Dialogue system using extended domain and natural language recognition method and computer-readable medium thereof

• US 9,495,958 B2
• Filed: 12/04/2013
• Issued: 11/15/2016
• Est. Priority Date: 01/04/2010
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Kim, Hong Won, Han, Woo Sup
• Primary Examiner(s): MCFADDEN, SUSAN IRIS
• Application Number:
  US14/096,628
• Publication Number:
  US 20140095152A1
• Time in Patent Office: 1,077 Days
• Field of Search: 704/9
• US Class Current: 1/1
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding

Abstract

A dialog system uses an extended domain in order to have a dialog with a user using natural language. If a dialog pattern actually input by the user is different from a dialog pattern predicted by an expert, an extended domain generated in real time based on user input is used and an extended domain generated in advance is used to have a dialog with the user.

Claims

1. A method of recognizing natural language in a dialogue system, the method comprising:

by at least one hardware processor interfaced with a memory including instructions which control the at least one hardware processor to;

input voice data of a user;

parse the voice data of the user to generate a parsed input set;

determine whether the parsed input set is included in an initial domain including an initial input set generated before dialogue with the user and in an extended domain including an extended input set generated upon dialogue with the user;

provide an understood result comprising content based on at least one of the initial input set and the extended input set according to a result of the determination; and

output the understood result.

2. The method according to claim 1, wherein when the parsed input set is not included in the initial domain, the extended domain is used.

3. The method according to claim 1, wherein the extended domain is generated by generating the extended input set using associated language data of a certain word of the parsed input set.

4. The method according to claim 3, wherein:

the associated language data includes at least one of a synonym, a hyponym and a hypernym associated with the certain word, and the extended input set is generated by changing at least one word of the parsed input set to the associated language data.

5. The method according to claim 4, wherein the instructions further control the at least one hardware processor to apply a weight to the extended input set in order to restrict a size of the generated extended domain, and to stop the generation of the extended domain if the weight applied to the extended input set is greater than a predefined value.

6. The method according to claim 5, wherein the weight is differently applied depending on which word of the parsed input set is changed, or depending on a number of words of the parsed input set are changed.

7. The method according to claim 6, wherein a weight applied to the extended input set generated by changing a certain word of the parsed input set to a hyponym is greater than a weight applied to the extended input set generated by changing the certain word to a synonym, and a weight applied to the extended input set generated by changing the certain word to a hypernym is greater than the weight applied to the extended input set generated by changing the certain word to the hyponym.

17. At least one non-transitory computer readable medium comprising computer readable instructions that control at least one processor to implement the method of claim 1.

8. A method of recognizing natural language in a dialogue system, the method comprising:

by at least one hardware processor interfaced with a memory including instructions which control the at least one hardware processor to;

input voice data of a user;

parse the voice data of the user to generate a parsed input set;

generate an extended domain including an extended input set generated using an initial input set generated by predicting meanings necessary for a use environment and expressions corresponding thereto and associated language data of the initial input set before dialogue with a user;

determine whether the parsed input set is included in an initial domain including the initial input set and in the extended domain;

provide an understood result comprising content based on at least one of the initial input set and the extended input set according to a result of the determination; and

output the understood result.

18. At least one non-transitory computer readable medium comprising computer readable instructions that control at least one processor to implement the method of claim 8.

9. A dialogue system using an extended domain, the dialogue system comprising:

at least one hardware processor interfaced with a memory including instructions which control the at least one hardware processor to;

parse voice data of a user to generate a parsed input set;

determine whether the parsed input set is included in an initial domain including an initial input set generated before dialogue with the user and in an extended domain including an extended input set generated upon dialogue with the user;

provide an understood result comprising content based on at least one of the initial input set and the extended input set according to a result of the determination; and

output the understood result.

10. The dialogue system according to claim 9,

wherein, if the parsed input set is not included in the initial domain, the at least one processor makes a request for generation of the extended domain.

11. The dialogue system according to claim 10, wherein the instructions further control the at least one hardware processor to:

maintain an associated word bank having associated language data of a certain word configuring the parsed input set;

generate the extended input set using the associated language data; and

make a request for generation of the extended input set to the input set generator according to the request for the generation of the extended domain.

12. The dialogue system according to claim 11, wherein the associated language data includes at least one of a synonym, a hyponym and a hypernym associated with the certain word, and

at least one word of the parsed input set is changed to the associated language data.

13. The dialogue system according to claim 11, wherein the instructions further control the at least one hardware processor to:

restrict a size of the extended domain, and

apply a weight to the extended input set and makes a request for stoppage of the generation of the extended input set if the weight applied to the extended input set is greater than a predefined value.

14. The dialogue system according to claim 13, wherein the weight is differently applied depending on which word of the parsed input set is changed, or depending on a number of words of the parsed input set are changed.

15. The dialogue system according to claim 13, wherein the instructions further control the at least one hardware processor to apply to the extended input set generated by changing a certain word of the parsed input set to a hyponym a weight greater than a weight applied to the extended input set generated by changing the certain word to a synonym, and applies to an extended input set generated by changing the certain word to a hypernym a weight greater than the weight applied to the extended input set generated by changing the certain word to the hyponym.

16. A dialogue system using an extended domain, the dialogue system comprising:

at least one hardware processor interfaced with a memory including instructions which control the at least one hardware processor to;

parse voice data of a user to generate a parsed input set;

generate an extended domain including an extended input set generated using an initial input set generated by predicting meanings necessary for a use environment and expressions corresponding thereto and associated language data of the initial input set before dialogue with a user;

determine whether the parsed input set is included in an initial domain including the initial input set and in the extended domain;

provide an understood result comprising content based on at least one of the initial input set and the extended input set according to a result of the determination;

and

output the understood result.

19. A dialogue system using an extended domain, the dialogue system comprising:

at least one hardware processor interfaced with a memory including instructions which control the at least one hardware processor to;

parse voice data of a user to generate a parsed input set;

determine whether the parsed input set is included in an initial domain including an initial input set generated before dialogue with the user and in an extended domain including an extended input set generated upon dialogue with the user;

provide an understood result comprising content based on at least one of the initial input set and the extended input set according to a result of the determination;

generate a response sentence based on the understood result; and

generate user language according to the response sentence by using the initial domain and the extended domain.