Apparatus for providing voice dialogue service and method of operating the same

• US 20070208556A1
• Filed: 08/28/2006
• Published: 09/06/2007
• Est. Priority Date: 03/03/2006
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Cho, Jeong Mi, Kang, In Ho, Kwak, Byung Kwan
• Granted Patent: US 7,734,461 B2
• Time in Patent Office: Days
• US Class Current: 704/9
• CPC Class Codes: G06F 40/205 Parsing
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A speech dialogue service apparatus including: a language analysis module tagging a part of speech (POS) of each respective word included in a sentence recorded in a predetermined text, syntactically analyzing the sentence by classifying a meaning of each respective word, and generating at least one semantic frame corresponding to the sentence according to a result of the syntactical analysis; and a dialogue management module analyzing an intention of the sentence corresponding to the at least one respective semantic frame, and generating a system response corresponding to the sentence intention by selecting a predetermined sentence intention according to whether an action corresponding to the intention of the respective sentence can be performed.

Claims

1. A speech dialogue service apparatus comprising:

a language analysis module tagging a part of speech (POS) of each respective word included in a sentence recorded in a predetermined text, syntactically analyzing the sentence by classifying a meaning of each respective word, and generating at least one semantic frame corresponding to the sentence according to a result of the syntactical analysis; and

a dialogue management module analyzing an intention of the sentence corresponding to each at least one semantic frame, and generating a system response corresponding to the sentence intention by selecting a predetermined sentence intention according to whether an action corresponding to the intention of the respective sentence can be performed.

2. The apparatus of claim 1, further comprising a speech recognition module recognizing an utterance of a user and converting speech of the utterance into a text,

wherein the text comprises at least one natural language uttered by the user.

3. The apparatus of claim 1, wherein the language analysis module comprises:

a POS tagging unit analyzing a morpheme of the sentence and tagging the POS of each respective word;

a syntax analysis unit classifying each respective word for each meaning and parsing the sentence into at least one phrase; and

a frame analysis unit classifying each respective phrase into a theme, a parameter, and an action, and generating at least one semantic frame corresponding to the sentence.

4. The apparatus of claim 3, wherein the dialogue management module comprises:

a context information unit restoring the sentence by converting a respective phrase of the at least one semantic frame into a valid phrase or a default value, with reference to a predetermined database or by analogizing a phrase determined to be omitted from the sentence; and

a user intention analysis unit generating at least one action list by analyzing an intention of the user according to the restored sentence and selecting a predetermined action by determining whether a respective action can be performed.

5. The apparatus of claim 4, wherein the database comprises:

a reference database maintaining a reference table in which one of the theme, the parameter, and the action is established as a reference domain and at least one valid phrase or default value corresponding to other domains in addition to the reference domain established with respect to a predetermined phrase is recorded; and

a focus stack in which a result of analyzing previous utterance contents in response to at least one user is recorded.

6. The apparatus of claim 4, wherein the dialogue management module further comprises a response control unit providing a dialog including inquiry content with respect to the phrase or the phrase determined to be omitted to the user when the context information unit cannot convert the phrase included in the sentence into the valid phrase or the default value, or cannot analogize the phrase determined to be omitted from the sentence.

7. The apparatus of claim 4, further comprising a context model database storing arguments corresponding to at least one phrase, and a sub-dialogue that is a combination of the arguments according to the previous utterance contents of the user in response to a respective user,

wherein the user intention analysis unit extracts the argument corresponding to the respective phrase included in the sentence with reference to the context model database and analyzes the intention of the user according to the sentence with reference to the sub-dialogue associated with the extracted argument.

8. The apparatus of claim 7, wherein the user intention analysis unit generates and records the sub-dialogue corresponding to the extracted argument in the context model database when the sub-dialogue associated with the extracted argument does not exist in the context model database.

9. The apparatus of claim 4, wherein the user intention analysis unit checks a present state of a device that will perform an action corresponding to the action list and determines whether the action can be performed.

10. A speech dialogue service method comprising:

tagging a POS of each respective word included in a sentence recorded in a predetermined text;

syntactically analyzing the sentence by-classifying a meaning of each respective word;

generating at least one semantic frame corresponding to the sentence according to a result of the syntactical analysis;

analyzing an intention of the sentence corresponding to each respective semantic frame;

selecting a predetermined sentence intention according to whether an action corresponding to the intention of the sentence can be performed; and

generating a system response corresponding to the sentence intention.

11. A computer-readable recording medium in which a program for executing a speech dialogue service method is recorded, the method comprising:

tagging a POS of each respective word included in a sentence recorded in a predetermined text;

syntactically analyzing the sentence by classifying a meaning of each respective word;

generating at least one semantic frame corresponding to the sentence according to a result of the syntactical analysis;

analyzing an intention of the sentence corresponding to each respective semantic frame;

selecting a predetermined sentence intention according to whether an action corresponding to the intention of the sentence can be performed; and

generating a system response corresponding to the sentence intention.

12. A method of providing a speech dialogue service, comprising:

recognizing and converting uttered speech into text;

resolving an intention of the speech by analyzing a sentence in the text by tagging a part of speech (POS) of each word of the sentence, parsing the sentence into at least one phrase by classifying a meaning of each word and combining each word whose meaning is classified, generating at least one semantic frame corresponding to the sentence according to the meaning of each word in the sentence, determining an intention of the sentence corresponding to the at least one semantic frame, and generating a system response corresponding to the sentence intention by selecting a predetermined sentence intention according to whether an action corresponding to the determined intention can be executed;

analyzing the intention and performing plan management with respect to the execution of the user'"'"'s intended command according to the intention analysis, when analysis of the intention is completed; and

inquiring about the intention of the speech when the intention is not correctly analyzed or when the determined intention cannot be executed.

13. A computer-readable recording medium in which a program for executing a method of providing a speech dialogue service is recorded, the method comprising:

recognizing and converting uttered speech into text;

resolving an intention of the speech by analyzing a sentence in the text by tagging a part of speech (POS) of each word of the sentence, parsing the sentence into at least one phrase by classifying a meaning of each word and combining each word whose meaning is classified, generating at least one semantic frame corresponding to the sentence according to the meaning of each word in the sentence, determining an intention of the sentence corresponding to the at least one semantic frame, and generating a system response corresponding to the sentence intention by selecting a predetermined sentence intention according to whether an action corresponding to the determined intention can be executed;

analyzing the intention and performing plan management with respect to the execution of the user'"'"'s intended command according to the intention analysis, when analysis of the intention is completed; and

inquiring about the intention of the speech when the intention is not correctly analyzed or when the determined intention cannot be executed.