Computing the value of information between concepts

• US 10,372,745 B2
• Filed: 10/03/2016
• Issued: 08/06/2019
• Est. Priority Date: 10/03/2016
• Status: Active Grant

• Current Assignee: Kyndryl Incorporated
• Original Assignee: International Business Machines Corporation
• Inventors: Lorge, Freddy
• Primary Examiner(s): Cheema, Azam M
• Application Number:
  US15/283,806
• Publication Number:
  US 20180096062A1
• Time in Patent Office: 1,037 Days
• CPC Class Codes: G06F 16/367 Ontology

Abstract

A method, and associated computer system and computer program product, for computing a value between two concepts in a schema containing concepts which are linked to each other through associations. In a schema S of n concepts C1, C2 . . . Ci . . . Cn, the concepts are linked by associations, each association having a semantic distance set in a range between a minimum and a maximum indicating the concepts are completely similar or dissimilar respectively. An information value is determined between concepts from their semantic distance and informational distance. For dissimilar concepts, the informational distance is computed according to a closeness of the concepts. For similar concepts, the informational distance is computed according to a remoteness of the concepts. Both the first and second functions increase with a number of links between C1 and another concept. The number of links is a topological distance between C1 and the other concept.

Claims

1. A method for computing a value between two concepts in a schema containing a plurality of concepts which are linked to each other through associations, said method comprising:

receiving, by one or more processors of a computer system, a result from a search engine resulting from a search query submitted to the search engine, said result comprising a schema S containing n concepts C1, C2 . . . Ci . . . Cn linked by associations, each such association forming a link in a chain connecting any two concepts of the n concepts, each such association having at least one semantic distance set in a range lying between a minimum semantic distance which indicates the two concepts are completely similar and a maximum semantic distance which indicates the two concepts are completely dissimilar, so that a semantic distance can be determined between said any two concepts of the n concepts by summing the at least one semantic distance between the two concepts, wherein the number of links in the chain that connects the concept C1 with another concept of the n concepts is a topological distance;

computing, by the or more processors, an informational distance between the concept C1 and the concept Ci+1 (i=1, 2, . . . , n−

1);

computing, by the one or more processors for i=1, 2, . . . , n−

1, an information value between the concept C1 and the concept Ci+1, wherein the information value between the concept C1 and the concept Ci+1 depends on the semantic distance between the concept Ci and the concept Ci+1 and the informational distance between the concept C1 and the concept Ci;

if the concept Ci and the concept Ci+1 are separated by an association with a semantic distance above a threshold value, and are hence dissimilar concepts, then said one or more processors computing the informational distance between the concept C1 and the concept Ci according to a closeness using a first function that increases with topological distance between the concepts C1 and Ci and gives a lower informational distance with lower semantic distance between the concepts C1 and Ci;

if the concept Ci and the concept Ci+1 are not separated by the association with a semantic distance above the threshold value, and are hence similar concepts, then said one or more processors computing the informational distance between the concept C1 and the concept Ci according to a remoteness using a second function that increases with topological distance between the concepts C1 and Ci and gives a lower informational distance with higher semantic distance between the concepts C1 and Ci; and

outputting, by the one or more processors, the information value between the concept C1 and the concept Ci+1.

2. The method of claim 1, wherein the information value between the concept C1 and the concept Ci+1 is proportional to the semantic distance between the concept Ci and the concept Ci+1.

3. The method of claim 2, wherein the information value between the concept C1 and the concept Ci+1 is inversely proportional to a sum of a constant and the informational distance between the concept C1 and the concept Ci.

4. The method of claim 1, wherein the schema S is provided with certain associations being fixed as indicating complete similarity between the concepts the certain associations interconnect, such that if such an association exists between a concept i and a concept i+1, then the information value between concept C1 and concept Ci+1 is set to be equal to the information value between concept C1 and concept Ci.

5. The method of claim 1, wherein the threshold value is set to 30, 40, 50, 60 or 70 percent of the maximum semantic distance to within plus/minus 5 percent.

6. The method of claim 1, wherein the first function has informational distance values less than or equal to those of the second function for all semantic distance values.

7. The method of claim 1, wherein the first function provides a value that increases monotonically with semantic distance.

8. The method of claim 1, wherein the second function provides a value that decreases monotonically with semantic distance.

9. The method of claim 1, wherein the information value is proportional to at least one value selected from the group consisting of an average of the semantic distances, an inverse of the semantic distances between the concepts in a sub-chain Cx . . . Ci+1, and a combination thereof, wherein 1≤

x≤

i+1 and the length of the sub-chain Cx . . . Ci+1 is a pre-determined constant.

10. A computer program product, comprising one or more computer readable hardware storage devices having computer readable program code stored therein, said program code containing instructions executable by one or more processors of a computer system to implement a method for computing a value between two concepts in a schema containing a plurality of concepts which are linked to each other through associations, said method comprising:

receiving, by the one or more processors, a result from a search engine resulting from a search query submitted to the search engine, said result comprising a schema S containing n concepts C1, C2 . . . Ci . . . Cn linked by associations, each such association forming a link in a chain connecting any two concepts of the n concepts, each such association having at least one semantic distance set in a range lying between a minimum semantic distance which indicates the two concepts are completely similar and a maximum semantic distance which indicates the two concepts are completely dissimilar, so that a semantic distance can be determined between said any two concepts of the n concepts by summing the at least one semantic distance between the two concepts, wherein the number of links in the chain that connects the concept C1 with another concept of the n concepts is a topological distance;

computing, by the or more processors, an informational distance between the concept C1 and the concept Ci+1 (i=1, 2, . . . , n−

1);

computing, by the one or more processors for i=1, 2, . . . , n−

1, an information value between the concept C1 and the concept Ci+1, wherein the information value between the concept C1 and the concept Ci+1 depends on the semantic distance between the concept Ci and the concept Ci+1 and the informational distance between the concept C1 and the concept Ci;

if the concept Ci and the concept Ci+1 are separated by an association with a semantic distance above a threshold value, and are hence dissimilar concepts, then said one or more processors computing the informational distance between the concept C1 and the concept Ci according to a closeness using a first function that increases with topological distance between the concepts C1 and Ci and gives a lower informational distance with lower semantic distance between the concepts C1 and Ci;

if the concept Ci and the concept Ci+1 are not separated by the association with a semantic distance above the threshold value, and are hence similar concepts, then said one or more processors computing the informational distance between the concept C1 and the concept Ci according to a remoteness using a second function that increases with topological distance between the concepts C1 and Ci and gives a lower informational distance with higher semantic distance between the concepts C1 and Ci; and

outputting, by the one or more processors, the information value between the concept C1 and the concept Ci+1.

11. The computer program product of claim 10, wherein the information value between the concept C1 and the concept Ci+1 is proportional to the semantic distance between the concept Ci and the concept Ci+1.

12. The computer program product of claim 11, wherein the information value between the concept C1 and the concept Ci+1 is inversely proportional to a sum of a constant and the informational distance between the concept C1 and the concept Ci.

13. The computer program product of claim 10, wherein the schema S is provided with certain associations being fixed as indicating complete similarity between the concepts the certain associations interconnect, such that if such an association exists between a concept i and a concept i+1, then the information value between concept C1 and concept Ci+1 is set to be equal to the information value between concept C1 and concept Ci.

14. The computer program product of claim 10, wherein the threshold value is set to 30, 40, 50, 60 or 70 percent of the maximum semantic distance to within plus/minus 5 percent.

15. The computer program product of claim 10, wherein the first function has informational distance values less than or equal to those of the second function for all semantic distance values.

16. The computer program product of claim 10, wherein the first function provides a value that increases monotonically with semantic distance.

17. The computer program product of claim 10, wherein the second function provides a value that decreases monotonically with semantic distance.

18. The computer program product of claim 10, wherein the information value is proportional to at least one value selected from the group consisting of an average of the semantic distances, an inverse of the semantic distances between the concepts in a sub-chain Cx . . . Ci+1, and a combination thereof, wherein 1≤

x≤

i+1 and the length of the sub-chain Cx . . . Ci+1 is a pre-determined constant.

19. A computer system, comprising one or more processors, one or more memories, and one or more computer readable hardware storage devices, said one or more hardware storage device containing program code executable by the one or more processors via the one or more memories to implement a method for computing a value between two concepts in a schema containing a plurality of concepts which are linked to each other through associations, said method comprising:

receiving, by the one or more processors, a result from a search engine resulting from a search query submitted to the search engine, said result comprising a schema S containing n concepts C1, C2 . . . Ci . . . Cn linked by associations, each such association forming a link in a chain connecting any two concepts of the n concepts, each such association having at least one semantic distance set in a range lying between a minimum semantic distance which indicates the two concepts are completely similar and a maximum semantic distance which indicates the two concepts are completely dissimilar, so that a semantic distance can be determined between said any two concepts of the n concepts by summing the at least one semantic distance between the two concepts, wherein the number of links in the chain that connects the concept C1 with another concept of the n concepts is a topological distance;

computing, by the or more processors, an informational distance between the concept C1 and the concept Ci+1 (i=1, 2, . . . , n−

1);

computing, by the one or more processors for i=1, 2, . . . , n−

1, an information value between the concept C1 and the concept Ci+1, wherein the information value between the concept C1 and the concept Ci+1 depends on the semantic distance between the concept Ci and the concept Ci+1 and the informational distance between the concept C1 and the concept Ci;

if the concept Ci and the concept Ci+1 are separated by an association with a semantic distance above a threshold value, and are hence dissimilar concepts, then said one or more processors computing the informational distance between the concept C1 and the concept Ci according to a closeness using a first function that increases with topological distance between the concepts C1 and Ci and gives a lower informational distance with lower semantic distance between the concepts C1 and Ci;

if the concept Ci and the concept Ci+1 are not separated by the association with a semantic distance above the threshold value, and are hence similar concepts, then said one or more processors computing the informational distance between the concept C1 and the concept Ci according to a remoteness using a second function that increases with topological distance between the concepts C1 and Ci and gives a lower informational distance with higher semantic distance between the concepts C1 and Ci; and

outputting, by the one or more processors, the information value between the concept C1 and the concept Ci+1.

20. The computer system of claim 19, wherein the information value between the concept C1 and the concept Ci+1 is proportional to the semantic distance between the concept Ci and the concept Ci+1.