Sound Processing Device Distributed Model Collaboration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sound processing devices fail to effectively collaborate to record and process overlapping soundscapes in public places, leading to loss of useful information and inefficiencies in tasks like sound triangulation and 3D reconstruction.
Innovation Solution
Implementing a distributed learning algorithm that allows sound processing devices to collaborate without exchanging actual recorded sound, where helper devices determine local adjustments to a sound processing model and share these with a main device, enabling real-time processing and refinement of the model based on local sound data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If sound processing devices exchange actual recorded sound in peer-to-peer fashion, then collaboration and information sharing are enabled, but communication load and privacy risks increase significantly
Solution Approach 1:
The patent extracts only the essential model adjustment parameters from the recorded sound data, separating the useful information (model adjustments) from the raw recorded sound. This allows devices to collaborate and improve the sound processing model without exchanging the actual recorded sound, thereby reducing communication load while preventing useful information loss.
Solution Approach 2:
Instead of exchanging original recorded sound data, the patent uses copies in the form of model adjustment parameters derived from the recorded sound. These parameters capture the essential information needed for model improvement without requiring transmission of the actual sound recordings, thus reducing communication overhead.
2Manufacturing precision
If sound processing devices exchange actual recorded sound, then model refinement is enabled, but privacy risks increase due to exposure of speech and sound features
Solution Approach 1:
The patent extracts only the necessary model adjustment parameters from the recorded sound data, separating the useful information needed for model refinement from the privacy-sensitive raw sound data. This extraction process enables model improvement while eliminating privacy risks associated with sharing actual recorded sound.
Solution Approach 2:
The sound processing model acts as an intermediary that receives local adjustments from multiple devices without directly exchanging raw recorded sound. This intermediary mechanism enables collaborative model refinement while protecting privacy by never transmitting the actual sound recordings containing sensitive speech and sound features.
3Quantity of substance
If multiple devices record overlapping soundscapes simultaneously, then comprehensive sound data is captured, but collaboration efficiency decreases due to lack of coordination
Solution Approach 1:
The patent merges the sound processing capabilities of multiple devices through a shared model that receives local adjustments from each device. Instead of coordinating complex real-time recording operations, devices independently process their local sound data and contribute model adjustments, which are then integrated to improve the overall sound processing model, thereby enhancing collaboration efficiency while maintaining comprehensive sound data coverage.
Data Source
AI summary
Examples relate to sound processing devices, to corresponding methods and computer programs for sound processing devices, and to devices, such as mobile devices or hearing aids, comprising a sound processing device. A sound processing device comprises at least one interface for communicating with one or more further sound processing devices. The sound processing device comprises processing circuitry, configured to obtain a sound processing model. The processing circuitry is configured to receive, from the one or more further sound processing devices, one or more local adjustments to the sound processing model determined by the one or more further sound processing devices based on sound recorded locally by the one or more further sound processing devices. The processing circuitry is configured to adjust the sound processing model based on the one or more local adjustments. The processing circuitry is configured to process sound recorded locally by the sound processing device using the sound processing model.