Method for Selecting a Fitting Agent and Determining Hearing Device Parameters

Abstract

The present invention discloses a fitting agent for a hearing device and related methods, wherein the fitting agent is configured to: initialize a user model including a user preference function; obtain main test settings of the hearing device; obtain secondary test settings of the hearing device; present the main test settings and the secondary test settings to a user; detect a user input of a preferred test setting indicating a preference for the main test setting or the secondary test setting; and update the user model based on hearing device parameters of the preferred test setting, wherein obtaining the secondary test settings includes: obtaining a candidate set of candidate test settings; determining an uncertainty parameter for each candidate test setting; and selecting the secondary test setting from the candidate set of candidate test settings based on the uncertainty parameter of the candidate test setting.

Description

Technical Field

[0001] The present disclosure relates to hearing devices and, in particular, to related tools, methods, and systems for determining, tuning, fitting, and optimizing one or more of the parameters of a hearing device. Accordingly, a fitting agent and related methods are provided, specifically a method for tuning the parameters of a hearing device. Background Art

[0002] The fitting and tuning of hearing devices or hearing aids has always been considered a cumbersome task for healthcare professionals (HCPs). Traditional methods for fitting the parameters of hearing devices rely on compensating for a user's hearing loss based on an audiogram by applying rules such as NAL-NL1 or NAL-NL2. However, these rules do not take into account specific user preferences.

[0003] Recent methods involve preference learning for hearing devices.

[0004] EP 3493555 A1 relates to a method and a hearing device for tuning the hearing device parameters of a hearing device. The method includes: initializing a model; obtaining an initial test setting defined by one or more initial test hearing device parameters; assigning the initial test setting as a primary test setting; obtaining a secondary test setting based on the model; outputting a primary test signal according to the primary test setting; outputting a secondary test signal according to the secondary test setting; detecting a user input of a preferred test setting; updating the model based on the primary test setting, the secondary test setting, and the preferred test setting; and updating the hearing device parameters of the hearing device based on the hearing device parameters of the preferred test setting according to a determination that a tuning criterion is met. Summary of the Invention

[0005] There are still challenges in terms of improved tools, methods, and devices that can improve the fitting and tuning of hearing device parameters.

[0006] The present invention discloses a fitting agent, for example for optimizing, determining, fitting, tuning and modeling one or more of the hearing device parameters of a hearing device, wherein the fitting agent is configured to: initialize a user model including a user preference function and / or a user response distribution; obtain a primary test setting of the hearing device; obtain a secondary test setting of the hearing device; present the primary test setting and the secondary test setting to a user; detect a user input of a preferred test setting indicating a preference for the primary test setting or the secondary test setting; and update the user model based on the hearing device parameters of the preferred test setting. Optionally, obtaining the secondary test setting includes obtaining a candidate set of candidate test settings; determining an uncertainty parameter for each candidate test setting; and selecting the secondary test setting from the candidate set of candidate test settings based on the uncertainty parameter of the candidate test setting.

[0007] There is also disclosed a method for optimizing, determining, fitting, tuning and modeling, such as determining the hearing device parameters of a hearing device, wherein the method includes: initialize a user model including a user preference function and / or a user response distribution; obtain a primary test setting of the hearing device; obtain a secondary test setting of the hearing device; present the primary test setting and the secondary test setting to a user; detect a user input of a preferred test setting indicating a preference for the primary test setting or the secondary test setting; and update the user model based on the hearing device parameters of the preferred test setting. In the method, optionally, obtaining the secondary test setting includes: obtaining a candidate set of candidate test settings; determining an uncertainty parameter for each candidate test setting; and selecting the secondary test setting from the candidate set of candidate test settings based on the uncertainty parameter of the candidate test setting.

[0008] An advantage of the present invention is to provide a simple and effective way to configure one or more hearing device parameters of a hearing device. Additionally, the present disclosure improves the modeling of the hearing parameter settings preferred by the user, and further enables the application of optimized settings in the hearing device to provide an improved hearing experience to the user.

[0009] Advantageously, the present invention utilizes the information provided in the user response to guide the learning process. The present disclosure provides a metric that can be effectively used to evaluate the learning performance in the absence of true user preference information, thereby allowing optimized user feedback and model optimization.

[0010] The present invention provides an efficient automatic search for optimal hearing device parameters by incorporating user feedback into the learning cycle. The present invention provides a fitting agent, device, and method that allow an authorized user to make a direct decision and have a direct impact on the fitting and / or tuning process to learn user preferences for hearing device parameters in an effective and minimally obtrusive manner.

[0011] In addition, the present invention allows hearing device parameters to be configured (such as fitted and / or tuned) during normal operating conditions and / or with a small amount of user input / interaction. Thus, a simple and smooth user experience of the hearing device is provided. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above and other features and advantages of the present invention will become apparent to those skilled in the art from the following detailed description of its exemplary embodiments with reference to the accompanying drawings, in which:

[0013] Figure 1 A hearing system according to the present disclosure is schematically shown,

[0014] Figure 2 The interaction between the user and the fitting agent is shown, and

[0015] Figures 3A to 3B is a flowchart of an exemplary method according to the present disclosure. DETAILED DESCRIPTION

[0016] In the following, various exemplary embodiments and details are described with reference to the relevant drawings. It should be noted that the drawings may or may not be drawn to scale, and elements of similar structure or function are denoted by the same reference numerals throughout the drawings. It should also be noted that the drawings are only intended to facilitate the description of the embodiments. They are not intended as an exhaustive description of the present invention or a limitation on the scope of the present invention. In addition, the embodiments shown do not necessarily show all aspects or advantages. Aspects or advantages described in connection with a particular embodiment are not necessarily limited to that embodiment and may also be practiced in any other embodiment, even if not so shown or if not so explicitly described.

[0017] A fitting agent is disclosed. The fitting agent or at least its first part may be implemented, for example, as an application in an accessory device (such as an electronic device). The accessory device includes an interface, a processor, and a memory. The accessory device may be, for example, or include a mobile phone such as a smart phone, a smart watch, a dedicated device, a computer (such as a laptop or a PC), or a tablet computer. The fitting agent or at least its second part may be implemented in a server device. The fitting agent or at least its third part may be implemented in a hearing device.

[0018] The present disclosure relates to a hearing system, a fitting agent, an accessory device, a hearing device of the hearing system, and related methods. The accessory device forms a connection between the accessory device and the hearing device. The accessory device is typically paired or wirelessly coupled to the hearing device. The hearing device may be a hearing aid, such as a behind-the-ear (BTE) type, in-the-ear (ITE) type, in-the-canal (ITC) type, receiver-in-the-canal (RIC) type, or receiver-in-the-ear (RITE) type hearing aid. The hearing device may be audible, such as a pair of earbuds or headphones. Typically, the hearing device system is owned and controlled by the hearing device user.

[0019] The hearing system may include a server device and / or a fitting device. The accessory device is controlled by a dispenser and is configured to determine configuration data, such as fitting parameters. The server device may be controlled by a hearing device manufacturer.

[0020] The fitting agent is configured to receive user input indicating a preferred test setting that detects a preference for the primary test setting or the secondary test setting. Thus, the hearing device and / or the accessory device implementing the fitting agent or at least a part of the fitting agent may include one or more user interfaces for receiving and / or detecting user input. For example, the hearing device may include a user interface for receiving user input. The user interface of the hearing device may include one or more buttons, an accelerometer, and / or a voice control unit. The accessory device may include a user interface. The user interface of the accessory device may include a touch-sensitive surface, such as a touch display, and / or one or more buttons. The user interface of the accessory device may include a voice control unit. The user interface of the hearing device may include one or more physical sliders, knobs, and / or buttons. The user interface of the accessory device may include one or more physical or virtual (on-screen) sliders, knobs, and / or buttons.

[0021] The method or at least a part of it may be executed in the hearing device. The method or at least a part of it may be executed in one accessory device or multiple accessory devices, such as in a smart phone optionally combined with a smart watch. Executing a part of the method in the accessory device (such as a smart phone optionally combined with a smart watch) may be advantageous in providing a smoother user input and user experience. Additionally, from the perspective of the hearing device, executing a part of the method in the accessory device may be advantageous in providing a more efficient method. The method or at least a part of it may be executed in the server device and / or the fitting device.

[0022] The present disclosure relates to a fitting agent for a hearing device, and more particularly to a fitting agent for optimizing, determining, fitting, and tuning one or more of the hearing device parameters of a hearing device.

[0023] The fitting agent is configured to initialize a user model that includes a user preference function. The user model may include a user response function, such as including a user response distribution.

[0024] When comparing two sets of hearing device parameter settings, the user model optionally represents a probabilistic description of the user response. Components of the model include a user preference function and a distribution of user responses to the presented parameter choices.

[0025] A vector of hearing device parameters is defined on an M - dimensional continuous and dense surface. In particular, the hearing device parameter x is optionally defined on an M - dimensional hypercube, i.e., x ∈ [0, 1] M . In one or more exemplary fitting agents / methods, the hearing device parameters can be normalized by their physical range. The fitting agent / method is configured to find an optimized / improved value of the hearing device parameters, also denoted as θ for a particular user. The term M for the hearing device parameters can be 1 and / or less than 100, such as in the range of 10 to 50. The number M of hearing device parameters can be greater than 20, such as in the range of 25 to 75.

[0026] Generally, the user preference function is unknown. The user preference function f(x; θ, Λ) can be a parametric function of the hearing device parameter x ∈ [0, 1] M with a known form but an unknown shape. The shape is optionally characterized by fitting or adjusting the parameter θ ∈ [0, 1] M and the scaling matrix Λ. The scaling matrix Λ can be a positive definite scaling matrix Λ. The scaling matrix can be a diagonal matrix

[0027] In one or more exemplary fitting agents / methods, a distribution prior can be applied to each element of Λ. For example, a log - normal distribution prior can be applied to each element of Λ, for example or a gamma distribution prior can be applied to each element of Λ, such as λ m ~Gamma(α m , β m ).

[0028] The scaling matrix Λ does not need to be a diagonal matrix. The scaling matrix Λ can be chosen as Λ = L′*L, where L is a lower triangular matrix (also known as the Cholesky decomposition of Λ). A Gaussian prior can be applied to each element of L, for example

[0029] The user preference function can be denoted as f or f(x; θ, Λ). The user preference function f(x; θ, Λ) can be given by:

[0030] f(x; θ, Λ) = -((x - θ) T Λ(x - θ)) s, (1)

[0031] where x is an M-dimensional vector optionally in the hypercube [0, 1] M and the vector represents (M) hearing device parameters of a device, θ is the maximizing argument of f, Λ is a positive definite M×M scaling matrix that characterizes a user's sensitivity to changes in hearing device parameters, M is an integer, and s is a real-valued exponent. The real-valued exponent s can be in the range from 0.01 to 0.99. The real-valued exponent s can be less than 0.5, such as in the range from 0.01 to 0.45. The real-valued exponent s can be greater than 0.5, such as in the range from 0.55 to 0.99.

[0032] The maximizing parameter θ may be subject to the following prior assumptions on the objective function f θ，Λ as follows:

[0033]

[0034] where is the cumulative density function of a probability distribution (such as a standard normal distribution), and is a sample from another probability distribution.

[0035] In one or more exemplary fitting agents / methods, the maximizing parameter θ may be subject to the following prior assumptions on the objective function f θ，Λ as follows:

[0036] where

[0037] where is the cumulative density function of a standard normal distribution, and is a sample from a normal distribution with mean vector μ and covariance matrix Σ. The values of the mean and covariance can be learned from user responses.

[0038] In one or more exemplary fitting agents / methods, the scaling matrix can be a transformed scaling matrix, for example, by applying a logarithmic transformation to the elements of the scaling matrix.

[0039] In one or more exemplary fitting agents / methods, the real-valued exponent s can be in the range from 0.25 to 0.75, such as 0.5.

[0040] Thus, the user preference function can be given as:

[0041]

[0042] The tuning parameter corresponds to the optimal location of the function θ, which is the optimal hearing device parameter for the user; and the range Λ around this optimum, which respectively characterizes the user's sensitivity to changes in the hearing device parameters. The tuning parameter is user-specific and needs to be learned to provide an optimal or improved user experience. This form of the user preference function has attractive properties that allow for a rapid reduction of the search space.

[0043] The user model is optionally based on a first hypothesis and a second hypothesis. The first hypothesis may be that the user preference function is a unimodal preference function representing the user's preference. The second hypothesis may be that the user may be uncertain about the preference, which can be represented by an additive random variable. A value function u(·) can be defined to model the user preference uncertainty. The value function u(·) can be defined as

[0044] u(x) = f(x; θ, Λ) + ε, (5)

[0045] where the user uncertainty error ε is optionally assumed to have a Gaussian distribution, taking care of, e.g., a standard Gaussian distribution, such as

[0046]

[0047] The variance of the Gaussian distribution can be a real value, e.g., in the range from 0 to 1 or greater than 1.

[0048] The user's trial definition includes a pair of test settings of a primary test setting and a secondary test setting. User responses for the pairwise comparison of test settings by the user are obtained from the trial, which is defined by a pair of test settings {x_ref, x_alt} of the test settings, where the primary test setting x_ref and the secondary test setting x_alt are respectively the so-called reference and alternative parameter suggestions. In other words, a trial T_n is performed, where each trial T_n includes a primary test setting a secondary test setting and a preferred test setting r n · or is defined by it. The index n - 1 refers to the previous trial T_n-1t.

[0049] The optional agent can be configured to present the primary test settings and the secondary test settings to the user. Presenting the primary test settings and the secondary test settings to the user can include transferring the primary test settings and the secondary test settings from an accessory device to the hearing device. Presenting the primary test settings and the secondary test settings to the user can include outputting a primary test signal according to the primary test settings, and / or outputting a secondary test signal according to the secondary test settings. In other words, the optional agent is optionally configured to output a primary test signal according to the primary test settings and output a secondary test signal according to the secondary test settings. The optional agent is configured to detect a user input of a preferred test setting indicating a preference for the primary test setting or the secondary test setting. The primary test setting x_ref is optionally initiated or obtained as the highest-rated recommendation / test setting from a previous trial, and this is challenged by the secondary test setting x_alt in a given trial. Depending on the corresponding test settings (hearing device parameters), the user response depends on the user's preference for the presented test signal. The user's preference for one or more hearing device parameters is given or quantified by a value function. The user input or user response R can take or be assigned a value in {0,1}, where a response of "1" to a given trial {x_ref,x_alt} means that the secondary test setting or hearing parameter x_alt is preferred to or as good as the primary test setting or hearing device parameter x_ref, otherwise 0, i.e.

[0050]

[0051] where ≥ means preferred or equivalent to, and > means strictly preferred.

[0052] The probability of a positive user response, i.e. R taking the value "1" in the trial {x_ref,x_alt}, is given by:

[0053]

[0054] where, Φ(·) is the cumulative distribution function (CDF) of the standard Gaussian distribution, and Note that x_ref and x can be used interchangeably herein ref to represent the primary test settings, and x_alt and x can be used interchangeably herein alt to represent the secondary test settings.

[0055] Therefore, R is a random variable with parameters Φ(f(x alt ; θ, Λ′)-f(x ref; a Bernoulli random variable with parameters (θ, Λ′)). Thus, given the user preference function and user response model defined by (1) and (4)-(8) above, the probability mass function (PMF) of the user response r to the trial {x_ref, x_alt} is given by:

[0056]

[0057] This user response model that associates binary observations with a continuous latent function (the user preference function) is also known as the Thurstone-Mosteller law of comparative judgment. In statistics, it is known as a binomial-probit regression model.

[0058] The fitting agent is configured to obtain a primary test setting of the hearing device (also denoted as x_ref or x ref ). The primary test setting x_ref is a vector including M hearing device parameters of the hearing device. The hearing device parameters may include one or more of filter coefficients, compressor settings, gains, or other parameters related to operations or signal processing in the hearing device.

[0059] The fitting agent is configured to obtain a secondary test setting of the hearing device (also denoted as x_alt or x alt ). The secondary test setting x_ref is a vector including M hearing device parameters of the hearing device. The hearing device parameters may include one or more of filter coefficients, compressor settings, gains, or other parameters related to operations or signal processing in the hearing device.

[0060] The fitting agent is configured to present the primary test setting and the secondary test setting to the user. Presenting the primary test setting and the secondary test setting to the user may optionally include outputting a primary test signal according to the primary test setting. Presenting the primary test setting and the secondary test setting to the user may optionally include generating a primary test signal in an accessory device according to the primary test setting, and streaming the primary test signal from the accessory device to the hearing device. Presenting the primary test setting and the secondary test setting to the user may optionally include transmitting a control signal indicating the primary test signal / primary test setting from the accessory device to the hearing device. The control signal may include the primary test setting. Presenting the primary test setting and the secondary test setting to the user may include generating the primary test signal in the hearing device according to the control signal (e.g., based on the primary test setting of the control signal).

[0061] Presenting the primary test settings and the secondary test settings to the user optionally includes outputting a secondary test signal according to the secondary test settings. Presenting the primary test settings and the secondary test settings to the user optionally includes generating a secondary test signal in an accessory device according to the secondary test settings, and streaming the secondary test signal from the accessory device to the hearing device. Presenting the primary test settings and the secondary test settings to the user optionally includes transmitting a control signal indicating the secondary test signal / secondary test settings from the accessory device to the hearing device. The control signal may include the secondary test settings. Presenting the primary test settings and the secondary test settings to the user may include generating the secondary test signal in the hearing device according to the control signal (e.g., the secondary test settings based on the control signal).

[0062] The fitting agent is configured to detect a user input of a preferred test setting indicating a preference for the primary test setting or the secondary test setting. In the fitting agent, detecting a user input of a preferred test setting indicating a preference for the primary test setting or the secondary test setting may include prompting the user for the user input, e.g., via a beep signal or a voice signal from the hearing device and / or a visual, tactile, and / or audio cue from the accessory device. Detecting the user input may be performed on the hearing device, e.g., by the user activating a button and / or an accelerometer in the hearing device (e.g., single or double tapping the hearing device housing). Detecting the user input may be performed on the accessory device, e.g., by the user selecting a user interface element representing the preferred test setting, e.g., on a touch-sensitive display of the user accessory device.

[0063] For an initial trial, the primary test settings and the secondary test settings may be initialized randomly. Multiple trials may be performed after the initial trial, where each trial includes obtaining test settings x_ref, x_alt, outputting a test signal, and detecting the preferred test setting.

[0064] The fitting agent is configured to update the user model based on the hearing device parameters of the preferred test setting. Updating the user model may include updating the user preference function based on one or more of the primary test settings, the secondary test settings, and the user input of the preferred test setting. In other words, the user preference function may be updated based on the test results including the primary test settings, the secondary test settings, and the preferred test settings of the primary test settings and the secondary test settings.

[0065] In one or more exemplary fitting agents and / or methods, updating the user model may be based on Bayesian inference. Updating the model may include updating one or more of the parameters of the user preference function and / or the user response model.

[0066] In a fitting agent and / or method, obtaining secondary test settings may include: obtaining a candidate set of candidate test settings; determining an uncertainty parameter (also denoted as UP) for each candidate test setting; and optionally selecting the secondary test settings from the candidate set of candidate test settings based on the uncertainty parameters of the candidate test settings.

[0067] The candidate set of candidate test settings (also denoted as CTS_set) may include a number N_C of candidate test settings (such as a plurality of candidate test settings). The candidate test settings are also denoted as CTS. The number N_C of candidate test settings may be at least 3. In one or more exemplary fitting agents / methods, the number N_C of candidate test settings may be at least 5 or even greater than 20. In other words, the candidate set of candidate test settings includes candidate test settings CTS_i, i = 1, 2,..., N_C, where i is the index of the candidate test settings in the candidate set. The number N_C of candidate test settings may depend on the number M of hearing device parameters, for example N_C may be greater than 3M, greater than 4M; or greater than 5M. The candidate test setting CTS_i is a vector including M hearing device parameters of the hearing device. The hearing device parameters of the candidate test setting CTS_i may include one or more of filter coefficients, compressor settings, gains, or other parameters related to operations or signal processing in the hearing device.

[0068] In one or more exemplary fitting agents / methods, determining the uncertainty parameter for each candidate test setting may include determining the mutual information I between the primary test setting and the corresponding candidate test setting. In other words, UP_i may be determined for CTS_i for i = 1, 2,..., N_C and is given by UP_i = I(x_ref, CTS_i).

[0069] The uncertainty parameter UP_i may be given as:

[0070]

[0071] where x ref is the primary test setting of the current trial, is the user response to the previous trials T_1 to T_n-1.

[0072] In one or more exemplary fitting agents / methods, determining the uncertainty parameter for each candidate test setting includes applying a predictor to one or more (such as each) candidate test settings. In other words, the predictor may be applied to each of the candidate test settings CTS_i, i = 1, 2,..., N_C. The predictor, also denoted as P_U or P u may be a user response distribution. The predictor may be based on posterior updates based on previous user responses and trials.

[0073] In one or more exemplary fitting agents / methods, the predictor is a weighted distribution function.

[0074] For example, the predictor P u may be given by:

[0075]

[0076] where w is a weighting function / distribution and r n is the user response.

[0077] For example, the predictor P u may be a weighted distribution where the weighting function is a posterior distribution of one or more of, for example, θ, Λ. Thus, the predictor P u may be given by:

[0078]

[0079] In one or more exemplary fitting agents / methods, selecting the secondary test setting from the candidates of the candidate test settings includes selecting the candidate test setting having the maximum uncertainty parameter as the secondary test setting. In other words, the secondary test setting may be selected as a candidate test setting in the candidate set where the model has the highest uncertainty or where the user model is least accurate, thereby resulting in fast hearing device parameter optimization and / or learning for the user model with few user interactions.

[0080] In one or more exemplary fitting agents / methods, obtaining the candidate set of candidate test settings includes obtaining a first candidate test setting by sampling a posterior distribution (such as the posterior distribution of hearing device parameters of a user preference function). Obtaining the candidate set of candidate test settings may include obtaining the candidate test settings of the candidate set by sampling a posterior distribution (such as the posterior distribution of hearing device parameters of a user preference function).

[0081] Obtaining the candidate set of candidate test settings may include generating one or more candidate test settings based on the posterior distribution. The posterior distribution may be the marginal posterior distribution of the optimal hearing device parameters.

[0082] In one or more exemplary fitting agents, obtaining the candidate set of candidate test settings may include sampling from an exact posterior distribution, from a sample posterior distribution (e.g., based on Monte-Carlo simulation), or an approximate posterior distribution.

[0083] Obtaining the candidate set of candidate test settings may include generating candidate test settings according to the following steps:

[0084]

[0085] In other words, candidate test settings can be generated according to the marginal posterior distribution of the optimal hearing device parameters.

[0086] In one or more exemplary fitting agents / methods, the fitting agent is configured to: determine whether a stop criterion is met; and based on the determination that the stop criterion is met, update the hearing device parameters of the hearing device based on the updated user model. The stop criterion evaluates whether the user model is accurate enough to provide useful hearing device parameters. In other words, when it is determined that the user model provides a high enough degree of certainty of user preferences, the stop criterion can be met.

[0087] Updating the hearing device parameters based on the updated user model can include obtaining user preferences for primary or secondary test settings, and updating the posterior distribution based on the observed user responses and test settings. The hearing device parameters can then be determined using maximum a posteriori estimation or alternatively as the posterior mean or median. In other words, updating the hearing device parameters of the hearing device can include determining the hearing device parameters by applying one of maximum a posteriori estimation, posterior mean estimation, and posterior median estimation.

[0088] In one or more exemplary fitting agents / methods, the fitting agent is configured to update the hearing device parameters of the hearing device after each update of the user model.

[0089] Updating the hearing device parameters can include determining the updated hearing device parameters based on the maximum a posteriori estimation, the mean value or the median of the posterior of the user preference function.

[0090] In one or more exemplary fitting agents / methods, updating the hearing device parameters of the hearing device includes transmitting the updated hearing device parameters from an accessory device to the hearing device.

[0091] In one or more exemplary fitting agents / methods, the stop criterion is based on a divergence, such as a normalized and / or weighted information divergence. The stop criterion can be based on the Kullback-Leibler (KL) divergence.

[0092] In one or more exemplary fitting agents / methods, the KL divergence is given by:

[0093]

[0094] where w is the weight distribution over φ. D_n is the normalized weighted KL divergence and is the KL divergence given by:

[0095]

[0096] Where P is the true user response distribution, Q is the user distribution assigned by the candidate proxy (which may be P_U), and n is the number of trials performed.

[0097] The stopping criterion can be based on a divergence parameter, such as the KL divergence given above. The divergence parameter can indicate the absolute divergence and / or the change in the divergence between the previous user model and the updated user model, such as the difference (D n-1 -D n ) or the ratio (D n / D n-1 ). In other words, determining whether the stopping criterion is met can include determining the divergence parameter and optionally determining whether the divergence parameter meets the stopping criterion, e.g., whether the divergence parameter is less than a threshold or whether the divergence parameter is greater than a threshold. Thus, the stopping criterion can be based on one or more thresholds.

[0098] In one or more exemplary candidate proxy / methods, the stopping criterion is based on the number of trials performed. For example, when the candidate proxy has performed 10 trials, the stopping criterion can be met.

[0099] In one or more exemplary candidate proxy / methods, the candidate proxy is configured to update the predictor based on the hearing device parameters of the preferred test settings. In other words, updating the model can include updating the user response model based on the primary test settings, the secondary test settings, and the preferred test settings of the primary and secondary test settings.

[0100] It should be noted that the description and features of the candidate proxy function also apply to the method, and vice versa.

[0101] Figure 1 is an overview of a hearing system with a candidate proxy according to the present disclosure. The hearing system includes a hearing device 2, an accessory device 4, and an optional server device / candidate device 5. The hearing device 2 includes a transceiver module 6 for (wireless) communication with the accessory device 4 and an optional contralateral hearing device ( Figure 1 not shown in the figure). The transceiver module 6 includes an antenna 8 and a transceiver 10 and is configured to receive wireless signals and / or transmit wireless signals to the accessory device 4 via a wireless connection 11. The hearing device 2 includes: a set of one or more microphones, which includes a first microphone 12 for providing a first microphone input signal 14; a processor 16 for processing the input signal including the first microphone input signal 14 according to one or more hearing device parameters and providing an electrical output signal 18 based on the input signal; an optional user interface 20 connected to the processor 16; and a receiver 22 for converting the electrical output signal 18 into an audio output signal.

[0102] The accessory device 4 is a smart phone and includes a user interface 24, which includes a touch display 26, a processor (not shown), and a memory (not shown).

[0103] In the hearing system 1, the fitting agent 27 is an application program installed in the memory of the accessory device 4.

[0104] The fitting agent 27 is a fitting agent for optimizing, determining, fitting, tuning, and modeling one or more of the hearing device parameters of a hearing device. The fitting agent 27 is configured to initialize a user model f(x; θ, Λ) including a user preference function; obtain a primary test setting x_ref of the hearing device; obtain a secondary test setting x_alt of the hearing device; present the primary test setting x_ref and the secondary test setting x_alt to the user, for example via the wireless connection 11; detect a user input R of a preferred test setting indicating a preference for the primary test setting x_ref or the secondary test setting x_alt; and update the user model, such as the user preference function f(x; θ, Λ) and / or the user response model (user response distribution), based on the hearing device parameters of the preferred test setting, such as based on one or more of R, x_ref, and x_alt, such as all of them.

[0105] In the fitting agent 27, obtaining the secondary test setting includes obtaining a candidate set CTS_set of candidate test settings CTS_i, i = 1, 2,..., N_C, where N_C is the number of candidate test settings in the CTS_set; determining an uncertainty parameter UP_i and / or a certainty parameter CP_i for each candidate test setting CTS_i, i = 1, 2,..., N_C; and selecting the secondary test setting x_alt from the candidate set of candidate test settings based on the uncertainty parameter UP_i and / or the certainty parameter CP_i, i = 1, 2,..., N_C, of the candidate test settings. In the fitting agent 27, determining the uncertainty parameter UP_i, i = 1, 2,..., N_C, includes applying a predictor or a posteriori as a weighted distribution function for each candidate test setting.

[0106] The fitting agent 27 is configured to select the secondary test setting from the candidate set of candidate test settings by selecting the candidate test setting for which the user preference model has the least knowledge of the user's preference. In other words, the fitting agent 27 is optionally configured to select the secondary test setting from the candidate set of candidate test settings by selecting the candidate test setting with the largest uncertainty parameter as the secondary test setting. In one or more exemplary fitting agents, the fitting agent is configured to determine the certainty parameter of each candidate test setting and select the candidate test setting with the smallest certainty parameter as the secondary test setting.

[0107] The candidate selector 27 is optionally configured to obtain a candidate set of candidate test settings (such as one or more candidate test settings, e.g., including a first candidate test setting) by sampling a distribution (such as a posterior distribution). Thus, in the candidate selector 27, obtaining a candidate set of candidate test settings optionally includes obtaining the first candidate test setting and / or a plurality of first candidate test settings (e.g., at least 3 candidate test settings) by sampling the posterior distribution of the optimal parameters of the hearing device.

[0108] The candidate selector 27 is optionally configured to determine whether a stop criterion is met; and based on the determination that the stop criterion is met, update the hearing device parameters of the hearing device based on the updated user model, e.g., by transmitting the updated hearing device parameters to the hearing device via the wireless connection 11. The stop criterion is optionally based on a divergence, such as a normalized weighted information divergence, such as the Kullback-Leibler (KL) divergence. For example, the stop criterion may be met when the divergence is less than a threshold or when the divergence difference between the current trial and the previous trial is less than a threshold or when the rate between the divergence of the current trial and the previous trial is less than a threshold. In other words, if the divergence is greater than the threshold, the candidate selector 27 is optionally configured to perform a trial (determine and present a test setting, detect a user response, and update the user model). In one or more exemplary candidate selectors, the stop criterion may be met when the user provides a user input indicating a desire to stop optimization, e.g., by detecting a user selection of a stop virtual button (not shown) on the user interface 26 of the accessory device 4 and / or when a preset number of user inputs for the preferred test setting have been made.

[0109] The candidate selector 27 is optionally configured to update the predictor based on the hearing device parameters of the preferred test setting. Updating the predictor may include updating a weighting function and / or a user response distribution.

[0110] The candidate selector 27 is configured to perform user model update or hearing device parameter optimization based on the determination that a start criterion is met. The start criterion may be met if a user input indicating a desire to start hearing device parameter optimization has been detected on the user interface 20 or the user interface 24 (e.g., by activating a virtual start button 28 on the accessory device 4). In one or more exemplary candidate selectors (such as the candidate selector 27), the start criterion is based on the time since the last user model update. For example, the candidate selector 27 may be configured to perform hearing device parameter optimization at least every 3 months. The start criterion may be based on a combination of user input and the time since the last model update.

[0111] In a specific implementation including the accessory device 4, the fitting agent 27 / accessory device 4 can be configured to send a control signal 30 to the hearing device 2, and the control signal 30 indicates the primary test settings and the secondary test settings, so that the hearing device 2 can output test signals accordingly.

[0112] The hearing device 2 (processor 16) is optionally configured to output a primary test signal according to the primary test settings via the receiver 22, and output a secondary test signal according to the secondary test settings via the receiver 22.

[0113] The fitting agent 27 (the hearing device 2 (processor 16) and / or the accessory device 4) is configured to detect a user input of a preferred test setting indicating a preference for the primary test setting or the secondary test setting. For example, by detecting a user input on the user interface 20, or by detecting a user selection of one of the primary virtual button 32 (the primary test setting is preferred) and the secondary virtual button 34 (the secondary test setting is preferred) on the user interface 24 of the accessory device 4.

[0114] It should be noted that the fitting agent 27 can be configured to detect a user input of a preferred test setting indicating a preference for the primary test setting or the secondary test setting, for example, by receiving a wireless input signal from a secondary accessory device (such as a smart watch including a user interface). Thus, a more convenient user input is provided, which in turn increases the user-friendliness of the fitting agent.

[0115] The fitting agent 27 implemented in the accessory device 4 is configured to update the user model based on the primary test settings, the secondary test settings, and the preferred test settings. In other words, the fitting agent is configured to update the user model based on the hearing device parameters of the preferred test settings. Updating the user model may include updating the user preference model and the user response model based on the primary test settings, the secondary test settings, and the preferred test settings. The fitting agent 27 / accessory device 4 can be configured to transmit the primary test settings, the secondary test settings, and the preferred test settings to the server device 5, and the server device updates the user model and transmits the updated user model to the fitting agent 27 / accessory device 4. Therefore, the fitting agent 27 / accessory device 4 can be configured to receive the updated model from the server device 5. In other words, the fitting agent 27 can be distributed on one or more of the accessory device 4, the hearing device 2, and the server device 5. Therefore, the fitting agent 27 may include a first part 27A implemented in the accessory device 4, an optional second part 27B implemented in the server device, and an optional third part 27C implemented in the hearing device 2, such as in the processor 16.

[0116] In a specific implementation including the accessory device 4, the accessory device 4 is configured to send a control signal 32 to the hearing device 2, for example, according to the determination of meeting a stop criterion, and the control signal 38 indicates the hearing device parameters of a preferred test setting, so that the hearing device can update and apply the preferred hearing device parameters in the hearing device.

[0117] Figure 2 An interaction between a user 40 in an environment 41 and a fitting agent 27 is shown. For each interaction or trial indexed by n, the fitting agent generates a primary test setting and a secondary test setting and presents 42 test settings to the user and For example, by controlling the hearing device to output a primary test signal and a secondary test signal respectively according to and indicating the primary test setting and the secondary test setting. The user evaluates the two test settings and and the fitting agent 27 receives and detects a user response 44, R indicating a preferred test setting of the primary test setting and the secondary test setting n . The fitting agent updates the user model based on R n , and and generates the n+1th trial by setting the preferred test setting of the nth trial as the primary test setting of the n+1th trial determining the secondary test setting as described herein and presenting 46 test settings to the user and

[0118] Figures 3A to 3B is a flowchart of an exemplary method according to the present disclosure. The method 100 is a method for optimizing, determining, fitting, tuning, and modeling one or more of them, such as determining the hearing device parameters of a hearing device, wherein the method includes: initializing a user model (S102), which optionally includes a user preference function and a user response model; obtaining a primary test setting of the hearing device (S104), for example, where a fitting agent implemented in one or more of the hearing device 2, the accessory device 4, and the server device 5 is used; obtaining a secondary test setting of the hearing device (S106); presenting the primary test setting and the secondary test setting to the user (S107); detecting a user input of a preferred test setting indicating a preference for the primary test setting or the secondary test setting (S112); and updating the user model based on the hearing device parameters of the preferred test setting (S114). Updating the user model (S114) optionally includes updating the predictor / user response model of the user model (S120).

[0119] Method 100 optionally includes determining whether a stop criterion is met (S116), and updating the hearing device parameters of the hearing device according to the updated model (S118), for example, as described above with respect to the fitting agent. If it is determined that the user model has not been sufficiently updated (the stop criterion is not met), then method 100 optionally proceeds to S104, i.e., further trials and user input for the test setup are desired or required for accurate and precise modeling of the user preference function.

[0120] In method 100, obtaining a secondary test setup (S104) includes obtaining a candidate set of candidate test setups (S130), for example at least 5 or at least 20, (such as at least αN_C) candidate test setups, where N_C is the number of hearing device parameters to be optimized, and α is optionally greater than 3; determining the uncertainty parameter and / or the certainty parameter for each candidate test setup (S134); and selecting a secondary test setup from the candidate set of candidate test setups based on the uncertainty parameter and / or the certainty parameter of the candidate test setup (S138).

[0121] Determining the uncertainty parameter and / or the certainty parameter (S134) may include applying a predictor for each candidate test setup (S136), and / or selecting a secondary test setup from the candidate set of candidate test setups (S138) may include selecting the candidate test setup with the maximum uncertainty parameter or with the minimum certainty parameter as the secondary test setup (S140).

[0122] In method 100, presenting the primary test setup and the secondary test setup to the user (S107) optionally includes outputting a primary test signal according to the primary test setup (S108) and / or outputting a secondary test signal according to the secondary test setup (S110).

[0123] The use of the terms "first", "second", "third" and "fourth", "primary", "secondary", "third", etc. does not imply any particular order, but is included to identify a single element. Additionally, the use of the terms "first", "second", "third", "fourth", "primary", "secondary", "third", etc. does not denote any order or importance, but rather the terms "first", "second", "third", "fourth", "primary", "secondary", "third", etc. are used to distinguish one element from another. Note that the words "first", "second", "third" and "fourth", "primary", "secondary", "third", etc. are used here and elsewhere for labeling purposes only and are not intended to indicate any particular spatial or temporal order.

[0124] The memory can be one or more of a buffer, flash memory, hard disk drive, removable media, volatile memory, non-volatile memory, random access memory (RAM), or other suitable devices. In a typical arrangement, the memory can include non-volatile memory for long-term data storage and volatile memory that serves as the system memory for the processor. The memory can exchange data with the processor via a data bus. The memory can be considered a non-transitory computer-readable medium.

[0125] The memory can be configured to store information in a portion of the memory.

[0126] Furthermore, the marking of a first element does not imply the existence of a second element, and vice versa.

[0127] It is understood that Figure 1-3B Some modules or operations are shown with solid lines and some modules or operations are shown with dashed lines. The modules or operations included in the solid lines are the modules or operations included in the broadest exemplary embodiment. The modules or operations included in the dashed lines can be included in additional modules or operations, as part of additional modules or operations, or as exemplary embodiments of additional modules or operations, which can be taken in addition to the modules or operations of the solid-line exemplary embodiment. It should be understood that these operations do not need to be performed in the order presented. Furthermore, it should be understood that not all operations need to be performed. The exemplary operations can be performed in any order and in any combination.

[0128] Note that the word "comprising" does not necessarily exclude the existence of other elements or steps in addition to the listed elements or steps.

[0129] Note that the word "a" or "an" before an element does not exclude the existence of a plurality of such elements.

[0130] It should also be noted that any reference signs do not limit the scope of the claims, the exemplary embodiments can be implemented at least in part by both hardware and software, and several "means", "units", "devices" can be represented by the same item of hardware.

[0131] The various exemplary methods, devices, and systems described herein are described in the general context of method-step processes, which in one aspect may be implemented by a computer program product embodied in a computer-readable medium and including computer-executable instructions (such as program code) executed by a computer in a networked environment. The computer-readable medium may include removable and non-removable storage devices, including but not limited to read-only memory (ROM), random access memory (RAM), compact discs (CDs), digital versatile discs (DVDs), and the like. In general, program modules may include routines, programs, objects, components, data structures, etc. that perform specified tasks or implement particular abstract data types. The computer-executable instructions, associated data structures, and program modules represent examples of program code for performing the steps of the methods disclosed herein. A particular sequence of such executable instructions or associated data structures represents an example of corresponding actions for implementing the functions described in such steps or processes.

[0132] Although the features have been shown and described, it will be understood that they are not intended to limit the claimed invention, and it will be apparent to those skilled in the art that various changes and modifications can be made without departing from the spirit and scope of the claimed invention. Accordingly, the specification and drawings are to be regarded as illustrative rather than restrictive. The claimed invention is intended to cover all alternatives, modifications, and equivalents.

[0133] List of reference numerals

[0134] 1: Hearing device system

[0135] 2: Hearing device

[0136] 4: Accessory device

[0137] 5: Server device

[0138] 6: Transceiver module

[0139] 8: Antenna

[0140] 10: Transceiver

[0141] 11: Wireless connection between hearing device and accessory device

[0142] 11A: Wireless connection between accessory device and server device

[0143] 12: First microphone

[0144] 14: First microphone input signal

[0145] 16: Processor

[0146] 18: Electrical output signal

[0147] 20: User interface

[0148] 22: Receiver

[0149] 24: User interface of the accessory device

[0150] 26: Touch display

[0151] 27: Optional agent

[0152] 27A: The first part of the optional agent

[0153] 27B: The second part of the optional agent

[0154] 27C: The third part of the optional agent

[0155] 28: Start button

[0156] 30: Control signal indicating the primary and secondary test settings

[0157] 32: Primary virtual button

[0158] 34: Secondary virtual button

[0159] 38: Control signal indicating the hearing device parameters of the preferred test settings

[0160] 40: User

[0161] 42: The nth trial, test settings and

[0162] 44: User response to the nth trial

[0163] 46: The (n + 1)th trial, test settings and

[0164] 100: Method for determining the hearing device parameters of a hearing device

[0165] S102: Initialize the user model

[0166] S104: Obtain the primary test settings

[0167] S106: Obtain the secondary test settings

[0168] S107: Present the primary test settings and the secondary test settings to the user

[0169] S108: Output the primary test signal

[0170] S110: Output the secondary test signal

[0171] S112: Detect user input

[0172] S114: Update user model

[0173] S116: Determine whether the stop criterion is met

[0174] S118: Update hearing device parameters

[0175] S120: Update predictor

[0176] S130: Obtain candidate set

[0177] S132: Obtain the first candidate test setting

[0178] S134: Determine uncertainty parameter

[0179] S136: Apply the predictor to each candidate test setting

[0180] S138: Select secondary test setting

[0181] S140; Select the candidate test setting with the maximum uncertainty parameter

Claims

1. An optimization agent, implemented in one or more of an electronic device, a server device, and a hearing device, for optimizing hearing device parameters of the hearing device, wherein the optimization agent is configured to: Initialize a user model including a user preference function and a user response distribution; Obtain main test settings of the hearing device; Obtain secondary test settings of the hearing device; Present the main test settings and the secondary test settings to the user; Detect a user input of a preferred test setting indicating a preference for the main test setting or the secondary test setting; And Update the user model based on the hearing device parameters of the preferred test setting, characterized in that obtaining the secondary test settings includes: Obtain a candidate set of candidate test settings; Determine an uncertainty parameter for each candidate test setting; and Select the secondary test setting from the candidate set of candidate test settings based on the uncertainty parameter of the candidate test setting, wherein selecting the secondary test setting from the candidate set of candidate test settings includes: selecting the candidate test setting with the largest uncertainty parameter as the secondary test setting.

2. The matching agent according to claim 1, wherein, Determining an uncertainty parameter for each candidate test setting includes: applying a predictor to each candidate test setting.

3. The matching agent according to claim 2, wherein The predictor is a weighted distribution function.

4. The matching agent according to any one of claims 1 to 3, wherein Obtaining a candidate set of candidate test settings includes: obtaining a first candidate test setting by sampling a posterior distribution.

5. The matching agent according to any one of claims 1 to 4, wherein, The optimization agent is configured to determine whether a stop criterion is met; and Based on the determination that the stop criterion is met, update the hearing device parameters of the hearing device based on the updated user model.

6. The matching agent according to claim 5, wherein The stop criterion is based on a normalized weighted information divergence, such as the Kullback-Leibler (KL) divergence.

7. The matching agent according to claim 2, wherein The optimization agent is configured to update the predictor based on the hearing device parameters of the preferred test setting.

8. The matching agent according to any one of claims 1 to 7, wherein The candidate set of candidate test settings includes at least five candidate test settings.

9. The matching agent according to any one of claims 1 to 8, wherein The user model is based on a first hypothesis and a second hypothesis, wherein the first hypothesis is that the user preference function is a unimodal preference function representing user preferences, and the second hypothesis is that the user may be uncertain about the preferences.

10. The matching agent according to any one of claims 1 to 9, wherein, The user preference function f(x, Λ, Θ) is given by: f(x,Λ,Θ)=-((x-Θ) T Λ(x-Θ)) s , where x is an M-dimensional vector in the hypercube [0, 1], which represents M hearing device parameters of the device, Θ is the maximum parameter of f, Λ is a positive definite M×M scaling matrix, which characterizes the user's sensitivity to changes in the hearing device parameters, T represents the transpose of the vector x - Θ, where M is an integer, and s is a real-valued exponent in the range from 0.01 to 0.

99. M ​ 11. A method for determining hearing device parameters of a hearing device, wherein, The method includes: Initialize a user model including a user preference function; Obtain main test settings of the hearing device; Obtain secondary test settings of the hearing device; Present the main test settings and the secondary test settings to the user; Detect a user input of a preferred test setting indicating a preference for the main test setting or the secondary test setting; and Update the user model based on the hearing device parameters of the preferred test setting, characterized in that obtaining the secondary test settings includes: Obtain a candidate set of candidate test settings; Determine an uncertainty parameter for each candidate test setting; and Select the secondary test setting from the candidate set of candidate test settings based on the uncertainty parameter of the candidate test setting, Among them, selecting the secondary test setting from the candidate set of the candidate test settings includes: selecting the candidate test setting with the largest uncertainty parameter as the secondary test setting.

Images