Natural language dialogue apparatus and method

• US 6,466,899 B1
• Filed: 03/14/2000
• Issued: 10/15/2002
• Est. Priority Date: 03/15/1999
• Status: Expired due to Term

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Sasajima, Munehiko, Yano, Takehide, Kono, Yasuyuki
• Primary Examiner(s): EDOUARD, PATRICK NESTOR
• Application Number:
  US09/525,052
• Time in Patent Office: 945 Days
• Field of Search: 704/1, 704/9, 704/10, 704/257, 704/270, 704/275, 707/531, 707/532, 707/533, 706/47, 706/49, 706/53, 706/54, 706/55
• US Class Current: 704/1
• CPC Class Codes: G06F 40/279 Recognition of textual entities
• CPC Class Codes: G06F 40/56 Natural language generation
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A natural language dialogue apparatus. A blank data memory stores blank data to blank a question element being a decision object for a value of truth or falsehood in a question to ask the value of truth or falsehood. If the input sentence is the question, a question data conversion section estimates the question element in the input sentence, and converts the input sentence to question data in which the question element is blanked by referring to the blank data memory. A problem solution section retrieves a solution for the question element blanked in the question data and generates a response in which an answer for the value of truth or falsehood is not included according to the solution.

Claims

1. A natural language dialogue apparatus, comprising:

input analysis means for analyzing an input sentence of natural language from a user;

blank data memory means for storing a plurality of blank data each including a retrieval key element and a blank object element for a question to ask a value of truth or falsehood, the blank object element representing a kind of object to be recognized;

question data conversion means for deciding whether the input sentence is a question to ask the value of truth or falsehood, for estimating an object to be recognized in the input sentence if the input sentence is the question to ask the value of truth or falsehood, for retrieving the blank data in which the retrieval key element is included in the input sentence from said blank data memory means, and for converting the input sentence to question data in which the object is blanked by referring to the blank object element of the retrieved blank data; and

problem solution means for retrieving a solution for the object blanked in the question data and for generating a response in which the solution is included and an answer for the value of truth or falsehood is not included.

2. The natural language dialogue apparatus according to claim 1,

wherein said blank data memory means stores a plurality of the blank data each including a question intention element and the blank object element corresponding to the retrieval key element, the retrieval key element representing a slot of arbitrary kind in the input sentence, the question intention element representing a kind of data necessary for the user in relation to the blank object element.

3. The natural language dialogue apparatus according to claim 2,

wherein said blank data memory means stores a plurality of the blank data corresponding to the same retrieval key element in order of priority, each blank data including the question intention element and the blank object element.

4. The natural language dialogue apparatus according to claim 3,

wherein said question data conversion means divides the input sentence into a plurality of slots of each kind if the input sentence is the question to ask the value of truth or falsehood.

5. The natural language dialogue apparatus according to claim 4,

wherein said question data conversion means retrieves the blank data corresponding to the retrieval key element that coincides with one slot in the input sentence from said blank data memory means, blanks another slot in the input sentence of which the kind is the same as the blank object element of the retrieved blank data, and converts a question intention to ask a value of truth or falsehood in the input sentence to the question intention element of the retrieved blank data.

6. The natural language dialogue apparatus according to claim 5,

wherein said question data conversion means retrieves the blank data in order of the priority if a plurality of the blank data corresponding to the coincided retrieval key element are stored in said blank data memory means.

7. The natural language dialogue apparatus according to claim 2,

wherein the retrieval key element is a verb as a movement element in the input sentence.

8. The natural language dialogue apparatus according to claim 7,

wherein said problem solution means previously stores a plurality of set course data from a start point to an arrival point along a driving route, each set course data including a plurality of slots of each kind necessary for actual driving on the driving route, one of the plurality of slots being the movement element.

9. The natural language dialogue apparatus according to claim 8,

wherein said problem solution means retrieves the set course data of which the movement element and at least one slot are equally included in the question data, and extracts another slot that is the same kind as the blanked slot included in the question data from the retrieved set course data as the solution.

10. The natural language dialogue apparatus according to claim 9,

wherein said question data conversion means blanks a slot of another kind in the question data if said problem solution means can not retrieve the set course data, and

wherein said problem solution means retrieves the set course data of which the movement element and at least one slot are equally included in the question data in which the slot of another kind is blanked.

11. The natural language dialogue apparatus according to claim 9,

wherein said problem solution means generates the response including the solution without the value of truth or falsehood by referring to the question intention element in the question data if said problem solution means can retrieve the set course data, and

wherein said problem solution means generates the response representing the value of falsehood in the input sentence if said problem solution means can not retrieve the set course data.

12. A method for communicating by natural language, comprising the steps of:

analyzing an input sentence of natural language from a user;

storing a plurality of blank data each including a retrieval key element and a blank object element for a question to ask a value of truth or falsehood, the blank object element representing a kind of object to be recognized;

deciding whether the input sentence is a question to ask the value of truth or falsehood;

estimating an object to be recognized in the input sentence if the input sentence is the question to ask the value of truth or falsehood;

retrieving the blank data of which the retrieval key element is included in the input sentence from the stored blank data;

converting the input sentence to question data in which the object is blanked by referring to the blank object element of the retrieved blank data;

retrieving a solution for the object blanked in the question data; and

generating a response in which the solution is included and an answer for the value of truth or falsehood is not included.

13. The method according to claim 12,

further comprising the step of:

storing a plurality of the blank data each including a question intention element and the blank object element corresponding to the retrieval key element, the retrieval key element representing a slot of arbitrary kind in the input sentence, the question intention element representing a kind of data necessary for the user in relation to the blank object element.

14. The method according to claim 13,

further comprising the step of:

storing a plurality of the blank data corresponding to the same retrieval key element in order of priority, each blank data including the question intention element and the blank object element.

15. The method according to claim 14,

further comprising the step of:

dividing the input sentence into a plurality of slots of each kind if the input sentence is the question to ask the value of truth or falsehood.

16. The method according to claim 15,

further comprising the steps of:

retrieving the blank data corresponding to the retrieval key element which coincides with one slot in the input sentence from the stored blank data;

blanking another slot in the input sentence that is the same kind as the blank object element of the retrieved blank data; and

converting a question intention to ask a value of truth or falsehood in the input sentence to the question intention element of the retrieved blank data.

17. The method according to claim 16,

further comprising the step of:

retrieving the blank data in order of the priority if a plurality of the blank data corresponding to the coincided retrieval key element are stored.

18. The method according to claim 13,

wherein the retrieval key element is a verb as a movement element in the input sentence.

19. The method according to claim 18,

further comprising the step of:

previously storing a plurality of set course data from a start point to an arrival point along a driving route, each set course data including a plurality of slots of each kind necessary for actual driving on the driving route, one of the plurality of slots being the movement element.

20. The method according to claim 19,

further comprising the steps of:

retrieving the set course data of which the movement element and at least one slot are equally included in the question data; and

extracting another slot that is the same kind as the blanked slot included in the question data from the retrieved set course data as the solution.

21. The method according to claim 20,

further comprising the steps of:

blanking a slot of another kind in the question data if the set course data is not retrieved; and

retrieving the set course data of which the movement element and at least one slot are equally included in the question data in which the slot of another kind is blanked.

22. The method according to claim 20,

further comprising the steps of:

generating the response including the solution without the value of truth or falsehood by referring to the question intention element in the question data if the set course data is retrieved; and

generating the response representing the value of falsehood in the input sentence if the set course data is not retrieved.

23. A computer readable memory containing computer readable instructions, comprising:

instruction means for causing a computer to analyze an input sentence of natural language from a user;

instruction means for causing a computer to store a plurality of blank data each including a retrieval key element and a blank object element for a question to ask a value of truth or falsehood, the blank object element representing a kind of object to be recognized;

instruction means for causing a computer to decide whether the input sentence is a question to ask the value of truth or falsehood;

instruction means for causing a computer to estimate an object to be recognized in the input sentence if the input sentence is the question to ask the value of truth or falsehood;

instruction means for causing a computer to retrieve the blank data of which the retrieval key element is included in the input sentence from the stored blank data;

instruction means for causing a computer to convert the input sentence to question data in which the object is blanked by referring to the blank object element of the retrieved blank data;

instruction means for causing a computer to retrieve a solution for the object blanked in the question data; and

instruction means for causing a computer to generate a response in which the solution is included and an answer for the value of truth or falsehood is not included.

24. A natural language dialogue apparatus, comprising:

an input analysis unit configured to analyze an input sentence of natural language from a user;

a blank data memory configured to store a plurality of blank data each including a retrieval key element and a blank object element for a question to ask a value of truth or falsehood, the blank object element representing a kind of object to be recognized;

a question data conversion unit configured to decide whether the input sentence is a question to ask the value of truth or falsehood, to estimate an object to be recognized in the input sentence if the input sentence is the question to ask the value of truth or falsehood, to retrieve the blank data of which the retrieval key element is included in the input sentence from said blank data memory, and to convert the input sentence to question data in which the object is blanked by referring to the blank object element of the retrieved blank data; and

a problem solution unit configured to retrieve a solution for the object blanked in the question data and to generate a response in which the solution is included and an answer for the value of truth or falsehood is not included.

25. The natural language dialogue apparatus according to claim 24,

wherein said blank data memory stores a plurality of the blank data each including a question intention element and the blank object element corresponding to the retrieval key element, the retrieval key element representing a slot of arbitrary kind in the input sentence, the question intention element representing a kind of data necessary for the user in relation to the blank object element.

26. The natural language dialogue apparatus according to claim 25,

wherein said blank data memory stores a plurality of the blank data corresponding to the same retrieval key element in order of priority, each blank data including the question intention element and the blank object element.

27. The natural language dialogue apparatus according to claim 26,

wherein said question data conversion unit divides the input sentence into a plurality of slots of each kind if the input sentence is the question to ask the value of truth or falsehood.

28. The natural language dialogue apparatus according to claim 27,

wherein said question data conversion unit retrieves the blank data corresponding to the retrieval key element that coincides with one slot in the input sentence from said blank data memory, blanks another slot in the input sentence of which the kind is the same as the blank object element of the retrieved blank data, and converts a question intention to ask a value of truth or falsehood in the input sentence to the question intention element of the retrieved blank data.

29. The natural language dialogue apparatus according to claim 28,

wherein said question data conversion unit retrieves the blank data in order of the priority if a plurality of the blank data corresponding to the coincided retrieval key element are stored in said blank data memory.

30. The natural language dialogue apparatus according to claim 25,

wherein the retrieval key element is a verb as a movement element in the input sentence.

31. The natural language dialogue apparatus according to claim 30,

wherein said problem solution unit previously stores a plurality of set course data from a start point to an arrival point along a driving route, each set course data including a plurality of slots of each kind necessary for actual driving on the driving route, one of the plurality of slots being the movement element.

32. The natural language dialogue apparatus according to claim 31,

wherein said problem solution unit retrieves the set course data of which the movement element and at least one slot are equally included in the question data, and extracts another slot that is the same kind as the blanked slot included in the question data from the retrieved set course data as the solution.

33. The natural language dialogue apparatus according to claim 32,

wherein said question data conversion unit blanks a slot of another kind in the question data if said problem solution unit can not retrieve the set course data, and

wherein said problem solution unit retrieves the set course data of which the movement element and at least one slot are equally included in the question data in which the slot of another kind is blanked.

34. The natural language dialogue apparatus according to claim 32,

wherein said problem solution unit generates the response including the solution without the value of truth or falsehood by referring to the question intention element in the question data if said problem solution unit can retrieve the set course data, and

wherein said problem solution unit generates the response representing the value of falsehood in the input sentence if said problem solution unit can not retrieve the set course data.