METHOD, PROGRAM AND SYSTEM FOR FINDING CORRESPONDENCE BETWEEN TERMS

• US 20120232884A1
• Filed: 03/07/2012
• Published: 09/13/2012
• Est. Priority Date: 03/08/2011
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Nasukawa, Tetsuya, Tsuboi, Yuta, Rudy, Raymond Harry, Zhang, Yao-zhong
• Granted Patent: US 8,682,641 B2
• Time in Patent Office: Days
• US Class Current: 704/8
• CPC Class Codes: G06F 40/42 Data-driven translation

Abstract

A computer implemented method, system, and product for finding correspondence between terms in two different languages. The method includes the steps of: creating a technical term set and a general term set for each of i) a first language and ii) a second language, creating two bipartite graphs, where each graph corresponds to one of the two languages, and connects the technical term set and general term set of each language, respectively, with weighted links based on corpus information, creating a third bipartite graph by creating weighted links between general terms in the first language and general terms in the second language by using a translation dictionary, creating an association matrix M corresponding to the three bipartite graphs, calculating a similarity matrix Q by calculation of an inverse matrix, and outputting correspondence between the technical term sets of the first and second language on basis of the similarity matrix.

Claims

1. A computer implemented method for finding correspondence between terms in two different languages, the method comprising the steps of:

creating a technical term set in a first language, a general term set in the first language, a technical term set in a second language and a general term set in the second language;

saving, in a storage unit of the computer, the term set of the first language, the general term set in the first language, the technical term set in the second language and the general term in the second language;

creating at least two bipartite graphs, wherein the first bipartite graph connects the technical term set and the general term set of the first language to each other with links on basis of corpus information, wherein the second bipartite graph connects the general term set and technical term set of the second language to each other with links the basis of corpus information, and wherein each of the links is weighted by a degree of association between terms;

creating a third bipartite graph by creating links between general terms in the first language and general terms in the second language by using a translation dictionary between general terms in the first language and general terms in the second language, each of the links being weighted by a degree of association between terms;

creating an association matrix M corresponding to the bipartite graphs between the technical term set and the general term set connected in each respective language and the bipartite graph between the general terms in the first language and the general terms in the second language;

calculating a similarity matrix Q by calculation of an inverse matrix of (I−

cM), where c is a positive number smaller than one; and

outputting correspondence between the technical term set in the first language and the technical term set in the second language on the basis of predetermined components of the similarity matrix Q,

wherein at least one step is executed on the computer.

2. The method according to claim 1, wherein the degree of association between the technical term set and the general term set corresponding to a same language in each of the bipartite graphs is either a co-occurrence frequency or a degree of co-occurrence correlation between terms.

3. The method according to claim 1, wherein the degree of association between the general terms in the first language and the general terms in the second language in the bipartite graph that links terms between the two languages is calculated on the basis of a listed sequence in the dictionary.

4. The method according to claim 1, wherein the association matrix M is normalized such that a sum of each row is equal to one.

5. The method according to claim 4, wherein the calculation of the similarity matrix Q includes calculating an inverse matrix of a matrix in which only a portion corresponding to both of the general term set in the first language and the technical term set in the first language and a portion corresponding to both of the general term set in the second language and the technical term set in the second language are left.

6. The method according to claim 1, wherein the c is a number from about 0.2 to about 0.3.

7. A computer readable non-transitory program product tangibly embodying computer readable instructions which, when executed, cause a computer to carry out steps of a method for finding correspondence between terms in two different languages, the method comprising the steps of:

creating a technical term set in a first language, a general term set in the first language, a technical term set in a second language and a general term set in the second language;

saving, in a storage unit of the computer, the term set of the first language, the general term set in the first language, the technical terms set in the second language and the general term in the second language;

creating at least two bipartite graphs, wherein the first bipartite graph connects the technical term set and the general term set of the first language to each other with links on the basis of corpus information, wherein the second bipartite graph connects the general term set and technical term set of the second language to each other with links on the basis of corpus information, and wherein each of the links is weighted by a degree of association between terms;

creating a third bipartite graph by creating links between general terms in the first language and general terms in the second language by using a translation dictionary between general terms in the first language and general terms in the second language, each of the links being weighted by a degree of association between terms;

creating an association matrix M corresponding to the bipartite graphs between the technical term set and the general term set connected in each respective language and the bipartite graph between the general terms in the first language and the general terms in the second language;

calculating a similarity matrix Q by calculation of an inverse matrix of (I−

cM) where c is a positive number smaller than one; and

outputting correspondence between the technical term set in the first language and the technical term set in the second language on the basis of predetermined components of the similarity matrix Q.

8. The program product according to claim 7, wherein the degree of association between the technical term set and the general term set corresponding to a same language in each of the bipartite graphs is either a co-occurrence frequency or a degree of co-occurrence correlation between terms.

9. The program product according to claim 7, wherein the degree of association between the general terms in the first language and the general terms in the second language in the bipartite graph that links terms between the two languages is calculated on the basis of a listed sequence in the dictionary.

10. The program product according to claim 7, wherein the association matrix M is normalized such that a sum of each row is equal to one.

11. The program product according to claim 10, wherein the calculation of the similarity matrix Q includes calculating an inverse matrix of a matrix in which only a portion corresponding to both of the general term set in the first language and the technical term set in the first language and a portion corresponding to both of the general term set in the second language and the technical term set in the second language are left.

12. The program product according to claim 7, wherein the c is a number from about 0.2 to about 0.3.

13. A computer system configured to find correspondence between terms in two different languages, the system comprising:

a unit configured to create a technical term set in a first language, a general term set in the first language, a technical term set in a second language and a general term set in the second language;

a storage unit for storing the term set of the first language, the general term set in the first language, the technical terms set in the second language and the general term in the second language;

a unit configured to create at least two bipartite graphs, wherein the first bipartite graph connects the technical term set and the general term set of the first language to each other with links on the basis of corpus information, wherein the second bipartite graph connects the general term set and technical term set of the second language to each other with links on the basis of corpus information, and wherein each of the links is weighted by a degree of association between terms;

a unit configured to create a third bipartite graph by creating links between general terms in the first language and general terms in the second language by using a translation dictionary between general terms in the first language and general terms in the second language, each of the links being weighted by a degree of association between terms;

a unit configured to create an association matrix M including the bipartite graphs between the technical term set and the general term set connected in each respective language and the bipartite graph between the general terms in the first language and the general terms in the second language;

a unit configured to calculate a similarity matrix Q by calculation of an inverse matrix of (I−

cM) where c is a positive number smaller than one; and

a unit configured to output correspondence between the technical term set in the first language and the technical term set in the second language on the basis of predetermined components of the similarity matrix Q.

14. The system according to claim 13, wherein the degree of association between the technical term set and the general term set corresponding to a same language in each of the bipartite graphs is either a co-occurrence frequency or a degree of co-occurrence correlation between terms.

15. The system according to claim 13, wherein the degree of association between the general terms in the first language and the general terms in the second language in the bipartite graph that links terms between the two languages is calculated on the basis of a listed sequence in the dictionary

16. The system according to claim 13, wherein the association matrix M is normalized such that a sum of each row is equal to one.

17. The system according to claim 16, wherein the calculation of the similarity matrix Q includes calculating an inverse matrix of a matrix in which only a portion corresponding to both of the general term set in the first language and the technical term set in the first language and a portion corresponding to both of the general term set in the second language and the technical term set in the second language are left.

18. The system according to claim 13, wherein the c is a number from about 0.2 to about 0.3.