KR20080052279A - Apparatus and method of dialogue tv agent service for providing daily information - Google Patents

Abstract

An interactive TV agency service method for providing real-life information and an apparatus therefor are provided to acquire real-life information updated on a web in real time and expand the acquired information to various region when necessary by performing operation for which a user wishes through an interactive interface including a voice, a remote controller, and a text input and providing a corresponding suitable response. A dialogue sentence analyzing unit(20) analyzes an input sentence received from a user. A dialogue management unit(30) analyzes the request of the user by using the analyzed result of the dialogue sentence analyzing unit. When a real-time information update is requested from the dialogue management unit, a real-time automatic knowledge constructing unit(70) extracts daily life information updated on a web in real time and stores the extracted information in a daily information database. A response generating unit(40) generates a response for the request of the user analyzed in the dialogue management unit and provides the response to the user. When it is impossible to respond to the request of the user or an error is generated in the response for the request of the user, an exception processing unit(80) recovers a system to allow the user to cancel the request of the user.

Description

Apparatus and method of dialogue TV agent service for providing daily information

The present invention provides an appropriate response to user requests input through various interfaces, and provides daily life information through an interactive TV agent by acquiring and realizing a knowledge source by obtaining various life information existing on the web in real time. A method and apparatus for

The present invention is derived from a study conducted as part of the IT new growth engine core technology development project of the Ministry of Information and Communication and the Ministry of Information and Telecommunication Research and Development. [Task Management No .: 2006-S-036-01, Task name: Large-capacity interactive for new growth engine industry] Development of distributed processing voice interface technology].

Republic of Korea Registration No. 10-0580619 is a method and apparatus for quickly and accurately managing the dialogue between a person and an agent using multi-modal information such as voice information, facial expression information and time delay information of the user, and a voice dialogue system using the same It is starting. The technique can draw the conversation accurately using the multi-modal information, but has a disadvantage that it cannot be used in an environment in which the multi-modal information is not available.

In addition, as a related art, there is disclosed a speech recognition system that allows an interactive TV to automatically select a channel in response to a user's voice request, and a conversation system that can obtain weather information from around the world from a telephone using a voice conversation. The disadvantage is that it uses only voice recognition and provides only a single area of information such as TV or weather.

 An object of the present invention for solving the above problems is that unlike the prior art provides only a single area information (for example, TV guide) to perform the operation desired by the user through a conversation interface including a remote control, text input as well as voice and And, by presenting the appropriate response, a conversation that enables you to obtain real-time information (e.g., TV guides, weather, schedules, news, stocks, traffic information, etc.) updated on the web in real time and expands to various areas as needed There is provided a type TV agent service apparatus and a method thereof.

According to an aspect of the present invention, there is provided a method of providing an interactive TV agent for providing real-life information, comprising: analyzing input text received from a user; Analyzing a user's request based on the analyzed input statement; Extracting daily life information updated on the web in real time and storing the same in a daily information database; Extracting information corresponding to the analyzed user's request from the daily information database; And outputting the extracted information through a TV.

      Preferably, the input text received from the user is input through the user's voice recognition, remote control operation or string input.

      The daily information database receives daily information that is updated on the web or personal schedule information that is updated in a personal information terminal in real time and builds a database for each information area. Preferably, the information area includes a TV guide, news, weather, personal schedule.

     Analyzing the input text received from the user, if there is an error in the received input text, correcting the input text; Performing language analysis through morphological analysis, entity name recognition, and syntax analysis of the input text; And extracting a speech act and a concept string from the language-analyzed input sentence based on a pre-stored speech act knowledge and a concept string knowledge database.

      The analyzing of the user's request may include analyzing an area of information desired by the user based on the analyzed input sentence and normalizing the input sentence; Analyzing abbreviations and substitute terms of the input sentence with reference to previously stored conversation data; Analyzing a user request using the dialogue act, the concept sequence, the information area, the substitute term, and the abbreviation word; Selecting an action corresponding to the analyzed user request; If the response to the user request is not available, causing the user to verify the user request; And if there is an error in the response to the user request, allowing the user to cancel the user request.

      The outputting of the extracted information through the TV is performed through any one of voice, TV screen, and TV control signal transmission.

      In accordance with another aspect of the present invention, there is provided a real-time information providing interactive TV agent service apparatus comprising: a dialogue analysis unit analyzing an input text received from a user;

A conversation manager for analyzing a user's request by using the analysis result of the conversation analyzer; A real-time automatic knowledge construction unit for extracting real-time information updated on the web in real time and storing it in a daily information database when a real-time information update request is received from the conversation manager; A response generation unit generating a response to the request of the user analyzed by the conversation manager and providing a response to the user; And an exception processing unit for enabling the user to cancel the user request and restoring the system when the user request is impossible to respond or when there is an error in the response to the user request.

      Preferably, the dialogue analysis unit comprises: an input signal recognition unit recognizing an input sentence received from a user; An error processing unit for correcting the input text when the input text has an error; A language analyzer configured to perform morphological analysis, entity name recognition, and syntax analysis on the input text; And a speech act and concept string analysis unit that analyzes the speech act and the concept string of the input statement.

       Preferably, the conversation manager comprises: a multi-region recognition unit for identifying an area of information desired by a user using a dialogue act and a concept string of the input sentence; A sentence normalizer for normalizing the input sentence; Discourse analysis unit for analyzing the substitute words and abbreviations of the input statement with reference to the previously stored conversation content data; A user request grasp unit for analyzing a user request using the dialogue act, the concept sequence, the information area, the substitute term, and the abbreviation word; And an operation instruction unit for instructing an operation according to the analyzed user request.

      Preferably, the response generator comprises a voice response generator for providing a voice response when the user request is analyzed to be acquisition of a single information; A screen response generation unit for providing a video response through a TV screen when the user request is analyzed to be acquisition of a plurality of information or video information; And a control signal generator for transmitting a control signal to the TV when the user request is analyzed to be TV control.

Preferably, the exception processing unit includes a dialog verification unit for verifying whether the user inputs a question answerable; An operation verification unit for determining whether an operation requested by the user is executable and allowing the user to verify the request of the user when the user request is not executable; And an error recovery unit for allowing a user to cancel the user request and restoring the system when there is an error in response to the user request.

Preferably, the real-time automatic knowledge building unit Web collection unit for collecting the web document in real time; A web document processing unit for parsing and analyzing web documents collected in the web document collecting unit to extract daily information; A terminal input / output unit configured to receive personal schedule information stored in the personal terminal; A personal information processing unit which extracts valid information by analyzing the information stored in the terminal input / output unit; A knowledge extracting unit for extracting daily life information from the information processed by the web document processing unit and the personal information processing unit; And an area knowledge storage unit that stores the information extracted by the area knowledge extraction unit as structured information.

As described above, according to the present invention, the user can conveniently and easily obtain daily life information and perform a desired action by conducting a natural conversation with the interactive TV through various types of conversation interfaces, and updating daily on the web. Various new information can be collected in real time. If a malfunction is performed, it can be easily restored and various types of exception processing can be performed.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

 1 is a block diagram showing the configuration of an embodiment of a real-time information providing interactive TV agent device according to the present invention.

     Referring to FIG. 1, the interactive TV agent device of the present invention includes a user input unit 10, a conversation analysis unit 20, a conversation management unit 30, a response generation unit 40, a query response unit 50, and real-time automatic The knowledge constructing unit 70 and the exception processing unit 80 are included.

      The user input unit 10 receives an input signal of at least one of a voice signal, a remote control signal, and a string signal from a user and transmits the input signal to the conversation text analyzing unit 20. The conversation analysis unit 20 analyzes the input signal and sends the language to the conversation management unit 30. The conversation management unit 30 uses the analysis result of the conversation analysis unit 20 to grasp the user's request and then the question answering unit ( 50) ask the correct answer. The question and answer unit 50 searches the daily information database 60 to extract the correct answer required by the user and sends it to the conversation manager 30. The conversation manager 30 provides a correct answer to the user through the response generator 40.

      When the response to the input statement input by the user is not possible, the exception processing unit 80 verifies the input statement by requesting the user for confirmation of the input statement. If an error occurs in the system operation, the exception processing unit 80 allows the user to cancel the operation or command and restores the system.

      The real-time automatic knowledge construction unit 70 extracts the daily life information updated on the web in real time when there is a real-time information update request from the conversation management unit 30 and updates it in the conversation management unit 30 and the dialogue text analysis unit 20. Communicate information The stored update information is constructed as a daily information database 60 for each area such as TV guide, weather, news, schedule, and the like, and is used for extracting the correct answer of the question answering unit 50.

      2 is a block diagram showing the detailed configuration of the dialogue analysis unit 20, the dialogue management unit 30 and the response generator 40 shown in FIG.

      Referring to FIG. 2, a detailed operation of the dialogue analyzing unit 20 will be described. First, the input signal recognizing unit 201 recognizes a voice, a remote control, a text signal, etc. input by a user. Includes natural (continuous) and isolated words. The following is an example of natural language (continuous word) and isolated word input.

-Natural language (continuous language) input sentence: What do you do drama on SBS tomorrow? Find a pro that comes out Yoo Jae Suk.

-Isolated phrase: SBS drama search tomorrow. Yoo Jae Suk Pro Search.

If there is an error in the input statement, the error processing unit 202 corrects the error and extracts the corrected input statement. The language analysis unit 203 performs linguistic analysis such as morpheme analysis, entity name recognition, and syntax analysis on the input text, and the speech act and concept string analysis unit 204 is based on a pre-stored speech act knowledge and concept string knowledge database. As a result, the speech act and the concept string, which are types of information desired by the user, are analyzed from the language-analyzed input statement. The processing result of the analyzed input text is transmitted to the conversation manager 30. Examples of speech acts and concept strings analyzed by pre-stored speech act knowledge and concept string knowledge data are as follows.

-User input: How is the weather in Daejeon today?

-Result of analysis of concept sequence: <Today: date> <Daejeon: location> How is the weather?

-Speech act analysis result: search_weather

      Referring to Figure 2 describes the detailed operation of the conversation management unit 30,

The multi-domain recognition unit 301 analyzes the user's desired daily life information using analyzed speech act and concept string information, and the sentence normalization unit 302 normalizes the input sentence. The discourse analysis unit 303 performs discourse analysis with reference to previously stored conversation flow data. The multi-domain recognition 301 and sentence normalization 302 are performed on the input text currently input, while the discourse analysis 303 reinforces the referenced information such as substitute words and abbreviations based on previously stored conversation contents. do.

      The user request grasping unit 304 grasps the user's needs using speech acts, concept strings, area information, substitute terms, and abbreviations extracted through speech act and concept string analysis 204 and discourse analysis 303. The user request is analyzed in the form of a frame and transmitted to the operation instruction unit 305. The following is an example of the final dialogue analysis result transmitted to the operation instruction unit 305.

-Previous ignition: back to KBS2.

Current speech: What are you doing this weekend?

Conceptual analysis: <Weekend: date> <Drama: genre> What are you doing?

Speech act analysis result: search_program

Area Recognition: TV Guide Area

Result of normalization of sentence: <This weekend = 2006/10/07 ~ 2006/10/08>

-Discourse Analysis: <In KBS2: Abbreviation> What are you doing this weekend?

Identify your intent:

frame slot value Program ? Channel KBS2 Date 2006/10/07 ~ 2006/10/08 Time ?

     When the user wants to obtain information by using the final analysis result of the input statement, the operation instruction unit 305 retrieves the correct answer by calling the query response unit 50, and verifies that the user's request is valid and an error occurs in the system. If so, the exception processing unit 80 is called. In addition, the final result of performing the desired operation is delivered to the user through the response generation unit 40, and performs a real-time automatic knowledge construction unit 70 and feedback to update the daily information in real time.

     Referring to FIG. 2, the detailed operation of the response generator 40 will be described. When the user performs a conversation for the purpose of obtaining single information, the sentence generator 401 generates a natural response and generates a voice response. The synthesis unit 403 presents the synthesized sound to the user. When a dialog is performed for the purpose of output presentation such as a plurality of information acquisition or news, the screen response generator 402 presents the image information on the TV screen. When the conversation for the purpose of controlling the TV, such as changing the TV channel, program recording reservation, changing the sound volume, etc. is performed, the control signal generator 404 transmits the TV control signal. The following is an example of the response generation result according to the dialog input by the user.

Example 1) User Input: Where do you play baseball today?

Dialogue Response: The Korean Series 3rd Game will air at 6 pm on KBS2.

Example 2)-User Input: Show me some economic news today.

Dialogue response: October 23, 2006 Economic News. Look at the screen.

Example 3)-User Input: Return to KBS2.

-Chat response: (change channel).

      3 is a block diagram showing the detailed configuration of the real-time automatic knowledge construction unit shown in FIG. When the real-time automatic knowledge construction unit 70 receives a real-time information update request from the conversation manager 30, the real-time automatic knowledge construction unit 70 acquires, processes, stores, and expands the multi-area daily life information updated on the web in real time.

      Referring to FIG. 3, the detailed operation of the real-time automatic knowledge building unit 70 will be described. The web document collecting unit 701 collects a web document including new information updated daily, in real time, and the web document processing unit 702 Parsing and analyzing the collected web documents to extract the necessary daily information. The personal schedule information stored in the PDA or the mobile phone is collected through the terminal input / output unit 703 and the personal information processing unit 704.

     The daily life information collected from the web document or the personal terminal is extracted through the knowledge extraction unit 705 for each region and stored as structured information through the knowledge storage unit 706 for each region. The stored information is constructed as a daily information database 60 for each area such as TV guide, weather, news, schedule, and the like, and is used for extracting the correct answer of the question answering unit 50.

      The updated information is transmitted to the input signal recognition unit 201 of the conversation analysis unit 20 by performing feedback through the conversation management unit 30 to inform the whole system of new knowledge update information.

      The daily information database 60 has a structure that is scalable, and can respond to requests for addition and expansion of a new area service other than the existing area.

      4 is a block diagram showing the detailed configuration of the exception processing unit shown in FIG.

The exception processing unit 80 instructs the user to respond to a command that cannot be executed or executes, or restores and verifies the system when the system malfunctions.

     Referring to Figure 4 describes the detailed operation of the exception processing unit 80,

The conversation verification unit 801 verifies whether the user inputs a question that can be answered using the conversation frame analyzed by the conversation management unit 30, and the operation verification unit 802 executes a command for performing a final task requested by the user. Verify that it is possible. If the user attempts a response or an impractical conversation, the system verifies the user by verifying that the user input statement is correct.

     The error recovery unit 803 allows a user to cancel a user request and restores the system when the system malfunctions. When the system is restored through the error restoration unit 803, the system is restored to the state before the current ignition of the user. The following is an example of verifying user request operation and recovering error.

Example 1)-User input 1: What did you do at 10 pm at MBC yesterday?

Dialogue 1: Jumong was broadcast.

User Input 2: Record that.

-Chat Response 2: It is the operation for the past time program and therefore cannot be performed. Please check and speak again.

Example 2)-User Input 1: Recorded 'Human Theater' today.

      -Dialogue 1: We will schedule a recording of the ‘Human Theater’ program to be broadcast at 9 pm on KBS2 tonight.

-User input 2: No that cancel.

-Chat Response 2: The recording schedule for 'Human Theater' has been cancelled.

      The present invention has been described above with reference to the preferred embodiments, and the person skilled in the art to implement the present invention may be embodied in a modified form within an equivalent range without departing from the essential characteristics of the present invention. It should be interpreted that it belongs to the scope of the invention.

Meanwhile, the above-described embodiments of the present invention can be written as a program that can be executed in a computer, and can be implemented in a general-purpose digital computer that operates the program using a computer-readable recording medium. The computer-readable recording medium may be, for example, a magnetic storage medium such as a ROM, a floppy disk, a hard disk, or the like, an optical read medium such as a CD-ROM, a DVD, or the like, and a carrier such as a transmission through the Internet. It includes a storage medium such as a wave.

1 is a block diagram of an interactive TV agent according to the present invention.

FIG. 2 is a block diagram showing the detailed configuration of the dialogue analyzing unit, the dialogue managing unit, and the response generating unit shown in FIG.

3 is a block diagram showing the detailed configuration of the real-time automatic knowledge building unit shown in FIG.

4 is a block diagram showing the detailed configuration of the exception processing unit shown in FIG.

Claims (15)

Analyzing the input text received from the user; Analyzing a user's request based on the analyzed input statement; Extracting daily life information updated on the web in real time and storing the same in a daily information database; Extracting information corresponding to the analyzed user's request from the daily information database; And Outputting the extracted information through a TV; Real-life information providing interactive TV agent service method comprising a.       The method of claim 1, The input received from the user A real-time information providing interactive TV agent service method characterized in that the input received through the user's voice recognition, remote control operation or string input. The method of claim 1, wherein the daily information database A real-time information providing interactive TV agent service method for receiving daily information updated on the web or personal schedule information updated in a personal information terminal in real time and storing the information in each information area.       The method of claim 3, wherein The information area is a real-time information providing interactive TV agent service method including a TV guide, news, weather, personal schedule. The method of claim 1, Analyzing the input text received from the user Correcting the input text when there is an error in the received text; Performing language analysis through morphological analysis, entity name recognition, and syntax analysis of the input text; And Extracting speech acts and concept strings from the linguistic analysis input sentences based on prestored speech act knowledge and concept string knowledge databases; Real-life information providing interactive TV agent service method comprising a. The method of claim 5, Analyzing the user's needs Analyzing a region of information desired by a user based on the analyzed input sentence and normalizing the input sentence; Analyzing abbreviations and substitute terms of the input sentence with reference to previously stored conversation data; Analyzing a user request using the dialogue act, the concept sequence, the information area, the substitute term, and the abbreviation word; And Selecting an action corresponding to the analyzed user request Real-life information providing interactive TV agent service method comprising a. The method of claim 6, If the response to the user request is not available, causing the user to verify the user request; And Enabling the user to cancel the user request if there is an error in response to the user request Real-life information providing interactive TV agent service method further comprising. The method of claim 7, wherein Outputting the extracted information through a TV A real-life information providing interactive TV agent service method, characterized in that it is made through any one of voice, TV screen and TV control signal transmission. A computer-readable recording medium having recorded thereon a program for executing the method according to any one of claims 1 to 8. A dialogue analyzer for analyzing an input sentence received from a user; A conversation manager that analyzes a user's request by using an analysis result of the conversation analyzer; A real-time automatic knowledge construction unit for extracting real-time information updated on the web in real time and storing it in a daily information database when a real-time information update request is received from the conversation manager; A response generation unit generating a response to the request of the user analyzed by the conversation manager and providing a response to the user; And An exception processing unit that enables the user to cancel the user request and restores the system when the user request is impossible to respond or when there is an error in the response to the user request. Real-life information providing interactive TV agent service device comprising a. The method of claim 10, wherein the dialogue analysis unit An input signal recognition unit recognizing an input text received from a user; An error processing unit to correct the input text when the input text has an error; A language analyzer configured to perform morphological analysis, entity name recognition, and syntax analysis on the input text; And Speech act and concept string analysis unit for analyzing the dialogue act and concept string of the input statement Real-life information providing interactive TV agent service device comprising a. The method of claim 11, wherein the conversation management unit A multi-domain recognizing unit which grasps an area of information desired by a user using a dialogue act and a concept string of the input text; A sentence normalizer for normalizing the input sentence; A discourse analysis unit for analyzing substitute terms and abbreviations of the input sentence with reference to previously stored conversation contents data; A user request grasp unit for analyzing a user request using the dialogue act, the concept sequence, the information area, the substitute term, and the abbreviation word; And An operation instruction unit for instructing an operation according to the analyzed user request Real-life information providing interactive TV agent service device comprising a. The method of claim 12, wherein the response generating unit A voice response generating unit for providing a voice response when the user request is analyzed to be acquisition of single information; A screen response generation unit for providing a video response through a TV screen when the user request is analyzed to be acquisition of a plurality of information or video information; And And a control signal generator for transmitting a control signal to a TV when the user request is analyzed to be TV control. The method of claim 13, wherein the exception handling unit A dialog verification unit for verifying whether a user inputs a question that can be answered; An operation verification unit for determining whether an operation requested by the user is executable and allowing the user to verify the request of the user when the user request is not executable; And If there is an error in response to the user request, the error recovery unit for allowing the user to cancel the user request and restore the system Real-life information providing interactive TV agent service device comprising a. 15. The method of claim 14, wherein the real-time automatic knowledge building unit A web document collecting unit collecting the web document in real time; A web document processing unit for parsing and analyzing web documents collected in the web document collecting unit to extract daily information; A terminal input / output unit configured to receive personal schedule information stored in the personal terminal; A personal information processing unit which extracts valid information by analyzing the information stored in the terminal input / output unit; A region-specific knowledge extracting unit for extracting daily life information from the information processed by the web document processing unit and the personal information processing unit; And Real-time information providing interactive TV agent service apparatus comprising a knowledge storage for each area for storing the information extracted by the knowledge extraction for each area as structured information.

Images