Neural network system and method for controlling information output based on user feedback

• US 8,260,733 B2
• Filed: 10/10/2006
• Issued: 09/04/2012
• Est. Priority Date: 02/02/1999
• Status: Expired due to Fees

• Current Assignee: Hyper Search LLC (IPValuation Partners, LLC)
• Original Assignee: Garbortese Holdings LLC (Intellectual Ventures LLC)
• Inventors: Pope, Ivan, Sullivan, Alan
• Primary Examiner(s): Chaki, Kakali
• Assistant Examiner(s): RIFKIN, BEN M
• Application Number:
  US11/544,858
• Publication Number:
  US 20070038588A1
• Time in Patent Office: 2,156 Days
• Field of Search: 706/15, 706/21
• US Class Current: 706/21
• CPC Class Codes: G06F 16/3326 using relevance feedback from the user, e.g. relevance feedback on documents, documents sets, document terms or passages
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/063 using electronic means
• CPC Class Codes: G06N 3/084 Backpropagation, e.g. using gradient descent
• CPC Class Codes: H04L 51/212 using filtering or selective blocking
• CPC Class Codes: H04L 67/02 based on web technology, e.g. hypertext transfer protocol [HTTP]
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 69/329 in the application layer [OSI layer 7]

Abstract

A system and method for controlling information output based on user feedback about the information is provided. At least one neural network module selects one or more of a plurality of objects to receive information from a plurality of information sources based at least in part on a plurality of inputs and a plurality of weight values during an epoch. The information sources may include electronic mail providers, chat participants, or page links. Recipients of the objects provide feedback about the information during an epoch. At the conclusion of an epoch, the neural network takes the feedback and generates a rating value for each of the plurality of objects. Based on the rating value and the selections made, the neural network redetermines the weight values within the network. The neural network then selects the objects to receive information during a subsequent epoch using the redetermined weight values and the inputs for that subsequent epoch.

Claims

1. A system, comprising:

means for conveying a first set of links to respective network locations in response to navigation to a first network location by a device;

means for receiving a selection of one or more links, of the set of links, during an epoch comprising a duration of a visit to the first network location;

means for assessing selection results of a plurality of epochs received from a plurality of devices, including means for altering at least one weight value associated with a respective at least one node of a neural network based on the selection results;

means for learning a probability, based on the at least one weight value, that navigation to the first network location is to be followed by selection of a first subset of the set of links; and

means for conveying a second list of links comprising the first subset of the set of links in response to a subsequent navigation to the first network location.

2. The system of claim 1, further comprising

means for accepting, as one or more input nodes of the neural network, one or more first attributes of the first network location,

wherein the means for conveying the second list of links includes means for selecting the first subset of the set of links based at least in part on a value of an output node of the neural network.

3. The system of claim 1, further comprising means for establishing an association between the first subset of the set of links and the first network location based on the probability.

4. The system of claim 1, wherein the means for assessing includes means for correlating the selection results with one or more attributes associated with the plurality of devices.

5. The system of claim 1, further comprising:

means for excluding a second subset of the set of links during the subsequent epoch based on a determined probability that navigation to the first network location is not followed by selection of the second subset of the set of links.

6. A method, comprising:

conveying, in response to selection of a first network location, a list of one or more links to respective one or more second network locations;

modifying, as facilitated by one or more processors, at least one weight value associated with a respective at least one node of a neural network based on selection of a link, of the one or more links, by a device;

learning, by the neural network based on the at least one weight value, that selection of the first network location during respective epochs is followed by selection, according to respective probabilities, of a first subset of the one or more links, the respective epochs comprising durations of time for visits to the first network location by the respective devices; and

conveying, during a subsequent visit to the first network location, a modified list of the one or more links comprising the first subset of the one or more links based on the determining.

7. The method of claim 6, wherein the conveying the first list comprises conveying the list of links to respective web pages.

8. The method of claim 6, further comprising establishing, based on the learning, an association between the first network location and the first subset of the one or more links for a subset of the devices having at least one common attribute.

9. The method of claim 6, further comprising receiving, from at least one of the devices, feedback indicating a relevance of at least one link of the one or more links.

10. The method of claim 9, further comprising establishing, by the neural network, an association between the first network location and the first subset of the one or more links based on the feedback.

11. The method of claim 6, further comprising receiving selection of the first subset of the one or more links via web site interfaces associated with the devices.

12. The method of claim 6, further comprising:

excluding a second subset of the one or more links from the modified list in response to determining that selection of the first network location is not followed, according to a probability, by selection of the second subset.

13. The method of claim 9, further comprising establishing an association between the feedback and one or more key words associated with the first network location.

14. The method of claim 13,

further comprising selecting the one or more links for inclusion in the second list based at least in part on the association between the feedback and the one or more key words.

15. The method of claim 6, further comprising accepting, as one or more input nodes of the neural network, one or more first attributes of the first network location, wherein the conveying the modified list includes selecting the first subset of the one or more links based at least in part on a value of an output node of the neural network.

16. A non-transitory computer-readable medium having stored thereon computer executable instructions that, in response to execution, cause a computing system to perform operations, including:

sending a set of links via a network, in response to navigation to a first network location, for presentation in a page of slots associated with the first network location, the set of links corresponding to a set of respective second network locations;

adjusting one or more weight values corresponding to respective one or more nodes of a neural network based on selections of one or more links, of the set of links, by a computing device during an epoch, the epoch comprising a duration of a visit to the first network location by the computing device;

learning, by the neural network based on the one or more weight values, that navigation to the first network location is likely to be followed by selection of a first subset of the set of links; and

selecting, by the neural network, the first subset of the set of links for inclusion in the page of slots for a subsequent epoch based on the learning.

17. The computer-readable medium of claim 16, wherein the selecting includes:

accepting, as a value of at least one input node of the neural network, at least one attribute of the first network location in response to navigation to the first network location during the subsequent epoch; and

selecting the first subset of the set of links based on at least one value of a respective at least one output node of the neural network.

18. The computer-readable medium of claim 16, the operations further including:

learning, by the neural network based on the one or more weight values, that navigation to the first network location is not followed by selection of a second subset of the set of links for a majority of multiple epochs; and

excluding the second subset of the set of links from the page of slots for the subsequent epoch.

19. The computer-readable medium of claim 16, wherein the learning includes learning, by the neural network based on the one or more weight values, that navigation to the first network location is likely to be followed by the selection of the first subset of the set of links for computing devices associated with least one common attribute;

and

the selecting includes selecting the first subset of the set of links for inclusion in the page of slots for the subsequent epoch for computing devices associated with the at least one common attribute.

20. The computer-readable medium of claim 16, the operations further including:

receiving feedback from the computing device indicating relevance of at least one of the one or more links,

wherein the selecting the first subset of the set of links for inclusion in the page of slots is further based on the feedback.

21. The computer-readable medium of claim 16, wherein the learning includes learning, by the neural network, that navigation to the first network location is likely to be followed by selection of the first subset of the set of links for computing devices having at least one common attribute;

and

wherein the selecting includes selecting the first subset of the set of links for the computing devices having the at least one common attribute.

22. A system, comprising:

a processor;

a memory communicatively coupled to the processor and having stored thereon computer-executable components that implement the system, including;

a neural network;

an application module configured to present a first list of links to web pages in response to navigation to a first web page by a first device and receive a selection of at least one link in the first list of links during an epoch, the epoch comprising a duration of a visit to the first web page; and

a neural network module configured to modify the one or more weight values of respective one or more neural network nodes of the neural network based on the selection of the at least one link during the epoch;

wherein the neural network is configured to learn, based on the one or more weight values, that navigation to the first web page is likely to be followed by selection of a subset of links of the list of links; and

wherein, in response to a subsequent navigation the first web page, the application module is further configured to present a second list of links comprising the subset of links based at least on at least one output of the neural network.

23. The system of claim 22, wherein the application module is further configured to receive, from the first device, feedback indicating a relevance of at least one of the first subset of links.

24. The system of claim 23, wherein the neural network is further configured to:

establish an association between the feedback and one or more keywords in a query used to invoke the first web page; and

determine whether to include the subset of links in the second list of links based on the association between the feedback and the one or more keywords in response to a subsequent query including the one or more keywords.

25. The system of claim 22, wherein the neural network is further configured to receive at least one attribute value of the first web page as a respective at least one input node value, and to select the second list of links based on at least one output node value.