IP Library Granted Patent US 7,383,542
Granted Patent B2
US 7,383,542 · App. 10/600,297 · Granted Jun 3, 2008

Adaptive machine translation service

View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 7,383,542
App. No.
10/600,297
Granted
Jun 3, 2008
Kind
B2
Abstract

An adaptive machine translation service for improving the performance of a user's automatic machine translation system is disclosed. A user submits a source document to an automatic translation system. The source document and at least a portion of an automatically generated translation are then transmitted to a reliable modification source (i.e., a human translator) for review and correction. Training material is generated automatically based on modifications made by the reliable source. The training material is sent back to the user together with the corrected translation. The user's automatic translation system is adapted based on the training material, thereby enabling the translation system to become customized through the normal workflow of acquiring corrected translations from a reliable source.

Claims (69)

1. A computer implemented method for providing information to an automatic machine translation system to improve translation accuracy, the method comprising:

receiving a collection of source text that is expressed in a first natural language;

receiving from the automatic machine translation system an attempted translation that corresponds to the collection of source text, wherein the attempted translation is expressed in a natural language other than the first natural language;

processing the attempted translation and the collection of source text to identify an error in the attempted translation; and

providing information to the automatic machine translation system to reduce the likelihood that the error will be repeated in subsequent natural language translations generated by the automatic machine translation system,

wherein providing information comprises providing information to be assimilated into the automatic machine translation system.

2. The method of claim 1 , wherein providing information to the automatic machine translation system comprises:

correcting the error; and

providing a corrected translation.

3. The method of claim 1 , wherein said receiving from the automatic machine translation system comprises receiving from a client upon which the automatic machine translation system is implemented.

4. The method of claim 3 , wherein receiving from a client comprises receiving by way of a computer network.

5. The method of claim 4 , wherein receiving by way of a computer network comprises receiving by way of the Internet.

6. The method of claim 1 , wherein said receiving from the automatic machine translation system comprises receiving from a server upon which the automatic machine translation system is implemented.

7. The method of claim 6 , wherein said receiving from a server comprises receiving by way of a computer network.

8. The method of claim 1 , wherein providing information to be assimilated comprises providing update information to be assimilated into a knowledge source associated with the automatic machine translation system.

9. The method of claim 1 , wherein providing information to be assimilated comprises providing update information to be assimilated into at least one translation correspondence associated with the automatic machine translation system.

10. The method of claim 1 , wherein providing information to be assimilated comprises providing update information to be assimilated into a collection of linguistic structures associated with the automatic translation system.

11. The method of claim 10 , wherein providing information to be assimilated comprises providing update information to be assimilated into a database of corresponding logical forms associated with the automatic machine translation system.

12. The method of claim 1 , wherein providing information to be assimilated comprises providing update information to be assimilated into a collection of statistical parameters associated with the automatic machine translation system.

13. The method of claim 1 , wherein providing information to be assimilated comprises providing update information to be assimilated into a collection of parsing information associated with the automatic machine translation system, the parsing information being information that enables a parser to provide analysis of a collection of segments.

14. The method of claim 1 , wherein providing information to be assimilated comprises providing update information to be assimilated into a collection of groups of corresponding words or phrases the associations associated with the automatic machine translation system.

15. The method of claim 1 , wherein providing information to be assimilated comprises providing bilingual corpora.

16. The method of claim 8 , wherein providing information to be assimilated comprises providing a bilingual corpus of one or more sentences.

17. A computer-implemented method for improving the performance of an automatic machine translation system, the method comprising:

employing the automatic machine translation system to generate a translation of a collection of source text, wherein the collection of source text is expressed in a first natural language and the translation is expressed in a natural language other than the first natural language;

transferring the collection of source text and at least a portion of the translation to a reliable modification source;

receiving from the reliable modification source an indication of an error in at least one portion of the translation; and

training the automatic machine translation system such that the error will be less likely to occur for subsequent translations generated by the automatic translation system,

wherein training the automatic machine translation system comprises updating a collection of parsing information associated with the automatic machine translation system, the parsing information being information that enables a parser to provide analysis of a collection of segments.

18. The method of claim 17 , further comprising:

generating a confidence metric representing a quality measurement with regard to the translation; and

selecting the a portion of the translation transferred to the reliable modification source based at least in part upon the confidence metric.

19. The method of claim 17 , wherein said transferring comprises transferring from a client computing device, upon which the automatic machine translation system is implemented, to a server computing device associated with the reliable modification source.

20. The method of claim 17 , wherein said transferring comprises transferring from a server, upon which the automatic machine translation system is implemented, to a server computing device associated with the reliable modification source.

21. The method of claim 17 , wherein training the automatic machine translation system comprises updating a knowledge source associated with the automatic machine translation system.

22. The method of claim 17 , wherein training the automatic machine translation system comprises updating at least one translation correspondence associated with the automatic machine translation system.

23. The method of claim 17 , wherein training the automatic machine translation system comprises updating a collection of linguistic structures associated with the automatic machine translation system.

24. The method of claim 23 , wherein training the automatic machine translation system comprises updating a database of corresponding logical forms associated with the automatic machine translation system.

25. The method of claim 17 , wherein training the automatic machine translation system comprises updating a collection of statistical parameters associated with the automatic machine translation system.

26. The method of claim 17 , wherein training the automatic machine translation system comprises updating a collection of corresponding word or phrase associations associated with the automatic machine translation system.

27. The method of claim 17 , wherein training the automatic machine translation system comprises providing bilingual corpora based on the error to the automatic machine translation system and enabling it to train itself based on the bilingual corpora.

28. The method of claim 17 , wherein training the automatic machine translation system comprises providing a bilingual corpus of one or more sentence pairs to the automatic machine translation system.

29. A method for improving the performance of an automatic machine translation system, the method comprising:

employing the automatic machine translation system to generate a translation of a collection of source text, a confidence metric being associated with portions of the translation, and wherein the collection of source text is expressed in a first natural language and the translation is expressed in a natural language other than the first natural language;

evaluating the confidence metric and selecting a low confidence portion of the translation;

transmitting the low confidence portion across a computer network to a reliable modification source;

utilizing the reliable modification source to generate a corrected version of the low confidence portion;

generating an updated database of translation knowledge based on the corrected version of the low confidence portion;

transmitting the updated database of translation knowledge across a computer network to the automatic machine translation system; and

incorporating the updated database of translation knowledge into the automatic machine translation system to enable the automatic machine translation system to subsequently translate with greater accuracy text similar to the low confidence portion,

wherein incorporating the database of translation knowledge comprises incorporating at least one updated linguistic structure.

30. The method of claim 29 , wherein utilizing the reliable modification source to generate a corrected version comprises utilizing a human translator.

31. The method of claim 29 , wherein transmitting across a computer network comprising transmitting across the Internet.

32. The method of claim 29 , wherein the automatic machine translation system is implemented on a client computing device.

33. The method of claim 29 , wherein incorporating the database of translation knowledge comprises incorporating at least one updated translation correspondence.

34. The method of claim 29 , wherein incorporating the database of translation knowledge comprises incorporating at least one update into a database of corresponding logical forms.

35. The method of claim 29 , wherein incorporating the database of translation knowledge comprises incorporating at least one update into a collection of statistical parameters.

36. The method of claim 29 , wherein incorporating the database of translation knowledge comprises incorporating at least one update into a collection of parsing information that enables a parser to provide analysis of a collection of segments.

37. The method of claim 29 , wherein incorporating the database of translation knowledge comprises incorporating at least one update into a collection of corresponding word or phrase associations.

38. A method for improving the performance of a self-customizing automatic machine translator, the method comprising:

implementing a first self-customizing automatic translator on a first computing device;

implementing a second self-customizing automatic translator on a second computing device;

providing a reliable translation source;

enabling communication between the first and second computing devices;

receiving at the second computing device a source text;

supplying the second computing device with a corrected version of an attempted translation produced by the reliable translation source, the attempted translation being an attempted translation of the source text, and wherein the source text is expressed in a first natural language and the attempted translation is expressed in a natural language other than the first natural language;

utilizing the second self-customizing automatic translator to process the source text and the corrected version of the attempted translation to produce training information for adapting the first self-customizing automatic translator to subsequently translate text similar to the source text with greater accuracy;

transferring the training information from the second computing device to the first computing device; and

assimilating the training information into the first self-customizing automatic translator to enable the first self-customizing automatic translator to subsequently translate with greater accuracy text similar to the source text.

Assignments (2)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Dec 9, 2014
From: MICROSOFT CORPORATION
To: MICROSOFT TECHNOLOGY LICENSING, LLC
Reel/Frame 034541/0477 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 20, 2003
From: RICHARDSON, STEPHEN D.; RASHID, RICHARD F.
To: MICROSOFT CORPORATION
Reel/Frame 014228/0564 →