Affinity Propagation in Adaptive Network-Based Systems

• US 20070203872A1
• Filed: 03/24/2007
• Published: 08/30/2007
• Est. Priority Date: 11/28/2003
• Status: Active Grant

• Current Assignee: Gula Consulting Limited Liability Company (Intellectual Ventures LLC)
• Original Assignee: Manyworlds Incorporated
• Inventors: Moneypenny, Naomi, Flinn, Steven
• Granted Patent: US 8,600,920 B2
• Time in Patent Office: Days
• US Class Current: 706/62
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/048 Fuzzy inferencing
• CPC Class Codes: G06N 7/02 using fuzzy logic computing arrangements based on biological models G06N3/00; computing arrangements using knowledge-based models G06N5/00

Abstract

Adaptive applications of affinity propagation are described to facilitate effective and computationally efficient means of clustering computer-based objects such as items of content, and/or to determine exemplars associated with a set of objects. Affinity propagation is also applied by the present invention to define system user affinity groups and/or exemplar users. The present invention applies usage behaviors as a basis for influencing clustering through methods such as initializing exemplar attractor values based on usage behaviors and/or basing similarity values between pairs of objects or users on usage behaviors associated with system objects, or usage behaviors that are associated with, directly or indirectly, specific system users.

Claims

1. An adaptive network-based system, comprising:

a structural aspect comprising a plurality of objects;

a usage aspect comprising a plurality of usage behaviors, wherein the plurality of usage behaviors are associated with one or more users; and

a function to identify a subset of the structural aspect by applying an affinity propagation process to the plurality of objects, wherein the affinity propagation process initializes an exemplar attractor value of at least one of the plurality of objects based on, at least in part, the one or more of the usage behaviors.

2. The system of claim 1, wherein a structural aspect comprising a plurality of objects comprises:

an object representing one or more users.

3. The system of claim 1, wherein a function to identify a subset of the structural aspect by applying an affinity propagation process to the plurality of objects, wherein the affinity propagation process initializes the exemplar attractor value of at least one of the plurality of objects based on, at least in part, with one or more of the usage behaviors comprises:

a function to identify the exemplar object that corresponds to the identified subset of the structural subset.

4. The system of claim 3, further comprising:

a function to deliver to a user the object identified as an exemplar.

5. The system of claim 3, further comprising:

a function to deliver a recommendation to a user, wherein the recommendation includes the object identified as an exemplar.

6. The system of claim 1, wherein a function to identify a subset of the structural aspect by applying an affinity propagation process to the plurality of objects, wherein the affinity propagation process initializes the exemplar attractor value of at least one of the plurality of objects based on, at least in part, the one or more of the usage behaviors comprises:

an affinity propagation function wherein the initialized similarity value between a first object of the plurality of objects and a second object of the plurality of objects is based, at least in part, on the one or more of the usage behaviors.

7. The system of claim 1, wherein a structural aspect comprising a plurality of objects comprises:

a relationship between a first object of the plurality of objects and a second object of the plurality of objects; and

a relationship indicator corresponding to the relationship.

8. The system of claim 7, further comprising:

an affinity propagation function wherein the similarity value between a pair of objects of the plurality of objects is initialized based on, at least in part, the relationship indicator corresponding to the said pair of objects.

9. The system of claim 7, further comprising:

a function to determine a relationship indicator associated with a pair of objects based on, at least in part, an affinity propagation message value associated with the said pair of objects.

10. An adaptive social system, comprising:

a usage aspect comprising a plurality of usage behaviors, wherein the plurality of usage behaviors are associated with a plurality of users; and

a function to identify a subset of users of the structural aspect by applying an affinity propagation process to the plurality of users, wherein the affinity propagation process initializes the exemplar attractor value of at least one user of the plurality of users based on, at least in part, the one or more of the usage behaviors.

11. The system of claim 10, further comprising:

an object with an association with a user;

one or more usage behaviors associated with the object; and

a function to associate the one or more usage behaviors associated with the object with the user.

12. The system of claim 10, wherein a function to identify a subset of users of the structural aspect by applying an affinity propagation process to the plurality of users, wherein the affinity propagation process initializes the exemplar attractor value of at least one user of the plurality of users based on, at least in part, the one or more of the usage behaviors comprises:

a function to identify the exemplar that corresponds to the identified subset of users.

13. The system of claim 10, further comprising:

a function to determine an affinity between a pair of users based on, at least in part, an affinity propagation message value associated with the said pair of users.

14. A computer-based adaptive network method, comprising:

accessing a structural aspect comprising a plurality of objects;

accessing a usage aspect comprising a plurality of usage behaviors, wherein the plurality of usage behaviors are associated with one or more users; and

identifying a subset of the structural aspect by applying an affinity propagation process to the plurality of objects, wherein the affinity propagation process initializes an exemplar attractor value of at least one of the plurality of objects based on, at least in part, the one or more of the usage behaviors.

15. The method of claim 14, wherein accessing a structural aspect comprising a plurality of objects comprises:

accessing an object representing one or more users.

16. The method of claim 14, wherein identifying a subset of the structural aspect by applying an affinity propagation process to the plurality of objects, wherein the affinity propagation process initializes the exemplar attractor value of at least one of the plurality of objects based on, at least in part, with one or more of the usage behaviors comprises:

identifying the exemplar object that corresponds to the identified subset of the structural subset.

17. The method of claim 16, further comprising:

receiving a representation of the object identified as an exemplar.

18. The system of claim 14, wherein identifying a subset of the structural aspect by applying an affinity propagation process to the plurality of objects, wherein the affinity propagation process initializes the exemplar attractor value of at least one of the plurality of objects based on, at least in part, the one or more of the usage behaviors comprises:

applying an affinity propagation function wherein the initialized similarity value between a first object of the plurality of objects and a second object of the plurality of objects is based, at least in part, on the one or more of the usage behaviors.

19. The method of claim 14, further comprising:

applying an affinity propagation function wherein the similarity value between a first object of the plurality of objects and a second object of the plurality of objects is initialized based on, at least in part, a relationship indicator associated with a relationship corresponding to a first object of the plurality of objects and a second object of the plurality of objects.

20. The method of claim 14, further comprising:

applying a function to determine a relationship indicator associated with a pair of objects based on, at least in part, an affinity propagation message value associated with the said pair of objects.