In-context exact (ICE) matching

• US 10,248,650 B2
• Filed: 05/09/2016
• Issued: 04/02/2019
• Est. Priority Date: 03/05/2004
• Status: Active Grant

• Current Assignee: SDL PLC (RWS Holdings Plc)
• Original Assignee: SDL PLC (RWS Holdings Plc)
• Inventors: Ross, Russell G., Gillespie, Kevin
• Primary Examiner(s): Colucci, Michael C
• Application Number:
  US15/150,377
• Publication Number:
  US 20160253319A1
• Time in Patent Office: 1,058 Days
• Field of Search: 704 9, 704260, 704251, 704 2, 704235, 704270, 704277, 704 4, 704 5, 341107, 715708, 715532, 715268, 707 3, 707797, 707 5
• CPC Class Codes: G06F 16/90344 by using string matching techniques
• CPC Class Codes: G06F 40/47 Machine-assisted translation, e.g. using translation memory

Abstract

Methods, systems and program product are disclosed for determining matching level of a text lookup segment with a plurality of source texts in a translation memory in terms of context. The invention determines exact matches for the lookup segment in the plurality of source texts, and determines, in the case that at least one exact match is determined, that a respective exact match is an in-context exact match for the lookup segment in the case that a context of the lookup segment matches that of the respective exact match. Degree of context matching required can be predetermined, and results prioritized. The invention also includes methods, systems and program products for storing a translation pair of source text and target text in a translation memory including context, and the translation memory so formed. The invention ensures that content is translated the same as previously translated content and reduces translator intervention.

Claims

1. A method for determining an in-context exact (ICE) match from context matching levels of a plurality of source texts stored in a translation memory to a lookup segment to be translated, the method comprising:

generating a preceding usage context hash code for a preceding segment, based on a text stream for the preceding segment next to the lookup segment;

generating a post usage context hash code for a post segment, based on a text stream for the post segment next to the lookup segment;

determining any exact matches for the lookup segment in the plurality of translation memory source texts;

calculating for each exact match a context matching level based on;

a match between the preceding usage context hash code for the lookup segment and a preceding usage context hash code generated for a segment of a translation memory source text, and

a match between the post usage context hash code for the lookup segment and a post usage context hash code generated for the segment of the translation memory source text; and

determining, for each exact match, if the segment of the translation memory source text providing the exact match is an ICE match for the lookup segment based on the calculated context matching level.

2. The method of claim 1, wherein the plurality of translation memory source texts providing the exact match includes at least two calculated context matching levels for the generated hash codes.

3. The method of claim 2, wherein the ICE match determining step indicates that a respective exact match is an ICE match for the lookup segment only in the case that each of the generated preceding and post usage context hash codes of the lookup segment matches that of the generated preceding and post usage context hash codes for the respective exact match of the translation memory.

4. The method of claim 3, wherein, in the case that greater than one ICE match is determined, the ICE match determining step includes prioritizing each ICE match according to a degree of context matching.

5. The method of claim 1, wherein the ICE match determining step indicates that a respective exact match is an ICE match for the lookup segment only in the case that both a preceding usage context hash code level and a post usage context hash code level of the lookup segment matches that of the respective exact match.

6. The method of claim 1, further comprising, in the case that greater than one ICE match is determined, prioritizing each ICE match according to a degree of context hash code matching.

7. The method of claim 6, wherein the prioritizing step includes:

first preferring an ICE match having both previous and post UC (usage context) levels that match those of the lookup segment over an ICE match having only one of the previous and post UC levels matching those of the lookup segment;

where the first preferring step is non-conclusive, second preferring an ICE match from a same asset as the lookup segment over an ICE match from a different asset; and

where the second preferring step is non-conclusive, third preferring an ICE match with a closest position to a position of the lookup segment within the asset.

8. The method of claim 7, further comprising allowing a user to select the ICE match based on the prioritization.

9. The method of claim 1, wherein the lookup segment includes a plurality of lookup segments that are substantially identical in terms of content, and wherein the ICE match determining step includes determining an ICE match for each lookup segment.

10. The method of claim 9, wherein at least one lookup segment has a different ICE match than at least one other lookup segment.

11. The method of claim 1, further comprising the step of allowing retrieval of at least one source text based on the context matching level.

12. A system for determining an in-context exact (ICE) match from context matching levels of a plurality of source texts stored in a translation memory to a lookup segment to be translated, the system comprising:

means for generating a preceding usage context hash code for a preceding segment, based on a text stream for the preceding segment next to the lookup segment;

means for generating a post usage context hash code for a post segment, based on a text stream for the post segment next to the lookup segment;

means for determining any exact matches for the lookup segment in the plurality of translation memory source texts;

means for calculating for each exact match a context matching level based on;

a match between the preceding usage context hash code for the lookup segment and a preceding usage context hash code generated for a segment of a translation memory source text, and

a match between the post usage context hash code for the lookup segment and a post usage context hash code generated for the segment of the translation memory source text; and

means for determining, for each exact match if the segment of the translation memory source text providing the exact match is an ICE match for the lookup segment based on the calculated context matching level.

13. The system of claim 12, wherein the ICE match determining means indicates that a respective exact match is an ICE match for the lookup segment only in the case that each of the calculated preceding and post usage context hash codes of the lookup segment matches that of the respective hash codes of the translation memory.

14. The system of claim 13, further comprising means for, in the case that greater than one exact match is determined, ranking each exact match according to a degree of context matching.

15. The system of claim 12, wherein the ICE match determining means indicates that a respective exact match is an ICE match for the lookup segment only in the case that both the preceding usage context level and the post usage context level of the lookup segment matches that of the respective exact match.

16. The system of claim 12, further comprising means for, in the case that greater than one ICE match is determined, prioritizing each ICE match according to a degree of context hash code matching.

17. The system of claim 16, wherein the prioritizing means:

first prefers an ICE match having both previous and post UC (usage context) levels that match those of the lookup segment over an ICE match having only one of the previous and post UC levels matching those of the lookup segment;

where the first preference is non-conclusive, second prefers an ICE match from a same asset as the lookup segment over an ICE match from a different asset; and

where the second preference is non-conclusive, third prefers an ICE match with a closest position to a position of the lookup segment within the asset.

18. The system of claim 17, further comprising means for allowing a user to select the ICE match based on the rank.

19. The system of claim 18, wherein the lookup segment includes a plurality of lookup segments that are substantially identical in terms of content, and wherein the ICE match determining means determines an ICE match for each lookup segment.

20. The system of claim 19, wherein at least one lookup segment has a different ICE match than at least one other lookup segment.

21. The system of claim 12, further comprising means for allowing retrieval of at least one source text based on the context matching level.

22. A program product stored on a non-transitory computer readable medium for determining an in-context exact (ICE) match from context matching levels of a plurality of translation memory source texts stored in a translation memory to a lookup segment to be translated, the computer readable medium comprising program code for performing the following steps:

generating a preceding usage context hash code for a preceding segment, based on a text stream for the preceding segment next to the lookup segment;

generating a post usage context hash code for a post segment, based on a text stream for the post segment next to the lookup segment;

determining any exact matches for the lookup segment in the plurality of translation memory source texts;

calculating for each exact match a context matching level based on;

a match between the preceding usage context hash code for the lookup segment and a preceding usage context hash code generated for a segment of a translation memory source text, and

a match between the post usage context hash code for the lookup segment and a post usage context hash code generated for the segment of the translation memory source text; and

determining, for each exact match, if the segment of the translation memory source text providing the exact match is an ICE match for the lookup segment based on the calculated context matching level.

23. The program product of claim 22, wherein the plurality of translation memory source texts providing the exact match includes at least two calculated context matching levels for the generated hash codes.

24. The program product of claim 23, wherein the ICE match determining step indicates if each exact match is an ICE match for the lookup segment only in the case that both the calculated preceding and calculated post context matching level of the lookup segment matches that of the respective exact match.

25. The program product of claim 24, further comprising, in the case that greater than one exact match is determined, ranking each exact match according to a degree of context matching.

26. The program product of claim 22, wherein the ICE match determining step indicates that a respective exact match is an ICE match for the lookup segment only in the case that both the preceding usage context hash code and the post usage context hash code of the lookup segment matches that of the respective exact match.

27. The program product of claim 23, further comprising, in the case that greater than one exact match is determined, ranking each exact match according to a degree of context hash code matching.

28. The program product of claim 26, wherein the prioritizing step includes:

first preferring an ICE match having both previous and post UC (usage context) hash codes that match those of the lookup segment over an ICE match having only one of the previous and post UC hash codes matching those of the lookup segment;

where the first preferring step is non-conclusive, second preferring an ICE match from a same asset as the lookup segment over an ICE match from a different asset; and

where the second preferring step is non-conclusive, third preferring an ICE match with a closest position to a position of the lookup segment within the asset.

29. The program product of claim 28, further comprising allowing a user to select the ICE match based on the prioritization.

30. The program product of claim 22, wherein the lookup segment includes a plurality of lookup segments that are substantially identical in terms of content, and wherein the ICE match determining step includes determining an ICE match for each lookup segment.

31. The program product of claim 30, wherein at least one lookup segment has a different ICE match than at least one other lookup segment.

32. The program product of claim 22, further comprising the step of allowing retrieval of at least one source text based on the context matching level.