Spoken dialogue interface apparatus and method

• US 7,725,322 B2
• Filed: 02/07/2006
• Issued: 05/25/2010
• Est. Priority Date: 02/15/2005
• Status: Expired due to Fees

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Lee, Jae-won, Kwak, Byung-kwan
• Primary Examiner(s): Dorvil; Richemond
• Assistant Examiner(s): He; Jialong
• Application Number:
  US11/348,301
• Publication Number:
  US 20060184370A1
• Time in Patent Office: 1,568 Days
• Field of Search: 704/275, 704/270, 704/9, 704/257, 704/258
• US Class Current: 704/275
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding

Abstract

The spoken dialogue interface apparatus according to an embodiment of the present invention includes a speech recognition module for recognizing a human'"'"'s speech from a sound signal; a user intention interpretation module for extracting a sentence from the recognized speech and interpreting a user'"'"'s intention based on the sentence; a user intention selection module for determining user intention using the interpreted user'"'"'s intention and a predetermined domain action frame; and a system response generation module for generating a system response sentence corresponding to the selected user intention, wherein the domain action frame includes service information which the user requests, and parameter information which is used to perform a service, and the domain action frame is constructed to have a hierarchical tree structure.

Claims

1. A spoken dialogue interface apparatus including a processor having computing device-executable instructions, the apparatus comprising:

a speech recognition module, controlled by the processor, for recognizing a human'"'"'s speech from a sound signal;

a user intention interpretation module, controlled by the processor, for extracting a sentence from the recognized speech and interpreting a user'"'"'s intention based on the sentence;

a user intention selection module, controlled by the processor, for determining user intention using the interpreted user'"'"'s intention and selecting a first predetermined domain action frame from a plurality of domain action frames stored in a domain action frame module; and

a system response generation module, controlled by the processor, for generation a system response sentence corresponding to the determined user intention and the selected first domain action frame;

wherein the first domain action frame includes service information which the user requests, and parameter information which is used to perform a service, and the first domain action frame is one of the plurality of domain action frames represented as a node in a hierarchical tree structure, and

wherein, when a corresponding system response sentence does not exist in the first domain action frame due to missing of some parameter information with respect to the first domain action frame, the system response generation module generates a response using a back-off response model storage module, the system response sentence being generated with reference to a second domain action frame which exists in a service provided by a predetermined domain and is represented as a node in an upper layer of the hierarchical tree structure.

2. The spoken dialogue interface apparatus as set forth in claim 1, wherein the system response sentences are set in the first or second domain action frame.

3. The spoken dialogue interface apparatus as set forth in claim 1, further comprising a speech synthesis module for transforming the system response sentence generated by the system response generation module into speech and outputting it to the user.

4. The spoken dialogue interface apparatus as set forth in claim 1, further comprising a system intention selection module for performing a service based on the user intention selected by the user intention selection module and transferring performance result;

wherein the system response generation module generates a system response sentence corresponding to the transferred results.

5. The spoken dialogue interface apparatus as set forth in claim 4, wherein the system response sentence is set in the first or second domain action frame.

6. The spoken dialogue interface apparatus as set forth in claim 4, further comprising a speech synthesis module for transforming the system response sentence generated by the system response generation module into speech and outputting it to the user.

7. A spoken dialogue interface method implemented by a spoken dialogue interface apparatus comprising a processor having computing device-executable instructions, the method comprising the steps of:

recognizing a human'"'"'s speech from a sound signal in a speech recognition module;

extracting a sentence from the recognized speech and interpreting the user'"'"'s intention based on the sentence;

determining the user intention using the interpreted user'"'"'s intention and selecting a first predetermined domain action frame from a plurality of predetermined domain action frames in a user intention selection module;

generating a system response sentence corresponding to the determined user intention and the selected first domain action frame in a system response generation module,

wherein the first domain action frame includes service information which the user requests, and parameter information which is used to perform a service, and the first domain action frame is one of the plurality of domain action frames represented as nodes in a hierarchical tree structure; and

generating, by way of the processor, the system response sentence with reference to a second domain action frame when a corresponding system response sentence does not exist in the first domain action frame due to missing of some parameter information with respect to the first domain action frame, the second domain action frame being a service provided by a predetermined domain represented as a node in the hierarchical tree structure and existing in an upper layer of the hierarchical tree structure in a back-off response model storage module.

8. The spoken dialogue interface method as set forth in claim 7, wherein the system response sentence is set in the first or second domain action frame.

9. The spoken dialogue interface method as set forth in claim 7, further comprising the step of transforming the system response sentence generated by the system response generation module into speech and outputting it to the user.

10. The spoken dialogue interface method as set forth in claim 7, wherein the generating step comprises the steps of;

performing a service based on the user intention selected by the user intention selection module and transferring performance results; and

generating a system response sentence corresponding to the transferred results.

11. The spoken dialogue interface method as set forth in claim 10, wherein the system response sentence is set in the first or second domain action frame.

12. The spoken dialogue interface method as set forth in claim 10, further comprising the step of transforming the system response sentence generated by the

system response generation module into speech and outputting it to the user.