In-Situ Hearing Aid Fitting With Server-Based Preference Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing in-situ fitting systems for hearing aids face challenges in efficiently personalizing settings to meet individual user preferences, particularly due to high processing and memory requirements, and the difficulty in accurately modeling user preferences.
Innovation Solution
The method involves an in-situ fitting system that uses a server connected to the hearing aid system to adapt settings based on evaluated parameter settings associated with specific users, employing techniques like matrix factorization to predict new settings and improve user satisfaction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional parameterized approaches are used to model user preferences, then the system can attempt to capture user preferences, but the processing and memory requirements become very high
Solution Approach 1:
The patent extracts the complex preference modeling computations from the hearing aid device itself and relocates them to an external server. The hearing aid only retains minimal local functionality for data collection and basic operations, while the server handles the computationally intensive matrix factorization and preference prediction algorithms.
Solution Approach 2:
The patent introduces an external server as an intermediary between the hearing aid and the user preference modeling process. This server acts as a mediator that receives data from multiple hearing aids, performs centralized computations using matrix factorization techniques, and returns optimized parameters to individual devices.
2Adaptability or versatility
If comprehensive personalization is implemented, then user satisfaction improves, but the time and complexity of the fitting process increases
Solution Approach 1:
The patent performs preliminary preference modeling by collecting and analyzing data from multiple users beforehand. The server builds preference models in advance using matrix factorization on aggregated data, so that when a new user or situation arises, the system can quickly retrieve and apply pre-computed parameters rather than performing full optimization in real-time.
Solution Approach 2:
The patent implements feedback loops where user responses to presented parameters are collected and used to refine preference models. The system presents parameter sets to users, captures their preferences or corrections, and uses this feedback to update the matrix factorization models, continuously improving personalization accuracy over time.
3Measurement precision
If detailed user preference modeling is attempted, then personalization accuracy improves, but the difficulty of detecting and measuring user preferences increases
Solution Approach 1:
The patent creates universal preference models that capture common patterns across multiple users through matrix factorization. By analyzing aggregated data from many users, the system identifies universal preferences and behaviors that can be applied to individuals, reducing the need for extensive individual testing while maintaining accuracy.
Solution Approach 2:
The patent uses partial information from multiple users to infer complete preference profiles. Rather than requiring exhaustive data from each individual, the system collects preferences from many users and uses matrix factorization to reconstruct individual preference models from partial data sets, achieving accurate personalization with less direct measurement per user.
Data Source
AI summary
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.