Serendipitous recommendations system and method

• US 8,645,292 B2
• Filed: 10/07/2011
• Issued: 02/04/2014
• Est. Priority Date: 03/29/2011
• Status: Active Grant

• Current Assignee: Manyworlds Incorporated
• Original Assignee: Manyworlds Incorporated
• Inventors: Moneypenny, Naomi Felina, Flinn, Steven Dennis
• Primary Examiner(s): CHEN, ALAN S
• Application Number:
  US13/268,047
• Publication Number:
  US 20120254079A1
• Time in Patent Office: 851 Days
• Field of Search: None
• US Class Current: 706/12
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 40/56 Natural language generation
• CPC Class Codes: G06K 7/10366 the interrogation device being adapted for miscellaneous applications
• CPC Class Codes: G06K 7/10475 arrangements to facilitate interaction with further interrogation devices, e.g. such that at least two interrogation devices may function and cooperate in a network of such devices
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• 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
• CPC Class Codes: G06Q 30/0631 Item recommendations
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

A computer-implemented serendipitous recommendations system and method generates recommendations for delivery to system users in accordance with settings of desired levels of serendipity, including serendipity levels established through use of serendipity tuning controls operable by users. The recommendations are informed by an interest affinity anomaly function that identifies contrasting interest affinities between recommendation recipients and other users. Explanations may be generated that provide reasons as to why a recommendation was delivered to a user, and the explanation may include a selection of phrases that are influenced by a serendipity level setting, and may include an expression of a level of confidence with regard to the recommendation.

Claims

1. A computer-implemented method, comprising:

contributing a plurality of behaviors to an affinity vector generation function executed on a processor-based computing device, wherein the affinity vector generation function generates a first affinity vector based, at least in part, on an inference from the plurality of behaviors;

receiving a recommendation from a computer-implemented recommendation function, wherein the recommendation function generates the recommendation based, at least in part, on an identification of contrasting corresponding affinity values between the first affinity vector and a second affinity vector, wherein the second affinity vector is associated with a user other than the recipient of the recommendation; and

tuning the recommendation function in accordance with a recommendation recipient'"'"'s setting of a computer-implemented tuning control, wherein the setting influences the application of the contrasting corresponding affinity values in generating the recommendation.

2. The method of claim 1, further comprising:

receiving the recommendation, wherein the second affinity vector is selected for comparison to the first affinity vector based, at least in part, on a relatively high positive correlation of a plurality of the second affinity vector'"'"'s affinity values with a corresponding plurality of affinity values of the first affinity vector.

3. The method of claim 1, further comprising:

receiving the recommendation, wherein a confidence level with regard to the contrasting affinity values is determined.

4. The method of claim 1, further comprising:

receiving an explanation comprising a reason for having received the recommendation.

5. The method of claim 1, further comprising:

receiving the recommendation, wherein the recommendation is generated in accordance with a probability distribution.

12. The method of claim 1, further comprising:

tuning the recommendation function in accordance with the recommendation recipient'"'"'s setting of the computer-implemented tuning control, wherein the setting influences the application of the contrasting corresponding affinity values in generating the recommendation, wherein the setting influences a confidence level.

6. A computer-implemented system, comprising:

an affinity vector generation function executed on a processor-based computing device, wherein the affinity vector generation function generates a first affinity vector based, at least in part, on an inference from a plurality of behaviors associated with a first user;

a computer-implemented recommendation function, wherein the recommendation function generates the recommendation based, at least in part, on an identification of contrasting corresponding affinity values between the first affinity vector and a second affinity vector, wherein the second affinity vector is associated with a second user; and

a user tuning control function that influences the application of the contrasting corresponding affinity values in generating the recommendation in accordance with a user'"'"'s setting of a tuning control.

7. The system of claim 6, further comprising:

an affinity vector comparison function that selects the second affinity vector for comparison to the first affinity vector based, at least in part, on a relatively high positive correlation of a plurality of the second affinity vector'"'"'s affinity values with a corresponding plurality of affinity values of the first affinity vector.

8. The system of claim 6, further comprising:

a confidence level function that determines a confidence level with regard to the contrasting corresponding affinity values.

9. The system of claim 6, further comprising:

a value of information function that determines an expected value of information of the recommendation.

10. The system of claim 6, further comprising:

the recommendation function, wherein the recommendation function generates the recommendation in accordance with a probability distribution.

11. The system of claim 10, further comprising:

the probability distribution, wherein the probability distribution is in accordance with a Levy Walk distribution.

13. The system of claim 6, further comprising:

the user tuning control function that influences the application of the contrasting corresponding affinity values in generating the recommendation in accordance with the user'"'"'s setting of a tuning control, wherein the user is a recipient of the recommendation.

14. The system of claim 6, further comprising:

the user tuning control function that influences the application of the contrasting corresponding affinity values in generating the recommendation in accordance with the user'"'"'s setting of a tuning control, wherein the user'"'"'s setting influences application of a probability distribution.

15. An article comprising a non-transitory computer-readable medium storing instructions for enabling a processor-based system to:

generate a first affinity vector based, at least in part, on an inference from a plurality of behaviors associated with a first user;

generate a recommendation based, at least in part, on an identification of contrasting corresponding affinity values between the first affinity vector and a second affinity vector, wherein the second affinity vector is associated with a second user; and

generate the recommendation in accordance with a recommendation recipient'"'"'s setting of a computer-implemented tuning control, wherein the recommendation recipient'"'"'s setting influences the application of the contrasting corresponding affinity values in generating the recommendation.

16. The article of claim 15, further comprising the non-transitory computer-readable medium storing instructions for enabling the processor-based system to:

determine a confidence level with regard to the contrasting corresponding affinity values.

17. The article of claim 15, further comprising the non-transitory computer-readable medium storing instructions for enabling the processor-based system to:

determine an expected value of information of the recommendation.

18. The article of claim 15, further comprising the non-transitory computer-readable medium storing instructions for enabling the processor-based system to:

generate the recommendation in accordance with a probability distribution.

19. The article of claim 15, further comprising the non-transitory computer-readable medium storing instructions for enabling the processor-based system to:

generate the recommendation in accordance with the recommendation recipient'"'"'s setting of a computer-implemented tuning control, wherein the recommendation recipient'"'"'s setting influences the application of the contrasting corresponding affinity values in generating the recommendation, wherein the recommendation recipient'"'"'s setting influences application of a confidence level.

20. The article of claim 15, further comprising the non-transitory computer-readable medium storing instructions for enabling the processor-based system to:

generate the recommendation in accordance with the recommendation recipient'"'"'s setting of the computer-implemented tuning control, wherein the recommendation recipient'"'"'s setting influences application of a probability distribution.