Vector space method for secure information sharing

• US 20040133574A1
• Filed: 01/07/2003
• Published: 07/08/2004
• Est. Priority Date: 01/07/2003
• Status: Abandoned Application

• Current Assignee: Science Applications International Corporation
• Original Assignee: Science Applications International Corporation
• Inventors: Bradford, Roger
• Application Number:
  US10/337,426
• Publication Number:
  US 20040133574A1
• Time in Patent Office: Days
• US Class Current: 707/8
• 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

A system and method for exchanging secure information is described. Sensitive documents are reduced to term vectors and document vectors. These vectors are mapped in to a latent semantic indexing space. Queries may be performed on the LSI space to determine if terms and documents are related.

Claims

1. A method for sharing information items in a secure setting comprising:

reducing a document to a vector representation in vector space;

reducing terms in said document to vector representations in said vector space;

receiving a query in said vector space;

alerting an entity controlling said document to said query.

2. The method according to claim 1, further comprising the step of releasing said document to an originator of said query.

3. The method according to clam 1, further comprising the step of withholding said document from an originator of said query.

4. A method for sharing documents in a secure setting comprising:

reducing a document to a vector representation in vector space;

receiving a query in said vector space;

alerting an entity controlling said document to said query.

5. The method according to claim 4, further comprising the step of releasing said document to an originator of said query.

6. The method according to clam 4, further comprising the step of withholding said document from an originator of said query.

7. A method for sharing documents in a secure setting comprising:

reducing a document to a vector representation in a first vector space;

reducing terms in said document to vector representations in said first vector space;

creating a second vector space from said first vector space without said document vectors;

receiving a query in said second vector space;

alerting an entity controlling said document to said query.

8. The method according to claim 7, further comprising the step of releasing said document to an originator of said query.

9. The method according to clam 7, further comprising the step of withholding said document from an originator of said query.

10. A process for adding documents to a latent semantic indexing space comprising the steps of:

receiving a latent semantic indexing space including vector representation of terms;

folding in documents based on the relationship between terms in said latent semantic indexing space and terms in said documents.

11. A system for securely sharing information comprising:

a latent semantic indexing space having document vectors;

an input for receiving a query;

a processor for processing said query and determining whether any of said document vectors are within a threshold of said query;

an output for notifying an administrator of a document corresponding to said document vector contained within said threshold.

12. The system of claim 11, wherein said latent semantic indexing space further includes term vectors and other documents, whose presence in said threshold does not notify said administrator.

13. A system for securely sharing information comprising:

a first latent semantic indexing space having term vectors;

a second latent semantic indexing space having document vectors;

an input for receiving a query in said first latent semantic indexing space;

an output for forwarding said query to said second latent semantic indexing space;

a processor for determining whether any of said document vectors in said second latent semantic indexing space are within a threshold of said query;

an output for notifying an administrator of a document corresponding to said document vector contained within said threshold.

14. The system of claim 13, wherein said first latent semantic indexing space further includes documents, whose presence in said threshold does not notify said administrator.

15. A process for preventing reconstruction of documents in a latent semantic indexing space, comprising the steps of:

receiving a document vector;

dithering at least one dimension of said document vector.

16. A process for obscuring the content of a database comprising the steps of:

adding second document vectors to first document vectors representing confidential documents, in order to obscure the contents of the confidential documents.

17. The process according to claim 16, wherein said second document vectors represent non-confidential documents.

18. The process according to claim 16, wherein said second document vectors are generated at random.

19. The process according to claim 16, wherein said second document vectors are generated to prevent topical disclosure of said first document vectors.

20. The process according to claim 16, wherein said second document vectors are ignored by a party originating said second document vectors.

21. A process of retrieving ranked documents comprising the steps of:

receiving a vector space including term vectors relating to a first set of documents, said vector space not including said first set of documents or document vectors relating to said first set of documents;

adding a second set of documents to said vector space by adding document vectors related to said second set of documents to said vector space;

querying said vector space with one or more terms;

receiving selected ones of said second set of documents, said selected ones of said second set of documents being ranked according at least to relationships between said query of one or more terms and said term vectors relating to said first set of documents.

22. The process according to claim 21, wherein said querying step further comprises the step of:

determining a position of said query based on the relationship said one or more terms have with said term vectors from said first set of documents.

23. The process according to claim 21, wherein said adding said second set of documents further comprises the steps of:

determining term vectors from said second set of documents; and

adding said term vectors from said second set of documents to said term vectors from said first set of documents in said vector space.

24. The process according to claim 21, wherein said adding said second set of documents further comprises the steps of:

determining term vectors from said second set of documents; and

adding said document vectors from said second set of documents to said vector space without modifying said term vectors from said first set of documents.