Natural-language processing system and dictionary registration system

• US 20090281786A1
• Filed: 09/06/2007
• Published: 11/12/2009
• Est. Priority Date: 09/07/2006
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Doi, Shinichi, Ando, Shinichi, Sadamasa, Kunihiko
• Granted Patent: US 9,575,953 B2
• Time in Patent Office: Days
• US Class Current: 704/2
• CPC Class Codes: G06F 40/00 Handling natural language data speech analysis or synthesis, speech recognition G10L
• CPC Class Codes: G06F 40/129 Handling non-Latin characters, e.g. kana-to-kanji conversion
• CPC Class Codes: G06F 40/242 Dictionaries
• CPC Class Codes: G06F 40/247 Thesauruses; Synonyms
• CPC Class Codes: G06F 40/49 using very large corpora, e.g. the web
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

Abstract

A natural-language processing system (10) includes a registration-candidate storage section (32) that stores therein registration-candidate dictionary data, a judgment means (22) that compares input data against the registration-candidate dictionary data to thereby judge whether or not the input data includes a word corresponding to the registration-candidate dictionary data, an inquiry means (23) that inquires to a user whether or not corresponding dictionary data is to be registered in a dictionary storage section (31) to accept a user'"'"'s instruction if it is judged that a corresponding word exists, a dictionary registration means (24) that registers the corresponding dictionary data in the dictionary storage section based on the input instruction, and a natural language processing means (25) that executes a natural-language processing onto the input data by using the dictionary data registered in the dictionary storage section.

Claims

1. A natural-language processing system that executes a natural-language processing onto input data, comprising:

a registration-candidate storage section that stores therein registration-candidate dictionary data;

a judgment section that compares the input data against the registration-candidate dictionary data, to judge whether or not the input data includes a word corresponding to the registration-candidate dictionary data;

an inquiry section that inquires to a user whether or not the registration-candidate dictionary data is to be registered in a dictionary storage section, to thereby accept a user'"'"'s instruction, if the judgment section judges that the corresponding word exists;

a dictionary registration section that registers the registration-candidate dictionary data in the dictionary storage section based on the instruction input to the inquiry section; and

a natural-language processing section that executes a natural-language processing onto the input data by using the dictionary data including the registration-candidate dictionary data when said dictionary registration section registers the registration-candidate dictionary data, and executes a natural-language processing onto the input data by using the dictionary data that does not include the registration-candidate dictionary data when said dictionary registration section does not register the registration-candidate dictionary data.

2. The natural-language processing system according to claim 1, wherein:

said natural-language processing section comprises an unknown-word detection function that detects an unknown word included in the input data; and

said judgment section starts operation if said natural-language processing section detects an unknown word included in the input data.

3. The natural-language processing system according to claim 2, wherein said judgment section compares a portion of the input data including, the unknown word against the registration-candidate dictionary data to thereby judge whether or not the corresponding word exists in said portion including the unknown word, if said natural-language processing section detects the unknown word included in the input data.

4. The natural-language processing system according to claim 1, wherein said natural-language processing section executes the natural-language processing onto the input data by using only the dictionary data registered in said dictionary storage section.

5. The natural-language processing system according to claim 1, wherein said natural-language processing section comprises a machine-translation section that converts a character string of a:

first language into another character string of a second language, and/or a voice-synthesis section that converts an input character string into an audio signal.

9. The dictionary registration, system according to claim 1, wherein:

the dictionary data includes dictionary data used in a machine-translation processing that converts a character string of an input first language into another character string of a second language, and/or dictionary data used in a voice-synthesis processing that converts an input character string into an audio signal.

10. The dictionary registration system according to claim 1, wherein:

the dictionary data includes dictionary data used in a machine-translation processing that converts, a character string of an input first language into another character string of a second language, dictionary data used in a kana-kanji conversion processing that converts an input kana character string into a kana-kanji-mixed character string, and/or dictionary data used in a speech-recognition processing that converts an input audio signal into a character string; and

said judgment section compares, upon comparing the input data against the registration-candidate dictionary data, the input data against a character string of the registration-candidate data that is stored in said registration-candidate storage section as a result of conversion result to thereby judge whether or not the corresponding word exists.

14. The natural-language processing method according to claim 1, wherein the natural-language processing uses only the dictionary data stored in the dictionary.

15. The natural-language processing method according to claim 1, wherein the natural-language processing includes a machine-translation processing that converts a character string of a first language into another character string of a second language, and/or a voice-synthesis processing that converts an input character string into an audio signal.

6. A dictionary registration system comprising:

a registration-candidate storage section that stores therein registration-candidate dictionary data;

a judgment section that compares input data against the registration-candidate dictionary data to thereby judge whether or not the input data includes a word corresponding to the registration-candidate dictionary data;

an inquiry section that inquires to a user whether or not the registration-candidate dictionary data is to be registered in a dictionary storage section to accept a user'"'"'s instruction, if said judgment section judges that the corresponding word exists; and

a dictionary registration section that registers the registration-candidate dictionary data in said dictionary storage section based on the instruction input to said inquiry section.

7. The dictionary registration system according to claim 6, further comprising an unknown-word detection section that detects an unknown word included in the input data, wherein:

said judgment section starts operation if said unknown-word detection section detects an unknown word included, in the input data.

8. The dictionary registration system according to claim 7, wherein, if said unknown-word detection section detects an unknown word included in the input data, said judgment section compares a portion of the input data including the unknown word against the registration-candidate dictionary data to thereby judge whether or not the corresponding word exists in said portion including the unknown word.

11. A natural-language processing method comprising:

comparing input data against registration-candidate dictionary data stored in a storage unit, to thereby judge whether or not the input data includes a word corresponding to the registration-candidate dictionary data;

inquiring to a user whether or not the registration-candidate dictionary data is to be registered in a dictionary to accept a user'"'"'s instruction, if it is judged that the corresponding word exists;

registering the registration-candidate dictionary data in the dictionary based on the input instruction; and

executing a natural-language processing onto the input data by using the dictionary data including the registration-candidate dictionary data when said dictionary registration section registers the registration-candidate dictionary data, and executing a natural-language processing onto the input data by using the dictionary data that does not include the registration-candidate dictionary data when said dictionary registration section does not register the registration-candidate dictionary data.

12. The natural-language processing method according to claim 11, further comprising:

judging whether or not the input data includes an unknown word; and

comparing, upon judging that the corresponding word exists in the input data, a portion of the input data including the unknown data against the registration-candidate dictionary data to thereby judge whether or not the corresponding word exists in the portion including the unknown word.

13. The natural-language processing method according to claim 12, further comprising comparing, upon judging that the unknown word exists in the input data, a portion of the input data including the unknown word against the registration-candidate dictionary data to thereby judge whether or not the corresponding word exists in said portion including the unknown word.

16. A dictionary registration method comprising:

comparing input data against registration-candidate dictionary data stored in a storage unit to thereby judge whether or not the input data includes a word corresponding to the registration-candidate dictionary data;

inquiring to a user whether or not the registration-candidate dictionary data is to be registered in a dictionary to accept a user'"'"'s instruction, if it is judged that the corresponding word exists; and

registering the registration-candidate dictionary data in the dictionary based on the input instruction.

17. The dictionary registration method according to claim 16, further comprising:

judging whether or not the input data includes an unknown word; and

comparing a portion of the input data including the unknown word against the registration-candidate dictionary data to thereby judge whether or not the corresponding word exists in said portion including the unknown word, if it is judged that the input data includes the unknown word.

18. The dictionary registration method according to claim 17, wherein, if it is judged that the input data includes the unknown word, a portion of the input data including the unknown word is compared against the registration-candidate dictionary data to thereby judge whether or not the corresponding word exists in said portion including the unknown word.

19. The dictionary registration method according to claim 16, wherein the dictionary data includes dictionary data used in a machine-translation processing that converts a character string of an input first language into another character string of a second language, and/or dictionary data used in a voice-synthesis processing that converts an input character string into an audio signal.

20. The dictionary registration method according to claim 16, wherein:

the dictionary data includes dictionary data used in a machine-translation processing that converts a character string of an input first language into another character string of a second language, dictionary data used in a kana-kanji conversion processing that converts an input kana character string into a kana-kanji-mixed character string, and/or dictionary data used in a speech-recognition processing that converts an input audio signal into a character string; and

said comparing of the input data against the registration-candidate-dictionary data compares the input data against a character string that is stored in the registration-candidate dictionary data as a result of conversion result, to thereby judge whether or not the corresponding word exists.

21. A recording medium encoded with a computer-readable program that causes a central processing unit (CPU) to operate, said program causes said CPU to execute the processings of:

storing registration-candidate dictionary data in a storage unit;

comparing input data against the registration-candidate dictionary data, and thereby judging whether or not the input data includes a word corresponding to the registration-candidate dictionary data;

inquiring to a user whether or not the registration-candidate dictionary data is to be registered in a dictionary to accept a user'"'"'s instruction, if said judging processing judges that the corresponding word exists;

registering the registration-candidate dictionary data in the dictionary based on an instruction input in said inquiring processing;

executing a natural-language processing onto the input data by using the dictionary data including the registration-candidate dictionary data when said dictionary registration section registers the registration-candidate dictionary data, and executing a natural-language processing onto the input data by using the dictionary data that does not include the registration-candidate dictionary data when said dictionary registration section does not register the registration-candidate dictionary data.

22. The medium according to claim 21, wherein:

said natural-language processing comprises judging whether or not the input data includes an unknown word; and

said program causes said CPU to execute said judging processing, if said natural-language processing judges that the unknown word exists in the input data.

23. The medium according to claim 22, said judging processing compares a portion of the input data including the unknown word against the registration-candidate dictionary data to thereby judge whether or not the corresponding word exists in said portion including the unknown word, if said natural-language processing judges that the unknown word exists in the input data.

24. The medium according to claim 21, wherein the natural-language processing executes the natural-language processing onto the input data by using only the dictionary data stored in the dictionary.

25. The medium according to claim 21, wherein the natural-language processing has a machine-translation function that converts a character string of a first language into another character string of a second language, and/or a voice-synthesis function that converts an input character string into an audio signal.

26. A recording medium encoded with a computer-readable program that causes a central processing unit (CPU) to operate, said program causes said CPU to execute the processings of:

storing registration-candidate dictionary data;

comparing input data against the registration-candidate dictionary data and judging whether or not the input data includes a word corresponding to the registration-candidate dictionary data;

inquiring to a user whether or not the registration-candidate dictionary data is to be registered in a dictionary, to accept a user'"'"'s instruction, if said judgment processing judges that the corresponding word exists; and

registering the registration-candidate dictionary data in the dictionary based on an instruction input in said inquiry processing.

27. The medium according to claim 26, further causing said CPU to execute the processing of judging whether or not the input data includes an unknown word, wherein:

said program causes said CPU to execute said judging processing, if said unknown-word detecting processing detects the unknown word included in the input data.

28. The medium according to claim 27, wherein, if said unknown-word detecting processing detects the unknown word included in the input data, said judging processing compares a portion of the input data including the unknown word against the registration-candidate dictionary data to thereby judge whether or not the corresponding word exists in said portion including the unknown word.

29. The medium according to claim 26, wherein the dictionary data includes dictionary data used in a machine-translation function that converts a character string of an input first language into another character string of a second language, and/or dictionary data used in a voice-synthesis function that converts an input character string into an audio signal.

30. The medium according to claim 26, wherein:

the dictionary data includes dictionary data used in a machine-translation function that converts a character string of an input first language into another character string of a second language, dictionary data used in a kana-kanji conversion function that converts an input kana character string into a kana-kanji-mixed character string, and/or dictionary data used in a speech recognition;

function that converts an input audio signal into a character string; and

wherein said judging processing compares, upon comparing the input data against the registration-candidate dictionary data, the input data against the registration-candidate dictionary data that is stored in the registration-candidate dictionary data as a result of conversion result, to thereby judge whether or not the corresponding word exists.