Joint embedding of corpus pairs for domain mapping

• US 10,579,940 B2
• Filed: 08/18/2016
• Issued: 03/03/2020
• Est. Priority Date: 08/18/2016
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Jagmohan, Ashish, Khabiri, Elham, Segal, Richard B., Vaculin, Roman
• Primary Examiner(s): Phan, Tuankhanh D
• Application Number:
  US15/240,649
• Publication Number:
  US 20180052849A1
• Time in Patent Office: 1,293 Days
• Field of Search: None
• CPC Class Codes: G06F 16/36 Creation of semantic tools, e.g. ontology or thesauri
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/08 Learning methods

Abstract

Techniques for outside-in mapping for corpus pairs are provided. In one example, a computer-implemented method comprises: inputting first keywords associated with a first domain corpus; extracting a first keyword of the first keywords; inputting second keywords associated with a second domain corpus; generating an embedded representation of the first keyword via a trained model and generating an embedded representation of the second keywords via the trained model; and scoring a joint embedding affinity associated with a joint embedding. The scoring the joint embedding affinity comprises: transforming the embedded representation of the first keyword and the embedded representation of the second keywords via the trained model; determining an affinity value based on comparing the first keyword to the second keywords; and based on the affinity value, aggregating the joint embedding of the embedded representation of the first keyword and the embedded representation of the second keywords within the second domain corpus.

Claims

1. A system, comprising:

a memory that stores computer executable components; and

a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise;

an extraction component that;

extracts a first keyword of a plurality of first keywords associated with a first corpus;

a computation component that;

generates a first embedded representation of the first keyword via a trained model and generates a second embedded representation of second keywords via the trained model, wherein the second keywords are associated with a second corpus; and

a scoring component that;

scores a joint embedding affinity associated with a joint embedding of the first embedded representation of the first keyword and the second embedded representation of the second keywords within the second domain corpus, resulting in a scoring of the joint embedding affinity, wherein the scoring of the joint embedding affinity comprises;

transforming the first embedded representation of the first keyword and the second embedded representation of the second keywords via the trained model;

determining an affinity value based on comparing the first keyword to the second keywords; and

based on the affinity value, aggregating the joint embedding of the first embedded representation of the first keyword and the second embedded representation of the second keywords within the second domain corpus.

2. The system of claim 1, wherein the plurality of first keywords comprises a second keyword associated with a user identity associated with a user profile and the second keywords comprise a plurality of user identities associated with a plurality of user profiles.

3. The system of claim 1, wherein the affinity value is generated in response to an unsupervised estimation, and wherein the unsupervised estimation comprises a cluster analysis of the plurality of first keywords and the second keywords.

4. The system of claim 1, wherein the plurality of first keywords comprises a second keyword associated with a plurality of web pages.

5. The system of claim 1, wherein the affinity value is related to a scoring of the joint embedding of the first embedded representation and the second embedded representation.

6. The system of claim 1, wherein the scoring the joint embedding affinity comprises, based on satisfaction of a defined condition associated with a frequency of the first keyword, outputting a keyword collection ranking, resulting in a ranked first keyword, to the second domain corpus.

7. A computer program product for managing a mapping process, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable to:

input a plurality of first keywords associated with a first domain corpus;

extract a first keyword of the plurality of first keywords;

input a plurality of second keywords associated with a second domain corpus;

generate a first embedded representation of the first keyword via a trained model and generate a second embedded representation of the second keywords via the trained model; and

score a joint embedding affinity associated with a joint embedding, wherein the scoring the joint embedding affinity comprises;

transforming the first embedded representation of the first keyword and the second embedded representation of the second keywords via the trained model;

determining an affinity value based on comparing the first keyword to the second keywords; and

based on the affinity value, aggregating the joint embedding of the first embedded representation of the first keyword and the second embedded representation of the second keywords within the second domain corpus.

8. The computer program product of claim 7, wherein the affinity value is related to a scoring of the joint embedding of the first embedded representation and the second embedded representation.

9. The computer program product of claim 8, wherein the program instructions are further executable to:

based on satisfaction of a defined condition associated with a frequency of the first keyword, output a keyword collection ranking, resulting in a ranked first keyword, to the second domain corpus.

10. The computer program product of claim 9, wherein the keyword collection ranking is ranked in a descending order.

11. The computer program product of claim 9, wherein the keyword collection ranking comprises taxonomy tag data associated with a taxonomy tag.

12. The computer program product of claim 9, wherein the program instructions are further executable to:

output the keyword collection ranking to the first domain corpus.

13. The computer program product of claim 7, wherein the plurality of first keywords comprises a second keyword associated with a user identity associated with a user profile and the plurality of second keywords comprises a plurality of user identities associated with a plurality of user profiles.

14. The computer program product of claim 7, wherein the affinity value is generated in response to an unsupervised estimation, and wherein the unsupervised estimation comprises a cluster analysis of the plurality of first keywords and the plurality of second keywords.

15. The computer program product of claim 7, wherein the plurality of first keywords comprises a second keyword associated with a plurality of web pages.

16. A computer-implemented method, comprising:

inputting, by a device operatively coupled to a processor, a plurality of first keywords associated with a first domain corpus;

extracting, by the device, a first keyword of the plurality of first keywords;

inputting, by the device, a plurality of second keywords associated with a second domain corpus;

generating, by the device, a first embedded representation of the first keyword via a trained model and generating a second embedded representation of the second keywords via the trained model; and

scoring, by the device, a joint embedding affinity associated with a joint embedding, wherein the scoring the joint embedding affinity comprises;

transforming the first embedded representation of the first keyword and the second embedded representation of the second keywords via the trained model;

determining an affinity value based on comparing the first keyword to the second keywords; and

based on the affinity value, aggregating the joint embedding of the first embedded representation of the first keyword and the second embedded representation of the second keywords within the second domain corpus.

17. The computer-implemented method of claim 16, wherein the affinity value is related to a scoring of the joint embedding of the first embedded representation and the second embedded representation.

18. The computer-implemented method of claim 17, further comprising:

based on satisfaction of a defined condition associated with a frequency of the first keyword, outputting, by the device, a keyword collection ranking, resulting in a ranked first keyword, to the second domain corpus.

19. The computer-implemented method of claim 16, wherein a keyword collection ranking comprises taxonomy tag data associated with a taxonomy tag.

20. The computer-implemented method of claim 16, further comprising:

generating, by the device, a stochastic gradient descent value related to the trained model.