Balancing the costs of sharing private data with the utility of enhanced personalization of online services

• US 8,346,749 B2
• Filed: 06/27/2008
• Issued: 01/01/2013
• Est. Priority Date: 06/27/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Krause, Ranier Andreas, Horvitz, Eric
• Primary Examiner(s): CHANNAVAJJALA, SRIRAMA T
• Application Number:
  US12/147,557
• Publication Number:
  US 20090327228A1
• Time in Patent Office: 1,649 Days
• Field of Search: 707705-713, 707/732, 707783-785, 707/719, 707/767, 702181-182, 705/14.73, 705/10, 705 1- 2, 705 26- 27, 709/203, 709215-220, 715/867, 715/759
• US Class Current: 707/706
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

Described herein is a method that includes determining a utility to a user of sharing a set of attributes for use in connection with personalized searching or other online services. The method further includes determining a cost to the user of sharing the set of attributes. The method also includes outputting an indication of a net benefit to the user of sharing the set of attributes based at least in part upon the determined utility and the determined cost.

Claims

1. A method comprising:

providing a set of attributes to a search engine for use in connection with personalized searching, wherein the set of attributes comprises a plurality of subsets of attributes, and wherein each attribute is a user-specific variable or a query-specific variable;

determining, for each subset of attributes, a respective value to a user of sharing a respective subset of attributes for use in connection with personalized searching, the respective value being indicative of an improvement in search results when considering the respective subset of attributes relative to search results when failing to consider the respective subset of attributes;

determining, for the respective subset of attributes, a respective cost to the user of sharing the respective subset of attributes for use in connection with personalized searching, the respective cost based on a probability that the user is identifiable based upon the respective subset of attributes;

utilizing a processor to compute a respective net benefit to the user of sharing the respective subset of attributes, the respective net benefit to the user being computed by the processor based at least in part upon the respective value and the respective cost to the user of sharing the respective subset of attributes; and

selecting a subset of attributes from the plurality of subset of attributes for use by the search engine when performing the personalized searching based at least in part upon the respective net benefit computed for the subset of attributes.

2. The method of claim 1, wherein the respective cost is based at least in part upon an indication of which attributes the user is willing to share with a search engine when the search engine executes searches responsive to receipt of queries from the user.

3. The method of claim 1, wherein the respective value is based at least in part upon an expected click entropy reduction, wherein the click entropy reduction indicates an expected change in click entropy between 1) when the subset of attributes is not used when a search is performed;

and

     2) when the subset of attributes is used when the search is performed.

4. The method of claim 1, further comprising:

receiving a profile from the user that includes a subset of attributes in the set of attributes that the user is willing to share; and

using the profile to perform a search when the user provides a query to a search engine.

5. The method of claim 1, further comprising:

receiving a query; and

responsive to receiving the query, prompting the user for at least one attribute that is determined to increase the net benefit to the user of sharing the at least one attribute together with previously shared attributes.

6. The method of claim 1, further comprising, for a single search session, using a profile when performing a personalized search.

7. The method of claim 1, further comprising:

receiving the set of attributes; and

selecting the subset of attributes based upon the respective net benefit being optimal relative to net benefits of other considered subsets of attributes.

8. The method of claim 7, further comprising:

receiving data from the user for at least one attribute in the subset of attributes; and

performing a search using the at least one attribute.

9. The method of claim 1, wherein the cost is based at least in part upon an expected number of violations of k-anonymity corresponding to the respective subset of attributes, wherein k-anonymity indicates that the subset of attributes is matched by k people.

10. A system comprising:

a processor; and

a memory that comprises a plurality of components that are executed by the processor, the plurality of components comprising;

a value of information determiner component configured to determine a value of information to a user of sharing one or more attributes that correspond to the user with a search engine that performs searches, the value of information based upon an expected reduction in click entropy from a base entropy achieved by using the one or more attributes when performing searches;

a cost determiner component configured to determine a cost to the user of sharing the one or more attributes, the cost based upon a probability of identifying the user based on the one or more attributes and a sensitivity to the user of sharing the one or more attributes; and

an analyzer component that is configured to;

receive the value of information determined by the value of information determiner component and the cost determined by the cost determiner component; and

output an indication of a net benefit to the user of sharing the one or more attributes in connection with personalized searching based at least in part upon the received value of information and the received cost.

11. The system of claim 10, the plurality of components further comprising a selector component configured to select a set of attributes in the one or more attributes that optimizes the net benefit to the user with respect to sharing the set of attributes based at least in part upon the output indication.

12. The system of claim 11, the plurality of components further comprising an information requester component configured to select information from the user to optimize the net benefit to the user with respect to personalized searching.

13. The system of claim 10, the plurality of components further comprising a profile generator component that is configured to:

provide information to the user;

receive information from the user; and

generate a user profile that includes information corresponding to at least one attribute in the one or more attributes that the user has authorized for use during a personalized search.

14. The system of claim 13, the plurality of components further comprising a search component configured to use the user profile to perform the personalized search when a query is received from the user.

15. The system of claim 13, wherein the user profile is authorized for use during a single search session.

16. The system of claim 13, wherein the user profile is retained on a client.

17. The system of claim 10, the plurality of components further comprising:

a logger component configured to generate logs of queries of the user and web sites visited by the user; and

a purger component configured to automatically purge data from the logs generated by the logger component to maintain restrictions in a user profile.

18. The system of claim 10, wherein the value of information determiner component determines the value of information through utilization of the following algorithm:

19. A computer-readable medium comprising instructions that, when executed by a processor, cause the processor to perform a method, the method comprising:

receiving a set of attributes that are indicative of an identity of a user;

selecting from the set of attributes a subset of attributes employable to perform a personalized search, the subset of attributes providing an optimal net benefit to the user for performing the personalized search, the optimal net benefit being based upon

     1) a value of information to the user of sharing the subset of attributes; and

     2) a cost to the user of sharing the subset of attributes, the value of information based upon a reduction in click entropy from a base entropy achieved when using the subset of attributes, and the cost based upon a probability that the user is identifiable based upon the subset of attributes;

receiving a query from the user; and

performing the personalized search responsive to receipt of the query from the user, the subset of attributes being used in connection with ranking search results that are presented to the user.

20. The computer-readable medium of claim 19, the method further comprising:

responsive to the cost to the user being above a user-defined threshold, automatically purging data associated with a user profile at a data store until a probability that the user is identifiable based upon the subset of attributes drops below a user-defined threshold.