Natural language processing system and method for expecting natural language information to be processed and for executing the processing based on the expected information

• US 6,029,123 A
• Filed: 12/12/1995
• Issued: 02/22/2000
• Est. Priority Date: 12/13/1994
• 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): Isen, Forester W.
• Assistant Examiner(s): Edouard, Patrick N.
• Application Number:
  US08/571,195
• Time in Patent Office: 1,533 Days
• Field of Search: 704/9, 704/10, 704/2, 704/1, 382/175, 382/229, 364/975, 395/12, 707/500
• US Class Current: 704/9
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation
• CPC Class Codes: G06F 40/279 Recognition of textual entities
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/274 Syntactic or semantic context, e.g. balancing

Abstract

A natural language processing system and method in which, in order to achieve a high accuracy of recognition of natural language information, a series of natural language information groups is progressively and sequentially recognized, with reference to a knowledge base which contains knowledge concerning the types of natural language information to be processed and knowledge concerning restriction in regard to the sequence of units of the series of natural language groups. In the course of the sequential recognition, the kind of the object to be recognized subsequently is expected based on the kinds of objects in the series of information groups which have been recognized, and a recognition result of the type which coincides with the expected type is delivered as a candidate output.

Claims

1. A natural language processing apparatus comprising:

recognition means for progressively and sequentially recognizing a series of natural language information groups of a plurality of types, each series comprising a sequence of words;

knowledge base means for storing knowledge of the types of said natural language information and restriction knowledge of restrictions on a sequence of words in a series of natural language information groups of the plurality of types; and

expectation means, which makes reference to the knowledge stored in said knowledge base means to supply a recognition candidate comprising a word expected to be next recognized by said recognition means, to said recognition means each time before said recognition means recognizes a next word, for use in recognizing the next word as one of the recognition candidates by said recognition means in processing the next word.

2. A natural language processing apparatus according to claim 1, wherein said recognition means selects a word from among the recognition candidates supplied by said expectation means.

3. A natural language processing apparatus according to claim 1, further comprising analysis means for analyzing natural language information obtained as a result of the recognition performed by said recognition means.

4. A natural language processing apparatus according to claim 1, wherein said expectation means generates a recognition candidate based on previous results of recognition by said recognition means by referring to the knowledge stored in said knowledge base means.

5. A natural language processing apparatus according to claim 1, wherein said recognition means recognizes character information of a natural language.

6. A natural language processing apparatus according to claim 1, wherein the restriction knowledge stored in said knowledge base means contains information concerning the physical position of said series of natural language information groups of different types with respect to each other on a physical medium.

7. A natural language processing apparatus according to claim 1, wherein said recognition means recognizes voice information of a natural language.

8. A natural language processing apparatus comprising:

input means for inputting natural language information comprising a series of natural language information groups of a plurality of types, each series comprising a sequence of words;

knowledge base means for storing knowledge of the types of said natural language information and restriction knowledge of restrictions on a sequence of words in a series of natural language information groups of the plurality of types;

expectation means for, prior to the inputting of a next word by said input means, generating an expected next word expected to be input next by said input means by making reference to the knowledge stored in said knowledge base means, and

correction means for correcting the next inputted word in accordance with the expected next word generated in said expectation means.

9. A natural language processing apparatus according to claim 8, further comprising analysis means for analyzing the natural language information obtained as a result of the correction performed by said correction means.

10. A natural language processing apparatus according to claim 8, wherein said expectation means generates an expected word, which is part of a series of natural language information groups based on a previously input word of the series by making reference to the knowledge stored in said knowledge base.

11. A natural language processing apparatus according to claim 8, wherein said input means inputs the results of recognition of character information of a natural language performed by a character recognition apparatus.

12. A natural language processing apparatus according to claim 8, wherein said input means inputs the results of recognition of voice information of a natural language performed by a voice recognition apparatus.

13. A natural language processing apparatus according to claim 8, wherein said input means inputs, as the results of recognition, a plurality of candidates for one object of recognition, and wherein said correction means performs correction by replacing the first one of said candidates with another.

14. A natural language processing apparatus according to claim 11, wherein restriction knowledge stored in said knowledge base means contains information concerning the physical position of said series of natural language information groups of different types with respect to each other on a physical medium.

15. A natural language processing apparatus according to claim 8, wherein said input means have keys through which characters are inputted.

16. A natural language processing method comprising:

a recognizing step for progressively and sequentially recognizing a series of natural language information groups of a plurality of types, each series comprising a sequence of words with recognizing means; and

an expecting step for referring to a knowledge base which contains knowledge of the types of said natural language information and knowledge of restrictions on a sequence of words in a series of natural language information groups of a plurality of types and for supplying a recognition candidate comprising a word expected to be next recognized in said recognizing step to said recognizing means each time before said recognizing step recognizes a next word, for use in recognizing the next word as one of the recognition candidates by said recognizing means in recognizing the next word.

17. A natural language processing method according to claim 16, wherein said recognizing step selects a word from among the recognition candidates supplied by said expectation step.

18. A natural language processing method according to claim 17, further comprising analyzing step in which the natural language information obtained as a result of recognition executed in said recognizing step is analyzed.

19. A natural language processing method according to claim 16, wherein said expecting step generates a recognition candidate based on the results of a previous recognition executed in said recognizing step, by making reference to the knowledge stored in said knowledge base.

20. A natural language processing method according to claim 16, wherein said recognizing step recognizes character information of a natural language.

21. A natural language processing method according to claim 16, wherein said expecting step makes reference to layout information concerning layout of said series of natural language information groups of a plurality of types on a sheet, said layout information being stored in said knowledge base and used as said information concerning restriction.

22. A natural language processing method according to claim 16, wherein said recognizing step recognizes voice information of a natural language.

23. A natural language processing method comprising:

an inputting step in which a series of natural language information groups of a plurality of types is inputted, each series comprising a sequence of words; and

an expecting step for, prior to the inputting of each word in said inputting step, generating an expected word expected to be input in said inputting step by referring to knowledge stored in a knowledge base which stores knowledge of the types of said natural language information and knowledge of restrictions on a sequence of words in a series of natural language information groups of the plurality of types; and

a correction step for correcting a next inputted word in accordance with the expected next word generated in said expecting step.

24. A natural language processing method according to claim 23, further comprising an analyzing step in which the natural language information obtained as the result of the correction executed in said correcting step is analyzed.

25. A natural language processing method according to claim 23, wherein said expecting step generates an expected word, which is part of a series of natural language information groups based on a previously input word of the series by making reference to the knowledge stored in said knowledge base.

26. A natural language processing method according to claim 23, wherein said inputting step inputs results of recognition of character information of a natural language performed in accordance with a character recognition technique.

27. A natural language processing method according to claim 23, wherein said inputting step inputs the results of recognition of voice information of a natural language performed in accordance with a voice recognition technique.

28. A natural language processing method according to claim 23, further comprising the step of recognizing words of the natural language information, wherein said inputting step inputs a plurality of candidates to be used in said recognizing step for recognizing one word, and wherein said correcting step executes correction by replacing the first one of said candidates with another of said candidates.

29. A natural language processing method according to claim 26, wherein said expecting step makes reference to layout information concerning layout of said series of natural language information groups of a plurality of types on a sheet, said layout information being stored in said knowledge base and used as said information concerning restriction.

30. A natural language processing method according to claim 24, wherein said inputting step inputs characters through keys.

31. A computer-readable storage medium storing a program for controlling a computer, said program comprising codes for causing the computer to perform:

a recognizing step for progressively and sequentially recognizing with recognizing means a series of natural language information groups of a plurality of types, each series comprising a sequence of words; and

an expecting step for referring to a knowledge base which contains knowledge of the types of said natural language information and knowledge of restrictions on a sequence of words in a series of natural language information groups of a plurality of types and for supplying a recognition candidate comprising a word expected to be next recognized in said recognizing step to said recognition means each time before said recognizing step recognizes a next word, for use in recognizing the next word as one of the recognition candidates by said recognition means in recognizing the next word.

32. A computer-readable storage medium storing a program for controlling a computer, said program comprising codes for causing the computer to perform:

an inputting step in which a series of natural language information groups of a plurality of types is inputted, each series comprising a sequence of words

an expecting step for, prior to the inputting of each word in said inputting step, generating an expected word expected to be input in said inputting step by referring to knowledge stored in a knowledge base which stores knowledge of the types of said natural language information and knowledge of restrictions on a sequence of words in a series of natural language information groups of the plurality of types; and

a correction step for correcting a next inputted word in accordance with the expected next word generated in said expecting step.