Genetic algorithms with robust rank estimation for hearing assistance devices

• US 8,359,283 B2
• Filed: 08/31/2009
• Issued: 01/22/2013
• Est. Priority Date: 08/31/2009
• Status: Active Grant

• Current Assignee: Starkey Laboratories Incorporated
• Original Assignee: Starkey Laboratories Incorporated
• Inventors: Baskent, Deniz, Durant, Eric
• Primary Examiner(s): Gaffin, Jeffrey A
• Assistant Examiner(s): Hill, Stanley K
• Application Number:
  US12/550,768
• Publication Number:
  US 20110055120A1
• Time in Patent Office: 1,240 Days
• Field of Search: 706 12- 13
• US Class Current: 706/13
• CPC Class Codes: G06N 3/126 Evolutionary algorithms, e.g. genetic algorithms or genetic programming

Abstract

Disclosed herein, among other things, is a method for fitting a hearing assistance device for a user using a genetic algorithm. Stimulus pairs are presented to the user using a computer, the stimulus pairs adapted to provide contrasting options for selection from a set of stimuli stored in the computer. Inputs are received from the user entered into the computer, including preference judgments of the user. A score is calculated for each stimulus of the pair using the computer to execute a rank agreement function to maximize agreement between scores and the preference judgments. A set of genes is selected based on the scores, where the set of genes correspond to hearing assistance device parameters. The set of genes is operated on with a genetic algorithm using the assigned scores to obtain a child set of genes. The child set is used to provide parameter values during fitting.

Claims

1. A method of fitting a hearing assistance device for a user, comprising:

presenting to the user using a computer stimulus pairs adapted to provide contrasting options for selection from a set of stimuli stored in the computer;

receiving inputs from the user entered into the computer, including preference judgments of the user;

calculating a score for each stimulus of the pair using the computer to execute a rank agreement function to maximize agreement between scores and the preference judgments;

selecting a set of genes based on the scores, wherein the set of genes correspond to hearing assistance device parameters;

operating on the set of genes with a genetic algorithm using the assigned scores to obtain a child set of genes; and

determining in which order to present various stimulus pairs from the set of stimuli to the user for paired preference judgments, by minimizing degrees of distance between elements of various possible stimulus pairs,

wherein the child set is used to provide parameter values during fitting of the hearing assistance device.

2. The method of claim 1, further comprising:

determining a number of stimulus pairs to present to the user based on a total number of stimuli in the set.

3. The method of claim 2, wherein determining a number of stimulus pairs to present includes using a logarithmic function.

4. The method of claim 1, wherein determining in which order to present various stimulus pairs includes using a symmetric matrix representing a number of comparisons needed to move between stimuli.

5. The method of claim 1, wherein determining in which order to present various stimulus pairs includes choosing a next pair to present to the user in a way that attempts to minimize a connection distance between stimuli presented.

6. The method of claim 1, wherein determining in which order to present various stimulus pairs includes choosing as a first pair to present to the user stimuli that have a weakest connection to the set of stimuli.

7. The method of claim 1, wherein calculating a score for each stimulus includes assigning scores to maximize a sum of a set of rank agreement values for each preference judgment made using a rank agreement function.

8. The method of claim 7, wherein the rank agreement function includes a function of difference in scores between stimuli being compared and preference strength.

9. The method of claim 1, wherein the rank agreement function includes a multidimensional unconstrained nonlinear optimization function.

10. The method of claim 1, wherein presenting stimulus pairs to the user includes presenting the user with a list of strength of preference judgments, the list including at least four options for the user to select from to provide feedback on which stimulus is preferred and the strength of preference.

11. A method for fitting a hearing assistance device for a user using a computer executed process, comprising:

selecting a first pair of stimuli from a population stored in computer memory by minimizing degrees of distance between elements of various possible stimulus pairs, the first pair including a first and second sound stimulus representing at least one device parameter;

presenting the user with the first and second sound stimulus using the hearing assistance device;

presenting the user with a list of strength of preference judgments using the computer, the list including at least four options for the user to select from to provide feedback on which stimulus is preferred and the strength of preference;

receiving a user selection; and

using the user selection and a genetic algorithm to rank stimuli in the population using the computer.

12. The method of claim 11, wherein the list includes at least seven options.

13. The method of claim 11, further comprising:

selecting a second pair of stimuli from the population, the second pair including a third and fourth sound stimulus representing at least one device parameter;

presenting the user with the third and fourth sound stimulus using the hearing assistance device;

presenting the user with a list of strength of preference judgments using the computer, the list including at least four options for the user to select from to provide feedback on which stimulus is preferred and the strength of preference;

receiving a second user selection;

repeating the steps of selecting pairs of stimuli, presenting the user with sound stimuli, presenting the user with a list of judgments and receiving user selections, until all combinations of pairs of stimuli from the population are completed; and

ranking stimuli in the population using the user selections.

14. The method of claim 13, wherein ranking stimuli in the population includes using a rating scale to assign points to the user selection based upon the strength of preference, and summing the points to provide an overall preference score for each stimuli.

15. The method of claim 11, wherein presenting a user with a list of strength of preference judgments includes presenting a user with a list including that the user prefers the first stimulus strongly better than the second stimulus, that the user prefers the first stimulus moderately better than the second stimulus, that the user prefers the first stimulus slightly better than the second stimulus, and that the user prefers the first stimulus the same as the second stimulus.

16. An apparatus for fitting a hearing assistance device for a user using a computer executing a genetic algorithm, comprising:

a computer memory having a first population stored therein, the first population including a plurality of stimuli;

a processor executing instructions adapted to select stimulus pairs from the first population to present to the user using the hearing assistance device, the stimulus pairs including a first stimulus and a second stimulus; and

a computer interface adapted to present preference judgment options to the user and to accept an input of a preference judgment between stimulus pairs;

wherein the processor executes instructions adapted to assign a score to stimuli to maximize a sum of a set of rank agreement values for each preference judgment made using a rank agreement function, and to perform a genetic algorithm using the scores to obtain a child set of stimuli, the child set used to provide parameter values during fitting of the hearing assistance device.

17. The apparatus of claim 16, wherein the preference judgment options include a list of strength of preference judgments, the list including at least four options for the user to select from to provide feedback on which of the stimuli is preferred and the strength of preference.

18. The apparatus of claim 17, wherein an option includes that the user prefers the first stimulus strongly better than the second stimulus.

19. The apparatus of claim 17, wherein an option includes that the user prefers the first stimulus moderately better than the second stimulus.

20. The apparatus of claim 17, wherein an option includes that the user prefers the first stimulus slightly better than the second stimulus.

21. The apparatus of claim 17, wherein an option includes that the user prefers the first stimulus the same as the second stimulus.

22. The apparatus of claim 17, wherein an option includes that the user prefers the second stimulus strongly better than the first stimulus.

23. The apparatus of claim 17, wherein an option includes that the user prefers the second stimulus moderately better than the first stimulus.

24. The apparatus of claim 17, wherein an option includes that the user prefers the second stimulus slightly better than the first stimulus.

25. A method of fitting a hearing assistance device for a user, comprising:

presenting to the user using a computer stimulus pairs adapted to provide contrasting options for selection from a set of stimuli stored in the computer;

receiving inputs from the user entered into the computer, including preference judgments of the user;

calculating a score for each stimulus of the pair using the computer to execute a rank agreement function to maximize agreement between scores and the preference judgments;

selecting a set of genes based on the scores, wherein the set of genes correspond to hearing assistance device parameters; and

operating on the set of genes with a genetic algorithm using the assigned scores to obtain a child set of genes,

wherein the child set is used to provide parameter values during fitting of the hearing assistance device, and

wherein calculating a score for each stimulus includes assigning scores to maximize a sum of a set of rank agreement values for each preference judgment made using a rank agreement function.

26. The method of claim 25, wherein the rank agreement function includes a function of difference in scores between stimuli being compared and preference strength.

27. A method of fitting a hearing assistance device for a user, comprising:

presenting to the user using a computer stimulus pairs adapted to provide contrasting options for selection from a set of stimuli stored in the computer;

receiving inputs from the user entered into the computer, including preference judgments of the user;

calculating a score for each stimulus of the pair using the computer to execute a rank agreement function to maximize agreement between scores and the preference judgments;

selecting a set of genes based on the scores, wherein the set of genes correspond to hearing assistance device parameters; and

operating on the set of genes with a genetic algorithm using the assigned scores to obtain a child set of genes,

wherein the child set is used to provide parameter values during fitting of the hearing assistance device, and

wherein the rank agreement function includes a multidimensional unconstrained nonlinear optimization function.

28. A method for fitting a hearing assistance device for a user using a computer executed process, comprising:

selecting a first pair of stimuli from a population stored in computer memory, the first pair including a first and second sound stimulus representing at least one device parameter;

presenting the user with the first and second sound stimulus using the hearing assistance device;

presenting the user with a list of strength of preference judgments using the computer, the list including at least four options for the user to select from to provide feedback on which stimulus is preferred and the strength of preference;

receiving a user selection;

using the user selection and a genetic algorithm to rank stimuli in the population using the computer;

selecting a second pair of stimuli from the population, the second pair including a third and fourth sound stimulus representing at least one device parameter;

presenting the user with the third and fourth sound stimulus using the hearing assistance device;

presenting the user with a list of strength of preference judgments using the computer, the list including at least four options for the user to select from to provide feedback on which stimulus is preferred and the strength of preference;

receiving a second user selection;

repeating the steps of selecting pairs of stimuli, presenting the user with sound stimuli, presenting the user with a list of judgments and receiving user selections, until all combinations of pairs of stimuli from the population are completed; and

ranking stimuli in the population using the user selections,

wherein ranking stimuli in the population includes using a rating scale to assign points to the user selection based upon the strength of preference, and summing the points to provide an overall preference score for each stimuli.