METHOD OF OPERATING AN IN-SITU FITTING SYSTEM AND AN IN-SITU FITTING SYSTEM

Abstract

A method of operating an in-situ fitting system (100) adapted to suggest an improved hearing aid parameter setting for a current user based on evaluated hearing aid parameter settings from a plurality of other users. The invention is also directed at an in-situ fitting system adapted to carry out said method.

Claims

1. A method of operating an in-situ fitting system adapted for adjusting a hearing aid parameter setting of a hearing aid system in response to a trigger event indicating that a first hearing aid parameter setting is not satisfactory, wherein adjustment of the first hearing aid parameter setting for a current user is carried out through the steps of: providing at least one server operationally connected with said hearing aid system; providing to said at least one server a first plurality of evaluated hearing aid parameter settings each associated with a specific hearing aid system user; using said first plurality of evaluated hearing aid parameter settings and their association with a specific hearing aid system user to predict a new hearing aid parameter setting adapted to replace said first hearing aid parameter setting for the current user, and providing the new hearing aid parameter setting to the hearing aid system and replacing said first hearing aid parameter setting with said new hearing aid parameter setting.

2. The method according to claim 1, wherein the step of using said first plurality of evaluated hearing aid parameter settings and their association with a specific hearing aid system user to predict a new hearing aid parameter setting adapted to replace said first hearing aid parameter setting for the current user comprises the further steps of: providing a ratings matrix representing said first plurality of evaluated hearing aid parameter settings each associated with a specific hearing aid system user; decomposing the ratings matrix into the product of two lower dimensionality matrices, predicting the new hearing aid parameter setting by at least partly multiplying said two lower dimensionality matrices in order to determine a vector representing the estimated ratings of the hearing aid parameter settings, for the current user and selecting the hearing aid parameter setting with the highest estimated rating as the new hearing aid parameter setting for the current user.

3. The method according to claim 2 wherein the step of decomposing the ratings matrix into the product of two lower dimensionality matrices is carried out under the requirement that all the elements of both the ratings matrix and said two lower dimensionality matrices are non-negative.

4. The method according to claim 1, wherein the step of using said first plurality of evaluated hearing aid parameter settings and their association with a specific hearing aid system user to predict a new hearing aid parameter setting adapted to replace said first hearing aid parameter setting for the current user comprises the further steps of: identifying a first plurality of specific hearing aid system users that are similar to the current user based on the similarity between evaluated hearing aid parameter settings for respectively the current user and the other specific hearing aid system users; and using the evaluated hearing aid parameter settings of the second plurality of specific hearing aid system users to predict the new hearing aid parameter setting for the current user.

5. The method according to claim 1, wherein a hearing aid parameter settings is considered to be evaluated in response to a trigger event from a group comprising: the hearing aid parameter setting has been stored in the hearing aid system, the hearing aid parameter setting has been selected or rated in an in-situ comparison between two different hearing aid parameter settings; and the value of an internal response function for the hearing aid parameter setting has been estimated with sufficient precision.

6. The method according to claim 1, comprising the further steps of: prompting the current user to evaluate said new hearing aid parameter setting; and providing said evaluation to said at least one server.

7. The method according to claim 1, wherein the step of providing to said at least one server a first plurality of evaluated hearing aid parameter settings each associated with a specific hearing aid system user comprises the further step of additionally associating with at least one specific hearing aid system user at least one characteristic selected from a group comprising age, gender, race, audiogram, type of hearing aid system worn, experience with wearing a hearing aid, native language and country of residence; and wherein the step of using said first plurality of evaluated hearing aid parameter settings and their association with a specific hearing aid system user to predict a new hearing aid parameter setting is adapted to additionally use the additionally associated data.

8. An in situ fitting system adapted for adjusting a hearing aid parameter setting of a hearing aid system in response to a trigger event indicating that a first hearing aid system setting is not satisfactory, wherein the adjustment of the hearing aid parameter setting is carried out through the steps of: providing at least one server operationally connected with said hearing aid system; providing to said at least one server a first plurality of evaluated hearing aid parameter settings each associated with a specific hearing aid system user; using said first plurality of evaluated hearing aid parameter settings and their association with a specific hearing aid system user to predict a new hearing aid parameter setting adapted to replace said first hearing aid parameter setting for the current user, and providing the new hearing aid parameter setting to the hearing aid system and replacing said first hearing aid parameter setting with said new hearing aid parameter setting.

9. A non-transitory computer readable medium carrying instructions which, when executed by a computer causes a method comprising the following steps to be performed: providing at least one server operationally connected with a hearing aid system; receiving by at least one server a first plurality of evaluated hearing aid parameter settings each associated with a specific hearing aid system user; using said first plurality of evaluated hearing aid parameter settings and their association with a specific hearing aid system user to predict a new hearing aid parameter setting adapted to replace said first hearing aid parameter setting for the current user, and providing the new hearing aid parameter setting to the hearing aid system and replacing said first hearing aid parameter setting with said new hearing aid parameter setting.

10. The non-transitory computer readable medium according to claim 9, wherein a software application is adapted to be downloaded from an external server and subsequently may be executed independently of the external server.

11. The non-transitory computer readable medium according to claim 9, wherein a software application is adapted to be executed at least partly from an external server and adapted to be accessed using a web browser of the computer.

12. The method according to claim 2, wherein the step of providing to said at least one server a first plurality of evaluated hearing aid parameter settings each associated with a specific hearing aid system user comprises the further step of additionally associating with at least one specific hearing aid system user at least one characteristic selected from a group comprising age, gender, race, audiogram, type of hearing aid system worn, experience with wearing a hearing aid, native language and country of residence; and wherein the step of using said first plurality of evaluated hearing aid parameter settings and their association with a specific hearing aid system user to predict a new hearing aid parameter setting is adapted to additionally use the additionally associated data.

13. The in situ fitting system according to claim 8 comprising the further steps of: providing a ratings matrix representing said first plurality of evaluated hearing aid parameter settings each associated with a specific hearing aid system user; decomposing the ratings matrix into the product of two lower dimensionality matrices, predicting the new hearing aid parameter setting by at least partly multiplying said two lower dimensionality matrices in order to determine a vector representing the estimated ratings of the hearing aid parameter settings, for the current user and selecting the hearing aid parameter setting with the highest estimated rating as the new hearing aid parameter setting for the current user.

14. The in situ fitting system according to claim 13 wherein the step of decomposing the ratings matrix into the product of two lower dimensionality matrices is carried out under the requirement that all the elements of both the ratings matrix and said two lower dimensionality matrices are non-negative.

15. The non-transitory computer readable medium according to claim 9 carrying instructions which, when executed by a computer causes a method comprising the following additional steps to be performed: providing a ratings matrix representing said first plurality of evaluated hearing aid parameter settings each associated with a specific hearing aid system user; decomposing the ratings matrix into the product of two lower dimensionality matrices, predicting the new hearing aid parameter setting by at least partly multiplying said two lower dimensionality matrices in order to determine a vector representing the estimated ratings of the hearing aid parameter settings, for the current user and selecting the hearing aid parameter setting with the highest estimated rating as the new hearing aid parameter setting for the current user.

16. The non-transitory computer readable medium according to claim 15 carrying instructions which, when executed by a computer causes a method comprising the following additional steps to be performed: wherein the step of decomposing the ratings matrix into the product of two lower dimensionality matrices is carried out under the requirement that all the elements of both the ratings matrix and said two lower dimensionality matrices are non-negative.

17. The non-transitory computer readable medium according to claim 15, wherein a software application is adapted to be downloaded from an external server and subsequently may be executed independently of the external server.

18. The non-transitory computer readable medium according to claim 16, wherein a software application is adapted to be downloaded from an external server and subsequently may be executed independently of the external server.

19. The non-transitory computer readable medium according to claim 15, wherein a software application is adapted to be executed at least partly from an external server and adapted to be accessed using a web browser of the computer.

20. The non-transitory computer readable medium according to claim 16, wherein a software application is adapted to be executed at least partly from an external server and adapted to be accessed using a web browser of the computer.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0025] By way of example, there is shown and described a preferred embodiment of this invention. As will be realized, the invention is capable of other embodiments, and its several details are capable of modification in various, obvious aspects all without departing from the invention. Accordingly, the drawings and descriptions will be regarded as illustrative in nature and not as restrictive. In the drawings:

[0026] FIG. 1 illustrates highly schematically a method of operating an in-situ fitting system according to an embodiment of the invention, and

[0027] FIG. 2 illustrates highly schematically an in-situ fitting system according to an embodiment of the invention.

DETAILED DESCRIPTION

[0028] According to an aspect of the invention it has been found that it provides a significant improvement for the user if the hearing aid system settings can be adapted to the user's current preferences (i.e. personalized). This is even more so because the user's preferences may vary significantly up to several times during a day, as a function of e.g. the time of day (morning, afternoon or evening) or the user's mood or the type of activity the user is engaged in.

[0029] As a consequence of these varying preferences of many users it provides a significant improvement for the user if the personalization can be carried out without having to spend too much time optimizing the settings.

[0030] Furthermore, it has been found that it is of significant importance that the personalization (i.e. the optimization of a hearing aid parameter setting) can be carried out without requiring the user to interact with the hearing aid system in a complex manner.

[0031] Reference is first made to FIG. 1 which illustrates highly schematically a method 100 of operating an in-situ fitting system according to a first embodiment of the invention.

[0032] According to the present embodiment the method is adapted for a adjusting a first hearing aid parameter setting of a hearing aid system in response to a trigger event indicating that the first hearing aid parameter setting is not satisfactory.

[0033] According to variations of the present embodiment the trigger event is selected from a group comprising: activating of a hearing aid system handle adapted to provide an indication that the current hearing aid parameter setting is not satisfactory, and detection that the cognitive stress experienced by the hearing aid system user is above a given threshold. According to more specific embodiment the handle is a button accommodated in a hearing aid of the hearing aid system or a handle implemented in a GUI of an external device, typically a smart phone, of the hearing aid system.

[0034] According to a first step 101 at least one server, operationally connected with the hearing aid system, is provided.

[0035] In variations a hearing aid or an external device of the hearing aid system is connected directly to the server using a wireless link to the internet, based on e.g. the 3G, 4G or upcoming 5G broadband cellular network technology. Alternatively, an external device such as a smart phone of the hearing aid system may be used as gateway for the hearing aid, all of which will be well known for the skilled person.

[0036] According to a second step 102 a first plurality of evaluated hearing aid parameter settings, each associated with a specific hearing aid system user, are provided to said at least one server.

[0037] Thus, in the present context a hearing aid parameter setting represents a set of selected values one for each of a corresponding set of parameters. According to a variation, the provided hearing aid parameter settings only represents a sub-set of all the parameters required to operate the hearing aid system.

[0038] According to a specific advantage of the present invention, the parameters, whose selected (i.e. preferred) values are provided to said at least one server, have been carefully selected due to their ability to represent general trends for all hearing aid system users. One example of such a set of parameters is a set of fine-tuning gains to be added or subtracted in a corresponding set of frequency bands. In more specific variations the number of frequency bands is three or four. However, more frequency bands such as between 10 and 20 may also be considered dependent primarily of the available processing power.

[0039] Basically, any type of hearing aid parameter is suitable for being adjusted in accordance with this method, thus e.g. noise reduction algorithms, beam forming algorithms, and compressor settings may be improved.

[0040] According to a more specific embodiment an evaluated hearing aid parameter setting may be obtained using the optimization method disclosed in WO-A1-2016004983 with the title “Method of optimizing parameters in a hearing aid system and a hearing aid system” and by the same applicant and which is hereby incorporated by reference. More specifically reference may be given to page 20, lines 15-27, which describes criteria for considering a hearing aid parameter setting to be preferred and therefore to be stored in a hearing aid. However, according to another specific variation, a hearing aid parameter setting may be considered evaluated already if the predicted internal response function (that may also be denoted the preference function) for a given hearing aid parameter setting is estimated with sufficient precision by said optimization method.

[0041] It is emphasized though that the present invention is generally independent on the specific method used to provide evaluated hearing aid parameter settings. More specifically the present invention is independent on whether said specific method to provide evaluated hearing aid parameter settings is probabilistic or not and independent on whether said specific method is parameterized or not.

[0042] According to yet another variation a hearing aid parameter setting is considered to have been evaluated if it fulfils at least one of: the setting has been stored in the hearing aid system, the setting has been rated in an in-situ comparison between two different hearing aid parameter settings and the value of a user's internal response function for the hearing aid parameter setting has been estimated with sufficient precision.

[0043] According to a third step 103 said first plurality of evaluated hearing aid parameter settings and their association with a specific hearing aid system user is used to predict a new hearing aid parameter setting adapted to replace said first hearing aid parameter setting for the current user.

[0044] Consider now an embodiment where a hearing aid system user wants to determine fine-tuning gains for say 3 frequency bands with a resolution of say 1 dB and an adjustment range of say +/−10 dB, which provides say 10 000 different hearing aid parameter settings that in the following for simplicity reasons may be denoted items.

[0045] Now, the connection between hearing aid system users (which in the following may simply be denoted users) and hearing aid parameter settings (i.e. items) can be represented as a matrix

[00001]$\underset{\_}{\underset{\_}{X}}=\begin{array}{c}\begin{array}{c}{\mathrm{user}}_1\\ {\mathrm{user}}_2\end{array}\\ {\mathrm{user}}_3\end{array}[\begin{array}{ccccc}{\mathrm{item}}_1& {\mathrm{item}}_2& {\mathrm{item}}_3& .Math.& \\ 1& & & .Math.& \\ & 1& 1& .Math.& \\ 1& 1& & .Math.& 1\end{array}],$

where a value of 1 at row 1, column 1 means that user.sub.1 has saved the hearing aid parameter setting represented by item.sub.1 in her hearing aid system at some point in time. Hearing aid parameters settings that have not been stored in the hearing aid system of a given user will, according to this specific variation, be assigned a value of zero. According to an alternative embodiment, the values in the matrix X (that in the following may be denoted ratings matrix) may also represent ratings in an absolute scale, which according to one example can be achieved by translating pairwise comparisons performed within a fine-tuning optimization algorithm to an absolute scale. According to a more specific example (using WO-A1-2016004983 also mentioned above) this translation can be performed by calculating the Gaussian Process mean for the pairwise evaluated settings to get scalar values, which are then interpreted as the user's rating of the hearing aid parameter setting. Therefore according to this embodiment the matrix X will contain ratings for the hearing aid parameter settings that the user has encountered while carrying out the fine-tuning optimization and zeros for unseen settings:

[00002]$\underset{\_}{\underset{\_}{X}}=\begin{array}{c}\begin{array}{c}{\mathrm{user}}_1\\ {\mathrm{user}}_2\end{array}\\ {\mathrm{user}}_3\end{array}[\begin{array}{ccccc}{\mathrm{item}}_1& {\mathrm{item}}_2& {\mathrm{item}}_3& .Math.& \\ 7.9& & & .Math.& 4.8\\ & 2.2& 5& .Math.& \\ 9.1& 3.4& & .Math.& 1\end{array}],$

[0046] Subsequently, the process of recommending a new hearing aid parameter setting to a current user can according to an embodiment be based on the nearest neighbor algorithm.

[0047] The first step of this algorithm is to measure the similarity of users based on their user preference vectors (i.e. the rows in the ratings matrix X) by determining a distance measure between the user preference vectors. According to variations the distance measure can be a Pearson correlation or cosine distance, but other distance measures may also be used. The next step is to choose the k nearest neighbors given the distance between user preference vectors. Finally, the predicted new setting (i.e. the best setting) for the current user can be determined by calculating e.g. a mean or medoid user preference vector for the k nearest neighbors and selecting as the new setting the setting that has the highest rating in the mean user preference vector.

[0048] However, the nearest neighbor algorithm may be considered disadvantageous if too many hearing aid parameter settings has not been rated, because the algorithm requires that the overlap of item ratings for each pair of users must have some minimal size. Consequently, according to another embodiment, in order to relieve this data sparsity problem matrix factorization techniques can be used. Matrix factorization algorithms work by decomposing the ratings matrix into the product of two lower dimensionality rectangular matrices, i.e.:

X.sub.N×I≈U.sub.N×FV.sub.F×I

wherein the U matrix, that in the following may be denoted user matrix, has dimensions N×F where the N matrix rows represent the number of hearing aid system users and the F matrix columns represent latent factors and wherein the V matrix, that in the following may be denoted settings matrix, has dimensions F×I where the F matrix rows represent the selected latent factors and where the I columns represent the hearing aid parameter settings.

[0049] The number of latent factors, F, must be chosen based on the specific application, but generally the latent factors provide a representation of the hearing aid parameter setting ratings of all considered users.

[0050] Given the two lower dimensionality matrices a rating for any given hearing aid parameter setting, even a previously un-evaluated setting for a user (such as the current user) can therefore be estimated as:

[00003]${\tilde{x}}_{n,i}=\underset{f=0}{\overset{F}{.Math.}}{u}_{n,f}{v}_{f,i}$

[0051] According to a more specific embodiment the matrix decomposition is carried out using non-negative matrix factorization, which requires that all the elements of both the ratings matrix and said two lower dimensionality matrices are non-negative.

[0052] According to a fourth step 104 the new hearing aid parameter setting is provided to the hearing aid system and used instead of the previous (i.e. the first) hearing aid parameter setting.

[0053] According to a variation of the present embodiment the method is adapted to additionally prompt the current user to evaluate said new hearing aid parameter setting and providing said evaluation to said at least one server. The evaluation may comprise an acceptance or rejection of the new setting or a comparison of the new setting with the previous setting.

[0054] This variation is especially advantageous because it provides that the disclosed methods used to predict a new hearing aid parameter setting receives feedback that can be used to improve performance. According to a specific variation this is achieved by carrying out a matrix factorization with the new data. According to another even more specific variation the difference between the estimated rating of the new setting and the user's actual evaluation determines whether the model behind the method of predicting a new setting needs to be updated, as one example by carrying out the matrix factorization with the new data.

[0055] According to another variation of the present embodiment the method is adapted to provide, to said at least one server, a plurality of additional data, each associated with a specific hearing aid system user, representing at least one characteristic selected from a group comprising age, gender, race, audiogram, type of hearing aid system worn, experience with wearing a hearing aid, native language and country of residence, and using said additional data to contribute to predicting a new hearing aid parameter setting by adding the plurality of additional data to the first plurality of evaluated hearing aid parameter settings. Hereby improved prediction may be achieved especially for a new hearing aid system user that have not yet evaluated any or only few hearing aid parameter settings.

[0056] According to a more specific variation a function capable of providing a new row (i.e. a new hearing aid system user) in the user matrix as a function of the associated additional data may be derived using the additional data for the other hearing aid system users. According to an even more specific variation said new row may be provided based on cluster analysis.

[0057] According to yet another embodiment improved predictions of a new best setting for a specific situation may be achieved by only considering hearing aid parameter settings that have been associated with said situation, wherein said situation is selected from a group of situations comprising an identified sound environment, an identified geographical location and a specific cognitive state of the hearing aid system user. According to another embodiment data representing at least one of said specific situations is associated with a corresponding hearing aid parameter setting and incorporated in the items used to construct the ratings matrix, such that each item no longer consists only of hearing aid parameter settings but also includes information identifying a specific situation. This embodiment is particularly advantageous because it enables the use of Matrix factorization methods to predict a preferred hearing aid parameter setting for a specific situation that a current user experiences for the first time.

[0058] Reference is now made to FIG. 2, which illustrates highly schematically an in-situ fitting system 200 according to a second embodiment of the invention.

[0059] The in-situ fitting system 200 comprises a hearing aid system 201 consisting of a left hearing aid 202-a and a right hearing aid 202-b and an external device, e.g. in the form of a smart phone 203 with a specific software application installed. Furthermore the in-situ fitting system 200 comprises an internet server 204 that is adapted to receive, over the internet, a plurality of evaluated hearing aid parameter settings, and adapted to transmit a new hearing aid parameter setting to said hearing aid system 201 in response to a trigger event.

[0060] In obvious variations the hearing aid system may consist of a single hearing aid (a so called monaural fitting) or may consist of both a left and a right hearing aid (a so called binaural fitting) and furthermore the hearing aid system may (or may not) include an external device 203.

[0061] According to another embodiment the in-situ fitting system does not comprise the hearing aids 202-a and 202-b, instead the in-situ fitting system is operationally connected with the hearing aids 202-a and 202-b either directly from the internet server or through the external device 203 that may operate as a gateway.

[0062] It is noted that the present invention does not require neither the use of probabilistic methods nor non-parameterized methods to provide evaluated hearing aid parameter settings, although these methods are generally preferred because they are more efficient than the alternative methods.

[0063] It is likewise noted that the present invention is independent on whether the parameters to be optimized are used to control how sound is processed in the hearing aid system or whether they are used to control how sound is synthetically generated by the hearing aid system.

[0064] The present invention is also independent on how the hearing aid system parameters are provided or offered or selected for optimization.

[0065] Generally, disclosed variations may be combined with all other disclosed variations unless the opposite is specifically mentioned.