Computer-implemented method, computer software and apparatus for use in a translation system

• US 9,400,786 B2
• Filed: 07/25/2013
• Issued: 07/26/2016
• Est. Priority Date: 09/21/2006
• Status: Active Grant

• Current Assignee: SDL Ltd.
• Original Assignee: SDL PLC (RWS Holdings Plc)
• Inventors: Lancaster, Mark, Gordon, Alastair, Mills, Keith
• Primary Examiner(s): SAINT CYR, LEONARD
• Application Number:
  US13/951,451
• Publication Number:
  US 20130346062A1
• Time in Patent Office: 1,097 Days
• Field of Search: 704 2- 10
• US Class Current: 1/1
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30
• CPC Class Codes: G06F 40/45 Example-based machine translation; Alignment
• CPC Class Codes: G06F 40/47 Machine-assisted translation, e.g. using translation memory
• 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

A computer-implemented method for use in natural language translation comprises performing in software processes, the steps of: comparing source material with stored material in a first natural language, said stored material having previously been translated from said first natural language to at least a second natural language, identifying at least a part of said source material which has a relationship with at least a part of said stored material, outputting said identified part of source material and said identified part of stored material in a form suitable for review by a user, and replacing said identified part of source material with said identified part of stored material to assist full translation of said source material from said first natural language to at least said second natural language.

Claims

1. A computer-implemented method for use in natural language translation, said method comprising performing in software processes, the steps of:

receiving source material segments being authored by a user in a first natural language;

comparing, utilizing a module stored in memory and communicatively coupled with a processor, the received source material segments with stored material segments in the first natural language, said stored material segments having previously been translated from said first natural language to target segments in a second natural language;

identifying a plurality of source material segments that are similar to stored material segments using a fuzzy match level;

outputting the plurality of identified source material segments and respective similar stored material segments in a form suitable for review by a user;

replacing identified source material segments with respective similar stored material segments to assist full translation of source material from said first natural language to at least said second natural language, wherein the replacing is carried out in response to input from a user and wherein the user chooses to amend the source material they are currently authoring according to stored material; and

translating the authored source material using machine translation to;

reuse similar stored material segments replaced from translation memory and respective stored target segments stored in translation memory, and translate the source material segments not replaced from translation memory.

2. The method according to claim 1, comprising conducting full translation of said source material from said first natural language to at least said second natural language with reference to said stored material segments which have previously been translated into at least said second natural language.

3. The method according to claim 2, wherein said step of conducting full translation of said source material from said first natural language to at least said second natural language comprises a human translator translating parts of said source material which were not identified as being similar to stored material segments using the fuzzy match level.

4. The method according to claim 2, wherein said step of conducting full translation of said source material from said first natural language to at least said second natural language comprises a machine translation process translating parts of said source material which were not identified as being similar to stored material segments using the fuzzy match level.

5. The method according to claim 2, wherein said step of conducting full translation of said source material from said first natural language to at least said second natural language comprises a human translator and a machine translation process translating parts of said source material which were not identified as being similar to stored material segments using the fuzzy match level.

6. The method according to claim 2, comprising outputting said full translation of said source material in a form suitable for review by a user.

7. The method according to claim 6, wherein parts of said source material which were translated with reference to said stored material segments which had previously been translated into at least said second natural language, parts of said source material which were translated by a human translator, and parts of said source material which were translated by a machine translation process are output in different forms.

8. The method according to claim 7, wherein said different forms comprise different colors.

9. The method according to claim 1, wherein said stored material segments comprise at least one translation memory.

10. The method according to claim 9, wherein said at least one translation memory contains a plurality of stored segment pairs, each of said stored segment pairs comprising source material in said first natural language and a corresponding translation in at least said second natural language.

11. The method according to claim 10, wherein each of said stored segments pairs corresponds to at least one of:

a paragraph,

a sentence, and

a phrase.

12. The method according to claim 1, further comprising dividing said source material segments into a plurality of source segments prior to said comparison of said source material segments with said stored material segments.

13. The method according to claim 12, wherein said comparing comprises comparing said plurality of source segments with said stored material segments in said first natural language from a plurality of stored segments.

14. The method according to claim 12, wherein said identified source material segments comprise one or more of said plurality of source segments.

15. The method according to claim 12, wherein said source material segments are divided into said plurality of source segments where at least one of:

a full stop,

an exclamation mark,

a question mark,

a colon,

a semicolon,

a tab character, or

a paragraph mark,

appears in said source material segments.

16. The method according to claim 1, wherein said method comprises comparing said source material segments with preferred terminology material.

17. The method according to claim 16, wherein said replacing of identified source material segments with respective similar stored material segments comprises replacing a part of said source material segments with a part of said preferred terminology material.

18. The method according to claim 1, wherein said method comprises comparing said source material segments with forbidden terminology material.

19. The method according to claim 1, wherein stylistic options are in said stored material segments.

20. The method according to claim 19, wherein said replacement comprises replacing a part of said source material segments according to at least one option from said stylistic options.

21. The method according to claim 19, wherein said stylistic options comprise at least one of:

clarity and conciseness,

length of a sentence,

a commonly misused word,

unnecessary wording, and

a combination of specific characters.

22. The method according to claim 19, wherein said stylistic options are user configurable.

23. The method according to claim 1, wherein said identifying of the plurality of source material segments that are similar to stored material segments using the fuzzy match level further comprises a correlation between said identified plurality of source material segments and said identified stored material segments.

24. The method according to claim 23, wherein said correlation comprises a correlation of at least a predetermined level.

25. The method according to claim 23, wherein said predetermined level is user configurable.

26. The method according to claim 1, wherein said outputting further comprises outputting data associated with said identifying of the plurality of source material segments that are similar to stored material segments using the fuzzy match level.

27. The method according to claim 1, comprising said user inputting said source material into a first software process.

28. The method according to claim 27, wherein said outputting and said replacing steps are carried out by said first software process.

29. The method according to claim 27, wherein said first software process comprises word processing software.

30. The method according to claim 27, wherein said comparing and said identifying steps are carried out by a second software process.

31. The method according to claim 30, wherein said first software process interfaces with said second software process via an Application Program Interface (API).

32. The method according to claim 30, wherein said stored material segments are accessed by said second software process.

33. A computer program product comprising a non-transitory computer-readable medium having computer readable instructions recorded thereon for natural language translation, the computer readable instructions being operative, when performed by a computerized device, to cause the computerized device to perform a method comprising:

receiving source material segments being authored by a user in a first natural language;

comparing the received source material segments with stored material segments in the first natural language, said stored material segments having previously been translated from said first natural language to target segments in a second natural language;

identifying a plurality of source material segments that are similar to stored material segments using a fuzzy match level;

outputting the plurality of identified source material segments and respective similar stored material segments in a form suitable for review by a user;

replacing identified source material segments with respective similar stored material segments to assist full translation of source material from said first natural language to at least said second natural language, wherein the replacing is carried out in response to input from a user and wherein the user chooses to amend the source material they are currently authoring according to stored material; and

translating the authored source material using machine translation to;

reuse similar stored material segments replaced from translation memory and respective stored target segments stored in translation memory, and

translate the source material segments not replaced from translation memory.

34. A system for redirecting translation that will be used for portions of source material to using machine translation including translation memory to facilitate natural language translation, the system comprising:

an authoring environment configured to receive source material being authored by a user in a first natural language;

a core component configured to;

receive the source material from the authoring environment,

compare segments of the received source material with stored material in the first natural language, said stored material having previously been translated from said first natural language to at least a second natural language and stored in a translation memory as segment pairs, each including a source segment and a translation segment,

identify a source material segment which is similar to a source segment of a segment pair in the translation memory,

compare the source material with stylistic options, and

output the identified source material segment and the identified similar source segment of the source segment pair to the authoring environment;

the authoring environment further configured to;

present the identified source material segment and the identified similar source segment of the source segment pair to the user,

replace the identified source material segment with the identified similar source segment of the source segment pair in response to input from the user to amend the source material the user is currently authoring, and to redirect translation of the source segment to be translated by machine translation using the translation memory during translation of the source material; and

a machine translator to perform machine translation of the source material, the machine translator configured to reuse portions of the source material replaced from translation memory and translate portions of the source material not replaced from translation memory.