Hearing Device Neural Networks With Mixed Data Types for Low-Power Processing

Solution Overview

Problem

Hearing devices implementing machine learning and deep neural networks (DNNs) are computationally costly and can negatively impact efficiency.

Innovation Solution

A hearing device that processes transducer input data using a neural network with different data types for network input and weights, enabling efficient computation by bypassing data type conversions and utilizing fixed point numbers for reduced computational cost and floating point numbers for improved precision and dynamic range.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If floating point numbers are used for neural network processing, then precision and dynamic range are improved, but computational cost and power consumption increase

Engineering Contradiction:

Improveaudio processing precisionVSAvoidprocessor power consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent applies different data types to different components of the neural network processing pipeline. Floating point numbers are used for network input data where high precision is needed, while fixed point numbers are used for neural network weights and internal computations where computational efficiency is prioritized. This local differentiation resolves the contradiction by optimizing each component for its specific requirements rather than using a uniform data type throughout.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The system dynamically switches between floating point and fixed point number representations based on the processing stage and computational requirements. The processor is configured to handle both data types and automatically select the appropriate representation, changing the numerical parameter format to balance between precision needs and computational efficiency at different stages of audio processing.

Inventive Principle:

Principle #35Parameter changes

2Adaptability or versatility

If data type conversion is performed between floating point and fixed point numbers, then compatibility between different data formats is achieved, but processing efficiency decreases

Engineering Contradiction:

Improvedata format compatibilityVSAvoiddata processing efficiency

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The system performs data type conversion in advance during the loading of neural network weights into the processor. By pre-converting and storing weights in the appropriate fixed point format, the system eliminates the need for costly conversions during the actual audio processing operations, thereby maintaining high processing efficiency while still supporting multiple data formats.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The processor acts as an intermediary that natively supports both floating point and fixed point number formats. This intermediary capability allows the system to accept data in various formats from different sources while internally processing using the most efficient representation, avoiding the need for conversion during the critical processing path.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Quantity of substance

If fixed point numbers are used for neural network weights, then memory efficiency and computational cost are reduced, but precision and dynamic range are limited

Engineering Contradiction:

Improvememory usageVSAvoidaudio processing precision

Core Design Contradiction:

Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent applies different data types to different components of the neural network processing pipeline. Floating point numbers are used for network input data where high precision is needed, while fixed point numbers are used for neural network weights and internal computations where computational efficiency is prioritized. This local differentiation resolves the contradiction by optimizing each component for its specific requirements rather than using a uniform data type throughout.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS20250310705A1Hearing device with low power neural network

Publication Date: 2025.10.02 GN HEARING AS

AI summary

A hearing device and related method is disclosed, the hearing device comprising a set of input transducers for provision of transducer input data, the set of input transducers comprising a first input transducer for provision of a first transducer input signal as part of the transducer input data; a processor for processing transducer input data and providing an electrical output signal based on the transducer input data; and a receiver for converting the electrical output signal to an audio output signal, wherein the processor is configured to apply a neural network to a network input based on the transducer input data for provision of a network output, the electrical output signal based on the network output, wherein the network input has a first data type and weights of the neural network have a second data type different from the first data type.