Text preparing system using knowledge base and method therefor

• US 5,802,504 A
• Filed: 06/21/1995
• Issued: 09/01/1998
• Est. Priority Date: 06/21/1994
• Status: Expired due to Term

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Jeyachandran, Suresh, Suda, Aruna Rohra
• Primary Examiner(s): Hafiz, Tariq R.
• Application Number:
  US08/492,930
• Time in Patent Office: 1,168 Days
• Field of Search: 395/10-13, 395/50-51, 395/60-61, 395/338, 395/772, 395/605, 395/54
• US Class Current: 706/11
• CPC Class Codes: G06F 40/166 Editing, e.g. inserting or deleting

Abstract

System and method for preparing a text by using a knowledge base, by which knowledge relating to input information is first retrieved from the knowledge base. Then, a plurality of kinds of information used as elements of a text to be generated is selected from the input information and the retrieved information relating thereto. Subsequently, a describing order, in which the selected elements of the text are described, is determined. Then, a text to be outputted is generated on the basis of the selected elements of the text, the describing order of which has been determined. Thereby, a text, whose contents and structure meet a user'"'"'s purpose, is generated from the input simple information.

Claims

1. A text preparing system comprising:

input means for inputting information;

a knowledge base for storing knowledge including a text preparation rule which defines kinds of contents in a text and an order of the contents in the text;

retrieval means for retrieving knowledge, which relates to the information inputted by said input means, from said knowledge base;

collection means for collecting information corresponding to the kinds of contents defined by the text preparation rule from the information inputted by said input means and the knowledge retrieved by said retrieval means;

order determining means for determining a describing order of the information collected by said collection means in accordance with the order of the contents in the text defined by the text preparation rule; and

generating means for generating the text containing the information collected by said collection means in the describing order determined by said order determining means.

2. The text preparing system according to claim 1, wherein said knowledge base has knowledge relating to a reader of a text to be generated by said generating means, wherein said collection means collects information corresponding to the kinds of contents defined by the text preparation rule, on the basis of the knowledge relating to the reader of the text.

3. The text preparing system according to claim 2, wherein said knowledge base stores a level of the reader'"'"'s understanding therein as the knowledge relating to the reader of the text.

4. The text preparing system according to claim 2, wherein said knowledge base stores a relation between an author and the reader of the text therein as the knowledge relating to the reader of the text.

5. The text preparing system according to claim 4, wherein the relation includes a difference in social position between the author and the reader of the text therein as the knowledge relating to the reader of the text.

6. The text preparing system according to claim 4, wherein a text to be generated by said generating means is a letter, wherein the relation between the author and the reader includes information relating to precedent correspondence between a sender and a receiver of the letter.

7. The text preparing system according to claim 6, wherein said knowledge base stores a text preparation rule in accordance with the precedent correspondence between the sender and the receiver.

8. The text preparing system according to claim 7, wherein the text preparation rule depends on absence of the precedent correspondence between the sender and the receiver.

9. The text preparing system according to claim 7, wherein the text preparation rule includes a text preparation rule that depends on a frequency of the precedent correspondence between the sender and the receiver.

10. The text preparing system according to claim 7, wherein the text preparation rule includes a text preparation rule that depends on the latest correspondence between the sender and the receiver.

11. The text preparing system according to claim 7, wherein the text preparation rule includes a text preparation rule that depends on whether or not the letter to be generated is a reply to a letter from the receiver.

12. The text preparing system according to claim 4, wherein a text to be generated by said generating means is a letter and wherein the relation between the author and the reader includes a closeness between a sender and a receiver of the letter.

13. The text preparing system according to claim 1, wherein said knowledge base has knowledge relating to a reader of a text to be generated by the generating means, wherein said order determining means determines the describing order of the information collected by said collecting means in accordance with the order of the contents in the text defined by the text preparation rule, on the basis of the knowledge relating to the reader of the text.

14. The text preparing system according to claim 13, wherein said knowledge base stores a level of the reader'"'"'s understanding therein as the knowledge relating to the reader of the text.

15. The text preparing system according to claim 13, wherein said knowledge base stores a relation between an author and the reader of the text therein as the knowledge relating to the reader of the text.

16. The text preparing system according to claim 15, wherein the relation includes a difference in social position between the author and the reader of the text therein as the knowledge relating to the reader of the text.

17. The text preparing system according to claim 1, wherein said knowledge base has knowledge relating to a document type of a text to be generated by said generating means, wherein said collection means collects information corresponding to the kinds of contents defined by the text preparation rule on the basis of the knowledge relating to the document type of the text.

18. The text preparing system according to claim 1, wherein said knowledge base has knowledge relating to a document type of a text to be generated by the generating means, wherein said order determining means determines the describing order of the information collected by said collection means in accordance with the order of the contents in the text defined by the text preparation rule, on the basis of the knowledge relating to the document type of the text.

19. The text preparing system according to claim 1, wherein said knowledge base includes knowledge relating to a document type and a tone of a text to be generated by said generating means and wherein said generating means generates the text in the tone corresponding to the document type of the text on the basis of the knowledge relating to the document type and the tone of the text.

20. A text preparing system comprising:

input means for inputting information;

a knowledge base for storing knowledge including a text preparation rule which defines kinds of contents in a text and an order of the contents in the text;

understanding means for retrieving knowledge, which relates to the information inputted by said input means, from said knowledge base, for understanding a user'"'"'s intention and for outputting information which relates to a text to be generated;

collection means for collecting information corresponding to the kinds of contents defined by the text preparation rule, from the information outputted by said understanding means;

order determining means for determining a describing order of the information collected by said collecting means in accordance with the order of the contents in the text defined by the text preparation rule; and

generating means for generating the text containing the information collected by said collection means in the describing order determined by said order determining means.

21. The text preparing system according to claim 20, wherein said input means inputs information represented in a natural language, wherein said understanding means has analysis means for analyzing a concept corresponding to the information inputted and represented in the natural language.

22. The text preparing system according to claim 20, wherein said knowledge base stores a history of operations performed by an operator of the system therein.

23. The text preparing system according to claim 20, wherein said knowledge base includes goal-plan knowledge relating to a goal and a plan and wherein said understanding means retrieves the goal-plan knowledge relating to the inputted information and understands the user'"'"'s intention on the basis of the retrieved goal-plan knowledge.

24. The text preparing system according to claim 23, wherein said understanding means infers a main goal of the letter and a main plan for the goal on the basis of the retrieved goal-plan knowledge.

25. The text preparing system according to claim 23, wherein the main goal and the main plan are included in the information collected by said collection means.

26. The text preparing system according to claim 25, wherein said order determining means determines the describing order of the information in order of the main goal, the main plan and other information.

27. The text preparing system according to claim 26, wherein the other information includes a sub goal.

28. A text preparing system comprising:

input means for inputting information;

a knowledge base for storing knowledge therein, said knowledge base stores a history of an operation of requesting a same text performed by an operator of the system;

understanding means for retrieving knowledge, which relates to the information input by said input means, from said knowledge base, for understanding a user'"'"'s intention and for outputting information which relates to a text to be generated;

selection means for selecting information, which should be used as an element of a text to be prepared, from the information outputted by said understanding means;

order determining means for determining a describing order in which the information selected by said selection means is described; and

generating means for generating the text, which should be outputted from the information which is selected by said selection means and has the describing order determined by said order determining means,

wherein said selection means refers to the history and selects an abundance of information correspondingly to a first request as information to be used as elements of the text but selects simple information correspondingly to a second request or later requests as the information to be used as the elements of the text.

29. A text preparing method comprising the steps of:

inputting information;

retrieving information relating to the inputted information from a knowledge base for storing knowledge including a text preparation rule which defines kinds of contents in a text and an order of the contents in the text;

collecting information corresponding to the kinds of contents defined by the text preparation rule, from the inputted information and the retrieved knowledge;

determining a describing order of the information collected in said collection step in accordance with the order of the contents in the text defined by the text preparation rule; and

generating a text containing the collected information in the describing order determined.

30. The text preparing method according to claim 29, wherein the knowledge base has knowledge relating to a reader of a text to be generated in said generating step, wherein in said collection step, the information corresponding to the kinds of contents defined by the text preparation rule, is collected on the basis of the knowledge relating to the reader of the text.

31. The text preparing method according to claim 30, wherein the knowledge relating to the reader of the text includes a level of the reader'"'"'s understanding.

32. The text preparing method according to claim 30, wherein the knowledge relating to the reader of the text includes a relation between an author and the reader of the text.

33. The text preparing method according to claim 32, wherein the knowledge relating to the reader of the text includes a difference in social position between the author and the reader of the text.

34. The text preparing method according to claim 32, wherein a text to be generated in said generating step is a letter and wherein the relation between the author and the reader includes information relating to precedent correspondence between a sender and a receiver of the letter.

35. The text preparing method according to claim 34, wherein said knowledge base stores a text preparation rule in accordance with the precedent correspondence between the sender and the receiver.

36. The text preparing method according to claim 35, wherein the text preparation rule depends on absence of the precedent correspondence between the sender and the receiver.

37. The text preparing method according to claim 35, wherein the text preparation rule includes a text preparation rule that depends on a frequency of the precedent correspondence between the sender and the receiver.

38. The text preparing method according to claim 35, wherein the text preparation rule includes a text preparation rule that depends on the latest correspondence between the sender and the receiver.

39. The text preparing method according to claim 35, wherein the text preparation rule includes a text preparation rule that depends on whether or not the letter to be generated is a reply to a letter from the receiver.

40. The text preparing method according to claim 32, wherein a text to be generated in said generating step is a letter and wherein the relation between the author and the reader includes a closeness between a sender and a receiver of the letter.

41. The text preparing method according to claim 29, wherein the knowledge base has knowledge relating to a reader of a text to be generated in said generating step, wherein in said determining step, the describing order of the information collected in said collection step in accordance with the order of the contents in the text defined by the text preparation rule, is determined on the basis of the knowledge relating to the reader of the text.

42. The text preparing method according to claim 41, wherein the knowledge base stores a level of the reader'"'"'s understanding therein as the knowledge relating to the reader of the text.

43. The text preparing method according to claim 41, wherein the knowledge base stores a relation between an author and the reader of the text therein as the knowledge relating to the reader of the text.

44. The text preparing method according to claim 43, wherein the relation includes a difference in social position between the author and the reader of the text therein.

45. The text preparing method according to claim 29 wherein the knowledge base has knowledge relating to a document type of a text to be generated in said generating step, wherein in said collection step information corresponding to the kinds of contents defined by the text preparation rule is collected on the basis of the knowledge relating to the reader of the text.

46. The text preparing method according to claim 29, wherein the knowledge base has knowledge relating to a document type of a text to be generated in said generating step, wherein in said order determining step, the describing order collected in said collection step in accordance with the order of the contents in the text defined by the text preparation rule, is determined on the basis of the knowledge relating to the document type of the text.

47. The text preparing method according to claim 29, wherein the knowledge base includes knowledge relating to a document type and a tone of a text to be generated in said generating step and wherein in said generating step the text is generated in the tone corresponding to the document type of the text on the basis of the knowledge relating to the document type and the tone of the text.

48. A text preparing method comprising the steps of:

inputting information;

understanding a user'"'"'s intention by retrieving information relating to the inputted information from a knowledge base for storing knowledge including a text preparation rule which defines kinds of contents in a text and an order of the contents in the text;

outputting information relating to a text according to the understood intention;

collecting information corresponding to the kinds of contents defined by the text preparation rule, from the outputted information;

determining a describing order of the collected information in accordance with the order of the contents in the text defined by the text preparation rule; and

generating a text containing the collected information in the describing order determined.

49. The text preparing method according to claim 48, wherein in the inputting step information is represented in a natural language, wherein the step of understanding the user'"'"'s intention has the sub-step of analyzing a concept corresponding to the information inputted and represented in the natural language.

50. The text preparing method according to claim 48, wherein the knowledge base stores a history of operations performed by an operator of a system for preparing the text.

51. The text preparing method according to claim 48, wherein the knowledge base includes goal-plan knowledge relating to a goal and a plan and wherein in said understanding step the goal-plan knowledge relating to the inputted information is retrieved and the user'"'"'s intention is understood on the basis of the retrieved goal-plan knowledge.

52. The text preparing method according to claim 51, wherein in said understanding step a main goal of the letter and a main plan for the goal are inferred on the basis of the retrieved goal-plan knowledge.

53. The text preparing method according to claim 51, wherein the main goal and the main plan are included in the information collected in said collection step.

54. The text preparing method according to claim 53, wherein in said order determining step the describing order of the information is determined in order of the main goal, the main plan and other information.

55. The text preparing method according to claim 54, wherein the other information includes a sub goal.

56. A text preparing method comprising the steps of:

inputting information;

understanding a user'"'"'s intention by retrieving information relating to the inputted information from a knowledge base for storing knowledge;

outputting information relating to a text, which should be generated, according to the understood intention;

selecting information, which should be used as an element of the text to be generated, from the outputted information;

determining an order in which the selected information is described; and

generating a text, which should be outputted, from the information which is selected and has the describing order determined,

wherein the knowledge base stores a history of an operation of requesting a same text performed by an operator, wherein in said step of selecting information, the history is referred to, and an abundance of information is selected correspondingly to a first request as information to be used as elements of the text but simple information is selected correspondingly to a second request or later requests as the information to be used as the elements of the text.

57. A computer-readable storage medium storing a text preparing program for controlling a computer to prepare a text, said program comprising codes for causing the computer to perform the steps of:

inputting information;

retrieving information relating to the inputted information from a knowledge base for storing knowledge including a text preparation rule which defines kinds of contents in a text and an order of the contents in the text;

collecting information corresponding to the kinds of contents defined by the text preparation rule, from the inputted information and the retrieved knowledge;

determining a describing order of the information collected in said collection step in accordance with the order of the contents in the text defined by the text preparation rule; and

generating a text containing the collected information in the describing order determined.

58. A computer-readable storage medium storing a text preparing program for controlling a computer to prepare a text, said program comprising codes for causing the computer to perform the steps of:

inputting information;

understanding a user'"'"'s intention by retrieving information relating to the inputted information from a knowledge base for storing knowledge including a text preparation rule which defines kinds of contents in a text and an order of the contents in the text;

outputting information relating to a text according to the understood intention;

collecting information corresponding to the kinds of contents defined by the text preparation rule, from the outputted information;

determining a describing order of the collected information in accordance with the order of the contents in the text defined by the text preparation rule; and

generating a text containing the collected information in the describing order determined.