Vector Space Method for Secure Information Sharing

• US 20090328226A1
• Filed: 06/05/2008
• Published: 12/31/2009
• Est. Priority Date: 01/07/2003
• Status: Active Grant

• Current Assignee: Relativity ODA LLC
• Original Assignee: Content Analyst Company, LLC (Relativity ODA LLC)
• Inventors: BRADFORD, Roger
• Granted Patent: US 8,024,344 B2
• Time in Patent Office: Days
• US Class Current: 726/26
• CPC Class Codes: G06F 21/6254 by anonymising data, e.g. decorrelating personal data from the owner's identification
• CPC Class Codes: G06Q 10/10 Office automation; Time management

Abstract

Presented are systems and methods for securely sharing confidential information. In such a method, term vectors corresponding to ones of a plurality of confidential terms included in a plurality of confidential documents is received. Each of the received term vectors is mapped into a vector space. Non-confidential documents are mapped into the vector space to generate a document vector corresponding to each non-confidential document, wherein the generation of each document vector is based on a subset of the received term vectors. At least one of the non-confidential documents is identified in response to a query mapped into the vector space.

Claims

1. A computer-based method for securely sharing confidential information, comprising:

(a) receiving term vectors corresponding to ones of a plurality of confidential terms included in a plurality of confidential documents;

(b) mapping each of the received term vectors into a vector space;

(c) mapping non-confidential documents into the vector space to generate a document vector corresponding to each non-confidential document, wherein the generation of each document vector is based on a subset of the received term vectors; and

(d) identifying at least one of the non-confidential documents in response to a query mapped into the vector space.

2. The method of claim 1, further comprising:

identifying at least one term contained in a subset of the non-confidential document, but not included in the plurality of confidential terms; and

mapping the at least one term into a corresponding at least one term vector in the vector space.

3. The method of claim 2, wherein mapping the at least one term into a corresponding at least one term vector in the vector space comprises:

generating the at least one term vector based on an average of a subset of document vectors corresponding to the subset of non-confidential documents that contain the at least one term.

4. A computer-based method for securely sharing confidential documents, wherein a non-confidential vector space is embodied in a computer readable medium, and the non-confidential vector space includes non-confidential term vectors and non-confidential document vectors corresponding to a collection of non-confidential documents, the computer-based method comprising:

(a) transmitting a subset of the non-confidential term vectors to a plurality of entities that each manage a confidential document database, each non-confidential term vector corresponding to a non-confidential term;

(b) receiving offset vectors from the plurality of entities responsive to confidential documents in the confidential document databases that contain the non-confidential terms;

(c) updating the term vectors in the non-confidential vector space responsive to the offset vectors;

(d) generating a vector representation of a query in the non-confidential vector space based on the updated term vectors; and

(e) identifying at least one non-confidential document based on a comparison between the vector representation of the query and the non-confidential document vectors.

5. The computer-based method of claim 4, further comprising:

transmitting a number of times each non-confidential term occurs in the collection of non-confidential documents.

6. The computer-based method of claim 4, wherein step (c) comprises:

updating the term vectors in the non-confidential vector space responsive to the offset vectors, such that a first term vector is updated based on a combination of offset vectors respectively received from each entity.

7. The computer-based method of claim 4, wherein step (c) comprises:

updating the term vectors in the non-confidential vector space responsive to the offset vectors, such that a new term vector is generated for each offset vector.

8. The computer-based method of claim 4, wherein step (b) further comprises:

receiving new term vectors from the plurality of entities.

9. The computer-based method of claim 8, wherein step (c) comprises:

updating the term vectors in the non-confidential vector space responsive to the offset vectors and the new term vectors.

10. A computer program product comprising a computer readable storage medium having computer readable code stored therein for securely sharing confidential information, the computer readable program code comprising:

computer readable first program code for causing the computer to receive term vectors corresponding to ones of a plurality of confidential terms included in a plurality of confidential documents;

computer readable second program code for causing the computer to map each of the received term vectors into a vector space;

computer readable third program code for causing the computer to map non-confidential documents into the vector space to generate a document vector corresponding to each non-confidential document, wherein the generation of each document vector is based on a subset of the received term vectors; and

computer readable fourth program code for causing the computer to identify at least one of the non-confidential documents in response to a query mapped into the vector space.

11. The computer program product of claim 10, further comprising:

computer readable fifth program code for causing the computer to identify at least one term contained in a subset of the non-confidential document, but not included in the plurality of confidential terms; and

computer readable sixth program code for causing the computer to map the at least one term into a corresponding at least one term vector in the vector space.

12. The computer program product of claim 11, wherein the computer readable sixth program code comprises:

code for generating the at least one term vector based on an average of a subset of document vectors corresponding to the subset of non-confidential documents that contain the at least one term.

13. A computer program product comprising a computer readable storage medium having computer readable code stored therein for securely sharing confidential information, wherein a non-confidential vector space is embodied in the computer readable medium, and the non-confidential vector space includes non-confidential term vectors and non-confidential document vectors corresponding to a collection of non-confidential documents, the computer readable program code comprising:

computer readable first program code for causing the computer to transmit a subset of the non-confidential term vectors to a plurality of entities that each manage a confidential document database, each non-confidential term vector corresponding to a non-confidential term;

computer readable second program code for causing the computer to receive offset vectors from the plurality of entities responsive to confidential documents in the confidential document databases that contain the non-confidential terms;

computer readable third program code for causing the computer to update the term vectors in the non-confidential vector space responsive to the offset vectors;

computer readable fourth program code for causing the computer to generate a vector representation of a query in the non-confidential vector space based on the updated term vectors; and

computer readable fifth program code for causing the computer to identify at least one non-confidential document based on a comparison between the vector representation of the query and the non-confidential document vectors.

14. The computer program product of claim 13, further comprising:

computer readable sixth program code for causing the computer to transmit a number of times each non-confidential term occurs in the collection of non-confidential documents.

15. The computer program product of claim 13, wherein the computer readable third program code comprises:

code for causing the computer to update the term vectors in the non-confidential vector space responsive to the offset vectors, such that a first term vector is updated based on a combination of offset vectors respectively received from each of the entities.

16. The computer program product of claim 13, wherein the computer readable third program code comprises:

code for causing the computer to update the term vectors in the non-confidential vector space responsive to the offset vectors, such that a new term vector is generated for each offset vector.

17. The computer program product of claim 13, wherein the computer readable second program code further comprises:

code for causing the computer to receive new term vectors from the plurality of entities.

18. The computer program product of claim 17, wherein the computer readable third program code comprises:

code for causing the computer to update the term vectors in the non-confidential vector space responsive to the offset vectors and the new term vectors.