Discovering and classifying situations that influence affective response

• US 20120290521A1
• Filed: 06/25/2011
• Published: 11/15/2012
• Est. Priority Date: 05/11/2011
• Status: Active Grant

• Current Assignee: Affectomatics ltd.
• Original Assignee: Affectivon Limited
• Inventors: Frank, Ari M., Thieberger, Gil, Thieberger Ben-Haim, Anat
• Granted Patent: US 8,965,822 B2
• Time in Patent Office: Days
• US Class Current: 706/45
• CPC Class Codes: G06F 3/013 Eye tracking input arrangements G06F3/015 takes precedence
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation

Abstract

Methods for detecting and validating an estimated situation using a situation-dependent predictor of a user response to token instances, including: receiving a temporal window of token instances and a putative situation for a user; predicting an expected response of the user to being exposed to the temporal window of token instances; receiving a value of a measurement channel of the user taken during exposure of the user to the temporal window of token instances; identifying that difference between the received value of the user measurement channel and the predicted expected response of the user is above a predefined threshold; and indicating that the putative situation is wrong.

Claims

1. (canceled)

2. A method for detecting a change in a situation of a user based on measurements of the user and a predictor of affective response to temporal windows of token instances, comprising:

receiving a first temporal window of token instances;

predicting a first expected response of the user to being exposed to the first temporal window of token instances;

wherein the first expected response is expressed in terms of a value of a user measurement channel;

receiving a first value of the user measurement channel taken during exposure of the user to the first temporal window of token instances;

receiving a second temporal window of token instances;

predicting a second expected response of the user to being exposed to the second temporal window of token instances;

wherein the second expected response is expressed in terms of a value of the user measurement channel;

receiving a second value of the user measurement channel taken during exposure of the user to the second temporal window of token instances;

identifying that difference between the first and the second expected responses is below a predefined threshold, while difference between the first and the second values of the user measurement channel is above the predefined threshold; and

signaling that the situation of the user has changed.

3. The method of claim 2, wherein predicting the first and the second expected responses comprises utilizing a situation-dependent predictor;

and the predictor receives a value corresponding to a putative situation of the user.

4. The method of claim 3, wherein predicting the expected response of the user to being exposed to the temporal window of token instances is done utilizing a regression model.

5. The method of claim 3, wherein predicting the expected response of the user to being exposed to the temporal window of token instances is done utilizing a Naive Bayes classifier.

6. The method of claim 3, wherein predicting the expected response of the user to being exposed to the temporal window of token instances is done utilizing a maximum entropy model classifier.

7. The method of claim 3, wherein predicting the expected response of the user to being exposed to the temporal window of token instances is done utilizing a neural network.

8. The method of claim 3, wherein predicting the expected response of the user to being exposed to the temporal window of token instances is done utilizing a random forest model.

9. The method of claim 2, further comprising predicting a third expected response to the second temporal window of token instances utilizing a value corresponding to a different situation;

identifying that the difference between the third expected response of the second value of the user measurement channel is below the predefined threshold; and

signaling that the user might be in the different situation.

10. A method for validating an estimated situation using a situation-dependent predictor of a user response to token instances, comprising:

receiving a temporal window of token instances and a putative situation for a user;

predicting an expected response of the user to being exposed to the temporal window of token instances;

wherein the expected response is expressed in terms of a value of a user measurement channel;

receiving a value of the user measurement channel taken during exposure of the user to the temporal window of token instances;

identifying that difference between the received value of the user measurement channel and the predicted expected response of the user is above a predefined threshold; and

indicating that the putative situation is wrong.

11. The method of claim 10, receiving a second temporal window of token instances and a second putative situation for the user;

predicting a second expected response of the user to being exposed to the second temporal window of token instances;

receiving a second value of the user measurement channel taken during the user'"'"'s exposure to the second temporal window of token instances;

identifying that difference between the second received value of the user measurement channel and the second predicted expected response of the user is below the predefined threshold; and

indicating that the second putative situation is valid.

12. The method of claim 10, wherein predicting the expected response comprises utilizing a situation-dependent predictor;

and the predictor receives a value corresponding to the user'"'"'s putative situation.

13. The method of claim 12, wherein predicting the expected response of the user to being exposed to the temporal window of token instances is done utilizing a regression model.

14. The method of claim 12, wherein predicting the expected response of the user to being exposed to the temporal window of token instances is done utilizing a Naive Bayes classifier.

15. The method of claim 12, wherein predicting the expected response of the user to being exposed to the temporal window of token instances is done utilizing a maximum entropy model classifier.

16. The method of claim 12, wherein predicting the expected response of the user to being exposed to the temporal window of token instances is done utilizing a neural network.

17. The method of claim 12, wherein predicting the expected response of the user to being exposed to the temporal window of token instances is done utilizing a random forest model.