NATURAL LANGUAGE DIALOGUE METHOD AND NATURAL LANGUAGE DIALOGUE SYSTEM

• US 20140188478A1
• Filed: 12/31/2013
• Published: 07/03/2014
• Est. Priority Date: 12/31/2012
• Status: Active Grant

• Current Assignee: VIA Technologies Incorporated (VIA Technologies)
• Original Assignee: VIA Technologies Incorporated (VIA Technologies)
• Inventors: Zhang, Guo-Feng
• Granted Patent: US 9,767,796 B2
• Time in Patent Office: Days
• US Class Current: 704/257
• CPC Class Codes: G06F 16/903 Querying for retrieval from the web G06F16/953
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/088 Word spotting

Abstract

A natural language dialogue method and a natural language dialogue system are provided. In the method, a first speech input is received and parsed to generate at least one keyword included in the first speech input, so that a candidate list including at least one report answer is obtained. According to a properties database, one report answer is selected from the candidate list, and a first speech response is output according to the report answer. Other speech inputs are received, and a user'"'"'s preference data is captured from the speech inputs. The user'"'"'s preference data is stored in the properties database.

Claims

1. A natural language dialogue system comprising:

a properties database;

a speech sampling module receiving a first speech input; and

a natural language comprehension system coupled to the properties database and the speech sampling module, the natural language comprehension system parsing a first keyword included in the first speech input to obtain a candidate list, the candidate list having at least one report answer, the natural language comprehension system selecting the at least one report answer from the candidate list according to the properties database and outputting a first speech response according to the at least one report answer, the speech sampling module receiving a plurality of speech inputs other than the first speech input, the natural language comprehension system capturing a user'"'"'s preference data from the speech inputs other than the first speech inputs and storing the user'"'"'s preference data into the properties database.

2. The natural language dialogue system as recited in claim 1, wherein the natural language comprehension system arranges the at least one report answer in the candidate list according to a priority and selects the at least one report answer from the candidate list according to the priority.

3. The natural language dialogue system as recited in claim 2, wherein the natural language comprehension system further determines the priority of the at least one report answer according to the user'"'"'s preference data stored in the properties database.

4. The natural language dialogue system as recited in claim 2, wherein the properties database further stores a user'"'"'s habit, and the natural language comprehension system determines the priority of the at least one report answer according to the user'"'"'s habit.

5. The natural language dialogue system as recited in claim 2, wherein the properties database further stores a public usage habit, and the natural language comprehension system determines the priority of the at least one report answer according to the public usage habit.

6. The natural language dialogue system as recited in claim 1, further comprising:

a speech synthesis database coupled to the natural language comprehension system, wherein the natural language comprehension system searches a first speech corresponding to the first keyword of the first speech input from the speech synthesis database, and the searched first speech acts as the first speech response corresponding to the at least one report answer.

7. The natural language dialogue system as recited in claim 1, the natural language comprehension system parsing a second speech input to obtain a second keyword, wherein the second speech input is received by the speech sampling module before the first speech response is output, and the natural language comprehension system selects one of the at least one report answer matching the first keyword and the second keyword.

8. The natural language dialogue system as recited in claim 1, the natural language dialogue system being suitable for a mobile communication apparatus.

9. A natural language dialogue method comprising:

receiving a first speech input;

parsing a first keyword included in the first speech input to obtain a candidate list, wherein the candidate list has at least one report answer;

selecting the at least one report answer from the candidate list according to a properties database;

outputting a first speech response according to the at least one report answer;

receiving a plurality of speech inputs other than the first speech input;

capturing a user'"'"'s preference data from the speech inputs other than the first speech input; and

storing the user'"'"'s preference data into the properties database.

10. The natural language dialogue method as recited in claim 9, further comprising:

arranging the at least one report answer in the candidate list according to a priority; and

selecting one of the at least one report answer from the candidate list according to the priority.

11. The natural language dialogue method as recited in claim 10, further comprising:

determining the priority of the at least one report answer according to the user'"'"'s preference data stored in the properties database.

12. The natural language dialogue method as recited in claim 10, further comprising:

storing a user'"'"'s habit into the properties database; and

determining the priority of the at least one report answer according to the user'"'"'s habit.

13. The natural language dialogue method as recited in claim 10, further comprising:

storing a public usage habit into the properties database;

determining the priority of the at least one report answer according to the public usage habit.

14. The natural language dialogue method as recited in claim 10, wherein the step of selecting the at least one report answer comprises:

in a speech synthesis database, searching a first speech corresponding to the first keyword included in the first speech input; and

selecting one of the at least one report answer conforming to the first speech.

15. The natural language dialogue method as recited in claim 10, wherein the step of selecting the at least one report answer further comprises:

parsing a second speech input to obtain at least one second keyword, wherein the second speech input is input before the first speech input; and

selecting one of the at least one report answer matching the first keyword and the at least one second keyword.

16. The natural language dialogue method as recited in claim 9, the natural language dialogue method being suitable for a mobile communication apparatus.

17. A language dialogue system comprising:

a properties database for storing a user'"'"'s preference data; and

a language comprehension system parsing a first keyword included in a first speech input to obtain a candidate list, wherein the candidate list has at least one first report answer, the language comprehension system selects one of the at least one first report answer from the candidate list according to the user'"'"'s preference data stored in the properties database and outputs a first speech response according to the selected one of the at least one first report answer, wherein the language comprehension system obtains a first user'"'"'s preference data according to the first speech input and revises the user'"'"'s preference data stored into the properties database according to the first user'"'"'s preference data.

18. The language dialogue system as recited in claim 17, wherein the language comprehension system determines a priority of the at least one first report answer according to the user'"'"'s preference data and selects one of the at least one first report answer according to the priority.

19. The language dialogue system as recited in claim 18, wherein the user'"'"'s preference data stored into the properties database comprises a correlation between a specific keyword and a user'"'"'s personal preference, a user'"'"'s personal usage habit, a public preference, or a public usage habit.

20. The language dialogue system as recited in claim 17, wherein the properties database comprises at least one record, each of the at least one record comprises a keyword and a corresponding user'"'"'s personal preference record data or a corresponding user'"'"'s personal dislike record data.

21. The language dialogue system as recited in claim 20, wherein the step of selecting the one of the at least one first report answer from the candidate list according to the user'"'"'s preference data stored in the properties database comprises:

searching one of the at least one record corresponding to the first keyword from the at least one record stored in the properties database; and

selecting the one of the at least one first report answer according to the user'"'"'s personal preference record data included in the corresponding record, the user'"'"'s personal dislike record data included in the corresponding record, or a combination of the user'"'"'s personal preference record data and the user'"'"'s personal dislike record data included in the corresponding record.

22. The language dialogue system as recited in claim 20, wherein the at least one record included in the properties database further comprises a user'"'"'s personal downloaded record corresponding to the keyword, a user'"'"'s personal cited record corresponding to the keyword, a user'"'"'s personal recommendation record corresponding to the keyword, a user'"'"'s personal comment record corresponding to the keyword, or a user'"'"'s personal referral record corresponding to the keyword.

23. The language dialogue system as recited in claim 22, wherein the at least one first report answer is selected according to the user'"'"'s personal downloaded record data included in the at least one record, the user'"'"'s personal cited record data included in the at least one record, the user'"'"'s personal recommendation record data included in the at least one record, the user'"'"'s personal comment record data included in the at least one record, the user'"'"'s personal referral record data included in the at least one record, or at least a combination of two of said user'"'"'s personal record data.

24. The language dialogue system as recited in claim 20, wherein the step of revising the user'"'"'s preference data stored into the properties database according to the first user'"'"'s preference data comprises:

searching one of the at least one record corresponding to the first keyword from the at least one record stored in the properties database; and

in the corresponding record of the first keyword, revising the user'"'"'s personal preference record data or the user'"'"'s personal dislike record data included in the corresponding record according to a preference indication corresponding to the first keyword included in the first user'"'"'s preference data.

25. The language dialogue system as recited in claim 20, further comprising if the one of the at least one record corresponding to the first keyword is not searched from the properties database, a new record is established in the properties database according to the first keyword, and the user'"'"'s personal preference record data corresponding to the new record and the user'"'"'s personal dislike record data corresponding to the new record are revised according to the user'"'"'s preference indication corresponding to the first keyword.

26. The language dialogue system as recited in claim 24, wherein when the at least one record stored in the properties database corresponds to the first keyword, it means that the keyword included in the at least one record stored in the properties database corresponds to the first keyword.

27. The language dialogue system as recited in claim 17, wherein the properties database comprises at least one record, each of the at least one record comprises a keyword and a corresponding public preference record data or a corresponding public dislike record data.

28. The language dialogue system as recited in claim 27, wherein the step of selecting the one of the at least one first report answer from the candidate list according to the user'"'"'s preference data stored in the properties database comprises:

searching one of the at least one record corresponding to the first keyword according to the keyword of the at least one record stored in the properties database;

selecting the one of the at least one first report answer according to the public preference record data included in the corresponding record, the public dislike record data included in the corresponding record, or a combination of the public preference record data and the public dislike record data included in the corresponding record.

29. The language dialogue system as recited in claim 27, wherein the at least one record included in the properties database further comprises a public downloaded record data corresponding to the keyword, a public cited record data corresponding to the keyword, a public recommendation record data corresponding to the keyword, a public comment record data corresponding to the keyword, or a public referral record data corresponding to the keyword.

30. The language dialogue system as recited in claim 29, wherein the at least one first report answer is selected according to the public downloaded record data included in the at least one record, the public cited record data included in the at least one record, the public recommendation record data included in the at least one record, the public comment record data included in the at least one record, the public referral record data included in the at least one record, or at least a combination of two of said public record data.

31. The language dialogue system as recited in claim 27, wherein the step of revising the user'"'"'s preference data stored into the properties database according to the first user'"'"'s preference data comprises:

searching one of the at least one record corresponding to the first keyword from the at least one record stored in the properties database; and

in the record corresponding to the first keyword, revising the public preference record data or the public dislike record data included in the corresponding record according to a user'"'"'s preference indication corresponding to the first keyword included in the first user'"'"'s preference data.

32. The language dialogue system as recited in claim 31, wherein if the one of the at least one record corresponding to the first keyword is not searched from the properties database, a new record is established in the properties database according to the first keyword, and the public preference record data corresponding to the new record and the public dislike record data corresponding to the new record are revised according to the user'"'"'s preference indication corresponding to the first keyword.

33. The language dialogue system as recited in claim 17, further comprising:

parsing a second speech input to obtain at least one second keyword, wherein the second speech input is input by a user after the first speech response is output; and

selecting a second report answer other than the at least one first report answer from the candidate list according to the second keyword.

34. The language dialogue system as recited in claim 33, further comprising the language comprehension system captures a second user'"'"'s preference data from the second speech input and revises the user'"'"'s preference data stored into the properties database according to the second user'"'"'s preference data.

35. The language dialogue system as recited in claim 17, wherein the step of revising the user'"'"'s preference data stored into the properties database according to the first user'"'"'s preference data comprises:

searching one of the at least one record corresponding to the first keyword from the at least one record stored in the properties database; and

in the record corresponding to the first keyword, revising the stored user'"'"'s preference data corresponding to the first keyword according to a user'"'"'s preference indication corresponding to the first keyword included in the first user'"'"'s preference data.

36. The language dialogue system as recited in claim 35, wherein if the one of the at least one record corresponding to the first keyword is not searched from the properties database, a new record is established in the properties database according to the first keyword, and the user'"'"'s preference record corresponding to the new record is revised according to the user'"'"'s preference indication corresponding to the first keyword.

37. A language dialogue method comprising:

receiving a first speech input;

parsing a first keyword included in the first speech input to obtain a candidate list, wherein the candidate list has at least one first report answer;

selecting the at least one first report answer from the candidate list according to a user'"'"'s preference data stored in a properties database;

outputting a first speech response according to the at least one first report answer;

obtaining a first user'"'"'s preference data according to the first speech response; and

revising the user'"'"'s preference data stored in the properties database according to the first user'"'"'s preference data.

38. The language dialogue method as recited in claim 37, further comprising arranging the at least one first report answer according to a priority determining by the user'"'"'s preference data and selecting one of the at least one first report answer according to the priority.

39. The language dialogue method as recited in claim 37, wherein the user'"'"'s preference data comprises a correlation between a specific keyword and a user'"'"'s personal preference, a user'"'"'s personal usage habit, a public preference, or a public usage habit.

40. The language dialogue method as recited in claim 37, wherein the properties database comprises at least one record, each of the at least one record comprises a keyword and a corresponding user'"'"'s personal preference record data or a corresponding user'"'"'s personal dislike record data.

41. The language dialogue method as recited in claim 40, wherein the step of selecting the one of the at least one first report answer from the candidate list according to the user'"'"'s preference data stored in the properties database comprises:

searching one of the at least one record corresponding to the first keyword from the at least one record stored in the properties database; and

selecting the one of the at least one first report answer according to the user'"'"'s personal preference record included in the corresponding record data, the user'"'"'s personal dislike record data included in the corresponding record, or a combination of the user'"'"'s personal preference record data and the user'"'"'s personal dislike record data included in the corresponding record.

42. The language dialogue method as recited in claim 40, wherein the at least one record included in the properties database further comprises a user'"'"'s personal downloaded record data corresponding to the keyword, a user'"'"'s personal cited record data corresponding to the keyword, a user'"'"'s personal recommendation record data corresponding to the keyword, a user'"'"'s personal comment record data corresponding to the keyword, or a user'"'"'s personal referral record data corresponding to the keyword.

43. The language dialogue method as recited in claim 42, wherein the at least one first report answer is selected according to the user'"'"'s personal downloaded record data included in the at least one record, the user'"'"'s personal cited record data included in the at least one record, the user'"'"'s personal recommendation record data included in the at least one record, the user'"'"'s personal comment record data included in the at least one record, the user'"'"'s personal referral record data included in the at least one record, or at least a combination of two of said user'"'"'s personal record data.

44. The language dialogue method as recited in claim 40, wherein the step of revising the user'"'"'s preference data stored into the properties database according to the first user'"'"'s preference data comprises:

searching one of the at least one record corresponding to the first keyword from the at least one record stored in the properties database; and

in the record corresponding to the first keyword, revising the user'"'"'s personal preference data record or the user'"'"'s personal dislike record data included in the corresponding record according to a user'"'"'s preference indication corresponding to the first keyword included in the first user'"'"'s preference data.

45. The language dialogue method as recited in claim 44, further comprising if the one of the at least one record corresponding to the first keyword is not searched from the properties database, a new record is established in the properties database according to the first keyword, and the user'"'"'s personal preference record data corresponding to the new record and the user'"'"'s personal dislike record data corresponding to the new record are revised according to the user'"'"'s preference indication corresponding to the first keyword.

46. The language dialogue method as recited in claim 44, wherein when the at least one record stored in the properties database corresponds to the first keyword, it means that the keyword included in the at least one record stored in the properties database corresponds to the first keyword.

47. The language dialogue method as recited in claim 37, wherein the properties database comprises at least one record, each of the at least one record comprises a keyword and a corresponding public preference record data or a corresponding public dislike record data.

48. The language dialogue method as recited in claim 47, wherein the step of selecting the one of the at least one first report answer from the candidate list according to the user'"'"'s preference data stored in the properties database comprises:

searching one of the at least one record corresponding to the first keyword according to the keyword of the at least one record stored in the properties database;

selecting the one of the at least one first report answer according to the public preference record data included in the corresponding record data, the public dislike record data included in the corresponding record data, or a combination of the public preference record data and the public dislike record data included in the corresponding record.

49. The language dialogue method as recited in claim 47, wherein the at least one record included in the properties database further comprises a public downloaded record data corresponding to the keyword, a public cited record data corresponding to the keyword, a public recommendation record data corresponding to the keyword, a public comment record data corresponding to the keyword, or a public referral record data corresponding to the keyword.

50. The language dialogue method as recited in claim 49, wherein the at least one first report answer is selected according to the public downloaded record data included in the at least one record, the public cited record data included in the at least one record, the public recommendation record data included in the at least one record, the public comment record data included in the at least one record, the public referral record data included in the at least one record, or at least a combination of two of said public record data.

51. The language dialogue method as recited in claim 47, wherein the step of revising the user'"'"'s preference data stored into the properties database according to the first user'"'"'s preference data comprises:

searching one of the at least one record corresponding to the first keyword from the at least one record stored in the properties database; and

in the record corresponding to the first keyword, revising the public preference record data or the public dislike record data included in the corresponding record according to a user'"'"'s preference indication corresponding to the first keyword included in the first user'"'"'s preference data.

52. The language dialogue method as recited in claim 51, wherein if the one of the at least one record corresponding to the first keyword is not searched from the properties database, a new record is established in the properties database according to the first keyword, and the public preference record data corresponding to the new record and the public dislike record data corresponding to the new record are revised according to the user'"'"'s preference indication corresponding to the first keyword.

53. The language dialogue method as recited in claim 37, further comprising:

parsing a second speech input to obtain at least one second keyword, wherein the second speech input is input by a user after the first speech response is output; and

selecting a second report answer other than the at least one first report answer from the candidate list according to the second keyword.

54. The language dialogue method as recited in claim 53, further comprising the language comprehension system captures a second user'"'"'s preference data from the second speech input and revises the user'"'"'s preference data stored into the properties database according to the second user'"'"'s preference data.

55. The language dialogue method as recited in claim 37, wherein the step of revising the user'"'"'s preference data stored into the properties database according to the first user'"'"'s preference data comprises:

searching one of the at least one record corresponding to the first keyword from the at least one record stored in the properties database; and

in the record corresponding to the first keyword, revising the stored user'"'"'s preference data corresponding to the first keyword according to a user'"'"'s preference indication corresponding to the first keyword included in the first user'"'"'s preference data.

56. The language dialogue method as recited in claim 55, further comprising if the one of the at least one record corresponding to the first keyword is not searched from the properties database, a new record is established in the properties database according to the first keyword, and the user'"'"'s preference record data corresponding to the new record is revised according to the user'"'"'s preference indication corresponding to the first keyword.

57. A properties database for recording a user'"'"'s usage habit, wherein when at least one user inputs request information to a language dialogue system, the properties database stores a user'"'"'s preference data included in the request information input by the at least one user, and each of at least one record of the properties database comprises:

a keyword field for storing at least one keyword; and

a user'"'"'s personal preference field for storing a preference indication corresponding to the at least one keyword,

wherein when the at least one user inputs the request information, the properties database compares a keyword included in the request information with the at least one keyword stored in the properties database, and if the keyword included in the request information matches one of the at least one keyword stored in the properties database, the preference indication corresponding to the one of the at least one keyword stored in the properties database is revised according to a preference indication included in the request information.

58. The properties database as recited in claim 57, wherein when the at least one user inputs the request information, the properties database compares the keyword included in the request information with the at least one keyword stored in the properties database and outputs the preference indication corresponding to the record matching the keyword included in the request information.

59. The properties database as recited in claim 57, wherein after the properties database compares the keyword included in the request information with the at least one keyword stored in the properties database and observes that the keyword included in the request information does not match any of the at least one keyword stored in the properties database, the properties database establishes a new record according to the keyword included in the request information, the keyword of the new record is the keyword included in the request information, and the user'"'"'s personal preference field of the new record stores data according to the preference indication included in the request information.

60. The properties database as recited in claim 57, wherein the at least one record stored in the properties database further comprises a public preference field for storing the preference indication corresponding to the at least one keyword of the at least one record, the public preference fields stores preference indications of a plurality of the at least one user, and the preference indications correspond to the at least one keyword of the at least one record.

61. The properties database as recited in claim 57, wherein the at least one record stored in the properties database further comprises a user'"'"'s personal dislike field, and when the preference indication included in the request information input by the at least one user indicates the at least one user dislikes the keyword included in the request information, the user'"'"'s personal dislike field is revised.

62. The properties database as recited in claim 61, wherein the at least one record stored in the properties database further comprises a public dislike field, and when the preference indication included in the request information input by the at least one user indicates the at least one user dislikes the keyword included in the request information, the public dislike field is revised, wherein the public dislike field stores dislike information from a plurality of the at least one user, and the dislike information corresponds to the at least one keyword of the at least one record.

63. The properties database as recited in claim 57, wherein the at least one record stored in the properties database further comprises one of following fields corresponding to the at least one keyword corresponding to the at least one record:

a personal downloaded record field;

a personal cited record field;

a personal recommendation record field;

a personal comment record field; and

a personal referral record field,

wherein if the keyword included in the request information matches the at least one keyword stored in the properties database, the properties database outputs data stored in the personal downloaded record field, the personal cited record field, the personal recommendation record field, the personal comment record field, and the personal referral record field corresponding to the at least one record.

64. The properties database as recited in claim 63, further comprising if the keyword included in the request information matches the at least one keyword stored in the properties database, the properties database further outputs the data stored in at least two of the personal downloaded record field, the personal cited record field, the personal recommendation record field, the personal comment record field, and the personal referral record field corresponding to the at least one record.

65. The properties database as recited in claim 63, wherein if the keyword included in the request information matches the at least one keyword stored in the properties database, the properties database revises the data stored in the personal downloaded record field, the personal cited record field, the personal recommendation record field, the personal comment record field, and the personal referral record field corresponding to the at least one record according to the preference indication included in the request information.

66. The properties database as recited in claim 57, wherein the at least one record stored in the properties database further comprises one of following fields corresponding to the at least one keyword of the at least one record:

a public downloaded record field;

a public cited record field;

a public recommendation record field;

a public comment record field; and

a public referral record field,

wherein if the keyword included in the request information matches the at least one keyword stored in the properties database, the properties database outputs data stored in the public downloaded record field, the public cited record field, the public recommendation record field, the public comment record field, and the public referral record field corresponding to the at least one record.

67. The properties database as recited in claim 66, wherein if the keyword included in the request information matches the at least one keyword stored in the properties database, the properties database further outputs the data stored in at least two of the public downloaded record field, the public cited record field, the public recommendation record field, the public comment record field, and the public referral record field corresponding to the at least one record.

68. The properties database as recited in claim 66, further comprising if the keyword included in the request information matches the at least one keyword stored in the properties database, the properties database revises the data stored in the public downloaded record field, the public cited record field, the public recommendation record field, the public comment record field, and the public referral record field corresponding to the at least one record according to the preference indication included in the request information.

69. A method of providing a corresponding report answer corresponding to user'"'"'s speech inputs in response to a user'"'"'s habit, the method being suitable for a mobile communication apparatus and comprising:

receiving a speech input from a user, the speech input comprising a keyword and a preference indication; and

outputting a speech response in response to a user'"'"'s preference data stored by the user, the user'"'"'s preference corresponding to the keyword, wherein the speech response is output according to a most popular user'"'"'s preference of the user'"'"'s preference,

wherein the user'"'"'s preference data is revised according to the preference indication included in the speech input.

70. The method as recited in claim 69, wherein the user'"'"'s preference data is a record indicating whether the user personally likes or dislikes the keyword.

71. The method as recited in claim 69, wherein the user'"'"'s preference is a record indicating whether the public likes or dislikes the keyword.

72. The method as recited in claim 69, the user'"'"'s preference data being stored into a record including a corresponding keyword, when the keyword included in the speech input does not match the corresponding keyword of the record, establishing a new record and storing the preference indication included in the speech input into the new record, wherein the keyword included in the speech input acts as the corresponding keyword of the new record, and the preference indication stored in the new record acts as a new user'"'"'s preference.