Deep Model Statistics Method for Machine Translation

• US 20090182549A1
• Filed: 02/18/2009
• Published: 07/16/2009
• Est. Priority Date: 10/10/2006
• Status: Active Grant

• Current Assignee: ABBYY Production LLC (ABBYY Software)
• Original Assignee: ABBYY Software
• Inventors: Zuev, Konstantin, Tuganbaev, Diar, Anisimovich, Konstantin, Selegey, Vladimir
• Granted Patent: US 8,145,473 B2
• Time in Patent Office: Days
• US Class Current: 704/4
• CPC Class Codes: G06F 40/268 Morphological analysis
• CPC Class Codes: G06F 40/284 Lexical analysis, e.g. tokenisation or collocates
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06F 40/44 Statistical methods, e.g. probability models
• CPC Class Codes: G06F 40/55 Rule-based translation
• CPC Class Codes: G06F 40/58 Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation

Abstract

In one embodiment, the invention provides a method for machine translation of a source document in an input language to a target document in an output language, comprising generating translation options corresponding to at least portions of each sentence in the input language; and selecting a translation option for the sentence based on statistics associated with the translation options.

Claims

1. A method for machine translation of a source document in an input language to a target document in an output language, comprising:

generating translation options corresponding to at least portions of each sentence in the input language; and

selecting a translation option for the sentence based on ratings associated with the translation options.

2. The method of claim 1, wherein the ratings comprise statistics.

3. The method of claim 2, wherein the statistics are used in generating the translation options.

4. The method of claim 3, wherein generating the translation options comprises syntactic analysis to determine surface elements associated with the sentence and semantic analysis to determine deep elements associated with the sentence.

5. The method of claim 4, wherein the surface elements are selected from the group consisting of lexemes, word combinations, lexical meanings, and surface slots.

6. The method of claim 4, wherein the deep elements are selected from the group consisting of semantic classes, and deep slots.

7. The method of claim 4, wherein the translation options comprise lexical choices in the output language corresponding to said portions of each sentence in the input language.

8. The method of claim 4, wherein the statistics comprise a combinatorial rating for the surface elements and deep elements.

9. The method of claim 8, wherein the combinatorial rating comprises a probability of objects of certain semantic classes being combined with objects of the same or another semantic class.

10. The method of claim 8, wherein the combinatorial rating comprises a probability of objects of certain semantic classes being combined with certain deep slots.

11. The method of claim 2, further comprising generating the statistics.

12. The method of claim 11, wherein generating the statistics comprising analyzing an assortment of texts to identify frequently occurring lexemes and lexeme combinations.

13. The method of claim 12, wherein the assortment of texts is selected from the group consisting of parallel texts, tagged texts, and domain-specific texts.

14. A computer storage medium having stored thereon, a sequence of instructions which when executed by a computer system, cause said computer system to perform a method for machine translation of a source document in an input language to a target document in an output language, comprising:

generating translation options corresponding to at least portions of each sentence in the input language; and

selecting a translation option for the sentence based on ratings associated with the translation options.

15. The computer storage medium of claim 14, wherein the ratings comprise statistics.

16. The computer storage medium of claim 15, wherein generating the translation option is based on the statistics.

17. The computer storage medium of claim 16, wherein generating the translation options comprises syntactic analysis to determine surface elements associated with the sentence and semantic analysis to determine deep elements associated with the sentence.

18. The computer storage medium of claim 15, wherein the surface elements are selected from the group consisting of lexemes, word combinations, lexical meanings, and surface slots.

19. The computer storage medium of claim 15, wherein the deep elements are selected from the group consisting of semantic classes, and deep slots.

20. The computer storage medium of claim 14, wherein the translation options comprise lexical choices in the output language corresponding to said portions of each sentence in the input language.

21. The computer storage medium of claim 17, wherein the statistics comprise a combinatorial rating for the surface elements and deep elements.

22. The computer storage medium of claim 21, wherein the combinatorial rating comprises a probability of objects of certain semantic classes being combined with objects of the same or another semantic class.

23. A computer system, comprising:

a processor; and

a storage medium coupled to the processor, the storage medium storing instructions which when executed by the processor cause the computer system to perform a method for machine translation of a source document in an input language to a target document in an output language, comprising;

generating translation options corresponding to at least portions of each sentence in the input language; and

selecting a translation option for the sentence based on ratings associated with the translation options.

24. The computer system of claim 23, wherein the ratings comprise statistics.

25. The computer system of claim 24, wherein generating the translation options is based on the statistics.

26. The computer system of claim 23, wherein generating the translation options comprises syntactic analysis to determine surface elements associated with the sentence and semantic analysis to determine deep elements associated with the sentence.