Method and system for transforming a source language linguistic structure into a target language linguistic structure based on example linguistic feature structures
First Claim
Patent Images
1. A computerized method of transforming an input source language linguistic feature structure (SLS) representing an expression in the source language to a target language linguistic feature structure (TLS) representing the expression in the target language, comprising:
- matching the SLS to at least one existing source language example linguistic feature structure (SEF);
aligning the SLS with the at least one SEF; and
generating the TLS using at least one target language example linguistic feature structure (TEF) corresponding to the at least one SEF.
1 Assignment
0 Petitions
Accused Products
Abstract
A system and method for transforming input source language linguistic structures (SLS) to target linguistic language linguistic structures (TLS) are described. In one embodiment, the SLS is matched to at least one existing source language example linguistic structure (SEF). The SLS is aligned with the at least one SEF, and the TLS is generated using at least one target language example linguistic structure (TEF) corresponding to the at least one SEF.
192 Citations
42 Claims
-
1. A computerized method of transforming an input source language linguistic feature structure (SLS) representing an expression in the source language to a target language linguistic feature structure (TLS) representing the expression in the target language, comprising:
-
matching the SLS to at least one existing source language example linguistic feature structure (SEF);
aligning the SLS with the at least one SEF; and
generating the TLS using at least one target language example linguistic feature structure (TEF) corresponding to the at least one SEF. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14)
applying transfer grammar rules to the SLS to create at least one SLS sub-structure.
-
-
3. The method of claim 2 wherein the transfer grammar rules are recursively applied to the SLS sub-structure from a top-most transfer rule until all SLS sub-structures within SLS are transferred to corresponding TEF sub-structures contained within the TEF.
-
4. The method of claim 2 further comprising:
compiling the transfer grammar rules to generate a transfer grammar.
-
5. The method of claim 2 wherein matching further comprises:
applying compiled context-free grammar rules to the at least one SLS sub-structure.
-
6. The method of claim 5 wherein the context-free grammar rules are applied recursively from specific SLS sub-structures to more general SLS sub-structures contained within the SLS.
-
7. The method of claim 2 wherein generating the TLS is applied recursively from larger SLS sub-structures to smaller SLS sub-structures.
-
8. The method of claim 2 further comprising:
combining at least two SLS sub-structures to calculate a match cost with the SEF sub-structure.
-
9. The method of claim 2 wherein aligning further comprises:
-
if the at least one SLS sub-structure is not aligned, performing said matching, aligning, and generating using the unaligned at least one SLS sub-structure as input until a matched, aligned sub-structure is found to produce at least one sub-TLS, and inserting the at least one resulting sub-TLS into a corresponding TEF sub-structure.
-
-
10. The method of claim 2 further comprising:
-
if the at least one SEF sub-structure is indexed and if the at least one SEF sub-structure is not aligned, deleting the corresponding at least one TEF sub-structure if the at least one TEF sub-structure is not marked “
don'"'"'t delete”
.
-
-
11. The method of claim 2 further comprising:
-
if the at least one SEF sub-structure is indexed and if the at least one SEF sub-structure is aligned, if the at least one SEF sub-structure is not similar to the corresponding at least one SLS sub-structure, performing said matching, aligning, and generating using the at least one SEF sub-structure, producing at least one resulting sub-TLS, and replacing the corresponding at least one TEF sub-structure with the at least one resulting sub-TLS.
-
-
12. The method of claim 1 wherein the SEF and corresponding TEF are maintained in a bilingual example database.
-
13. The method of claim 1 wherein matching further comprises:
if an exact match is found, transferring the matching SEF to the corresponding TEF to produce the TLS.
-
14. The method of claim 1 wherein matching further comprises:
-
generating a match cost; and
accepting the matching SEF if the match cost is within pre-defined limits.
-
-
15. A system for transforming an input source language linguistic feature structure (SLS) representing an expression in the source language to a target language linguistic structure (TLS) representing the expression in the target language, comprising:
-
means for matching the SLS to at least one existing source language example linguistic feature structure (SEF);
means for aligning the SLS with the at least one SEF; and
means for generating the TLS using at least one target language example linguistic feature structure (TEF) corresponding to the at least one SEF.
-
-
16. A computer readable medium comprising instructions, which when executed on a processor, perform a method for transforming an input source language linguistic feature structure (SLS) representing an expression in the source language to a target language linguistic feature structure (TLS) representing the expression in the target language, comprising:
-
matching the SLS to at least one existing source language example linguistic feature structure (SEF);
aligning the SLS with the at least one SEF; and
generating the TLS using at least one target language example linguistic feature structure (TEF) corresponding to the at least one SEF. - View Dependent Claims (17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29)
compiling the transfer grammar rules to generate a transfer grammar.
-
-
20. The medium of claim 17 wherein matching further comprises:
applying compiled context-free grammar rules to the at least one SLS sub-structure.
-
21. The medium of claim 20 wherein the context-free grammar rules are applied recursively from specific SLS sub-structures to more general SLS sub-structures contained within the SLS.
-
22. The medium of claim 17 wherein generating the TLS is applied recursively from larger SLS sub-structures to smaller SLS sub-structures.
-
23. The medium of claim 17 further comprising:
combining at least two SLS sub-structures to calculate a match cost with the SEF sub-structure.
-
24. The medium of claim 17 wherein aligning further comprises:
-
if the at least one SLS sub-structure is not aligned, performing said matching, aligning, and generating using the unaligned at least one SLS sub-structure as input until a matched, aligned sub-structure is found to produce at least one sub-TLS, and inserting the at least one resulting sub-TLS into a corresponding TEF sub-structure.
-
-
25. The medium of claim 17 further comprising:
-
if the at least one SEF sub-structure is indexed and if the at least one SEF sub-structure is not aligned, deleting the corresponding at least one TEF sub-structure if the at least one TEF sub-structure is not marked “
don'"'"'t delete”
.
-
-
26. The medium of claim 17 further comprising:
-
if the at least one SEF sub-structure is indexed and if the at least one SEF sub-structure is aligned, if the at least one SEF sub-structure is not similar to the corresponding at least one SLS sub-structure, performing said matching, aligning, and generating using the at least one SEF sub-structure, to produce at least one resulting sub-TLS, and replacing the corresponding at least one TEF sub-structure with the at least one resulting sub-TLS.
-
-
27. The medium of claim 16 wherein the SEF and corresponding TEF are maintained in a bilingual example database.
-
28. The medium of claim 16 wherein matching further comprises:
if an exact match is found, transferring the matching SEF to the corresponding TEF to produce the TLS.
-
29. The medium of claim 16 wherein matching further comprises:
-
generating a match cost; and
accepting the matched SEF if the match cost is within pre-defined limits.
-
-
30. An apparatus for transforming an input source language linguistic feature structure (SLS) representing an expression in the source language to a target language linguistic feature structure (TLS) representing the expression in the target language, comprising:
-
a transfer module configured to match the SLS to at least one existing source language example linguistic feature structure (SEF), align the SLS with the at least one SEF, and generate the TLS using at least one target language example linguistic feature structure (TEF) corresponding to the at least one SEF. - View Dependent Claims (31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41)
apply transfer grammar rules to the SLS to create at least one SLS sub-structure.
-
-
32. The apparatus of claim 31 wherein the transfer grammar rules are recursively applied to the SLS sub-structure from a top-most transfer rule until all SLS sub-structures within SLS are transferred to corresponding TEF sub-structures.
-
33. The apparatus of claim 31 further comprising:
a compiler configured to compile the transfer grammar rules to generate a transfer grammar.
-
34. The apparatus of claim 31 wherein the transfer module is further configured to apply compiled context-free grammar rules to the at least one SLS sub-structure.
-
35. The apparatus of claim 34 wherein the context-free grammar rules are applied recursively from specific SLS sub-structures to more general SLS sub-structures contained within the SLS.
-
36. The apparatus of claim 31 wherein the generation of TLS is applied recursively from larger SLS sub-structures to smaller SLS sub-structures.
-
37. The apparatus of claim 31 wherein the transfer module is further configured to combine at least two SLS sub-structures to calculate a match cost with the SEF sub-structure.
-
38. The apparatus of claim 30 wherein the SEF and corresponding TEF are maintained in a bilingual example database.
-
39. The apparatus of claim 30 wherein the transfer module is further configured to:
transfer the matching SEF to the corresponding TEF to produce the TLS if an exact match is found.
-
40. The apparatus of claim 30 further comprising:
a thesaurus matching system configured to generate a match cost.
-
41. The apparatus of claim 40 wherein transfer module is further configured to accept the matching SEF if the match cost is within pre-defined limits.
-
42. A memory for storing data for access by an application program being executed on a data processing system, comprising:
-
a data structure stored in said memory, said data structure including information resident in a database used by said application program and including;
a plurality of bilingual example pairs used for the matching of data, wherein each bilingual example pair includes a source language example linguistic feature structure representing an expression in the source language, and a target language example linguistic feature structure corresponding to the source language example linguistic feature structure.
-
Specification