ESTIMATING CONGNITIVE-LOAD IN HUMAN-MACHINE INTERACTION

• US 20130325482A1
• Filed: 02/07/2013
• Published: 12/05/2013
• Est. Priority Date: 05/29/2012
• Status: Abandoned Application

• Current Assignee: GM Global Technology Operations LLC (General Motors Company)
• Original Assignee: GM Global Technology Operations LLC (General Motors Company)
• Inventors: Tzirkel-Hancock, Eli, Tsimhoni, Omer
• Application Number:
  US13/761,541
• Publication Number:
  US 20130325482A1
• Time in Patent Office: Days
• US Class Current: 704/275
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 21/18 Details of the transformation process

Abstract

Estimating cognitive-load of a user in human-machine interaction by identifying an expression of cognitive-load within a user expression captured by a dialogue system and using a user model to estimate a level of the cognitive-load based on the expression of cognitive-load.

Claims

1. A method for estimating cognitive-load through human-machine interface, the method comprising,

performing computer-implemented steps of:

identifying an expression of cognitive-load within a user expression expressed by a user interacting with a dialogue system; and

using a user model to estimate a level of the cognitive-load experienced by the user interacting with the dialogue-system based on the expression of cognitive-load.

2. The method of claim 1, wherein the user model includes a dynamic Bayesian network.

3. The method of claim 1, wherein the expression of cognitive-load is selected from the group consisting of disfluency, a statement indicative of cognitive-load, and disfluency combined with a statement indicative of cognitive-load.

4. The method of claim 2, wherein the dynamic Bayesian network includes an observed user-dialogue act variable depending directly or indirectly on a cognitive-load variable.

5. The method of claim 4, wherein the cognitive-load variable depends on at least one previous dialogue-turn variable.

6. The method of claim 5, wherein the previous dialogue-turn variable includes at least one of any of the previous dialogue-turn variables selected from the group consisting of previous cognitive-load variable, previous user-goal variable, and previous machine-dialogue-action variable.

7. The method of claim 1, wherein the user expression is selected from the group consisting of a verbal expression, a head motion, a facial expression, a hand gesture, and an application of pressure to a steering wheel wherein the pressure exceeds a threshold pressure.

8. The method of claim 1, wherein the dialogue system includes a multi-modal dialogue system.

9. The method of claim 1, wherein the dialogue system receives input from at least one data capture device non-related to the dialogue system.

10. The method of claim 1, further comprising selecting a system-dialogue act at least partially based on goal probabilities determined by the user model.

18. The system of claim 10, wherein the dialogue system receives input from at least one data capture device non-related to the dialogue system.

11. A dialogue system for estimating cognitive-load of a user interacting with the system, the system comprising:

a processor configured to;

recognize an expression of cognitive-load in a user expression captured by the dialogue system; and

use a user model to estimate a level of the cognitive-load experienced by the user at least partially based on the expression of cognitive-load.

12. The system of claim 11, wherein the user model includes a dynamic Bayesian network.

13. The system of claim 12, wherein the expression of cognitive-load is selected from the group consisting of a verbal disfluency, a statement indicative of cognitive-load, and a verbal disfluency combined with a statement indicative of cognitive-load.

14. The system of claim 12, wherein the dynamic Bayesian network includes an observed user-dialogue-act variable depending directly or indirectly on a cognitive-load variable.

15. The system of claim 14, wherein the cognitive-load variable depends on at least one previous dialogue-turn variable.

16. The system of claim 15, wherein the previous dialogue-turn variable is selected from the group consisting of previous cognitive-load variable, previous user-goal variable, and previous machine-dialogue-act variable.

17. The system of claim 11, wherein the dialogue system includes a multi-modal dialogue system.

19. A non-transitory computer-readable medium having stored thereon instructions for estimating cognitive-load of a user interacting with a dialogue system which when executed by a processor causes the processor to perform a method comprising:

recognizing an expression of cognitive-load in a user expression captured by a dialogue system; and

using a user model to estimate a level of the cognitive-load experienced by the user at least partially based on the expression of cognitive-load.

20. The non-transitory computer-readable medium of claim 19, wherein the user model includes a dynamic Bayesian network.