9 is a configuration diagram of an information presentation system according to a second embodiment of the present invention.

FIG. 10 is a configuration diagram of an information providing system according to a third embodiment of the present invention.

FIG. 11 is a configuration diagram of a video search system according to a fourth embodiment of the present invention.

FIG. 12 is an explanatory diagram of a configuration of a template pattern according to a fourth embodiment of the present invention.

FIG. 13 is an explanatory diagram of metacharacters according to a fourth embodiment of the present invention.

FIG. 14 is an explanatory diagram of template switching according to a fourth embodiment of the present invention.

FIG. 15 is an example of a user interface screen according to a fourth embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Information presentation system 8 Input device 9 Output device 11 Conversation control script storage means 12 Input sentence control means 21 Presentation control means 22 Information storage means 81 Microphone 82 Voice recognition device 83 Auxiliary input device 91 Display device 92 Voice synthesizer 93 Speaker

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor: Kana Fuchi, 1-10-11 Kinuta, Setagaya-ku, Tokyo Japan Broadcasting Corporation Research Institute (72) Inventor: Masahiro Shibata, 1-10 Kinuta, Setagaya-ku, Tokyo No. 11 Japan Broadcasting Corporation Broadcasting Research Laboratory (72) Inventor Seiki Inoue 1-1-10 Kinuta, Setagaya-ku, Tokyo Japan Broadcasting Corporation Broadcasting Research Laboratory (72) Inventor Hirokazu Konishi 1-kinuta, Setagaya-ku, Tokyo 10-11 Japan Broadcasting Corporation Broadcasting Research Institute (72) Kinji Matsumura Inventor 1-10-11 Kinuta, Setagaya-ku, Tokyo Japan Broadcasting Research Institute F-term (reference) 5B075 ND16 NK32 PP02 PP03 PP07 PP24 PQ02 5E501 AA01 AC25 AC33 BA03 BA05 BA17 CA02 CA08 CB02 CB09 CB15 EA21 FA13 FA14 FA32

Claims (3)

[Claims] Request information input means for receiving a question sentence in which request information is expressed in a natural language, analyzing the question sentence, and acquiring response control information corresponding to the analyzed question sentence; And a presentation information producing means for producing presentation information by combining material information according to the acquired response control information to respond to the request information. 2. The request information input means receives the question sentence, morphologically analyzes the question sentence, and creates a word pattern corresponding to the question sentence, and a template corresponding to the request information. A conversation control script storing unit that stores a pattern and a conversation control script describing a process to be performed when the template pattern appears in the question sentence; wherein the input sentence analysis unit stores the conversation control script. 2. The method according to claim 1, further comprising: detecting a template pattern corresponding to the created word pattern from the means; and acquiring a conversation control script stored corresponding to the detected template pattern as the response control information.
Information presentation device as described. 3. A computer-readable recording medium storing an information presentation program for providing presentation information according to request information, wherein the program is configured to convert a question sentence expressed in a natural language requesting information into a morpheme. Analyzing, a process of creating a word pattern from the words of the question sentence obtained by the morphological analysis, a process of detecting a template pattern that matches the word pattern from template patterns stored in advance, A process of acquiring a conversation control script that is stored corresponding to the detected template pattern and describes a process to be performed when the template pattern appears in the question message; Specify the presentation information that is presented for the question sentence The material information combined according to the presentation control instruction,
A process for producing the presentation information including the response sentence.