Vehicle Model Syncing via Isometric Data Masking for Privacy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine-learned models in vehicles cannot be remotely trained using data from other vehicles due to privacy constraints, as traditional methods like homomorphic encryption and differential privacy face challenges in preserving data privacy while maintaining distance metrics, leading to impractical solutions.
Innovation Solution
Applying isometric transformations and masking techniques to vehicle data, allowing remote training of models while preserving privacy by transforming data points into a multi-dimensional space using rotation, reflection, and translation, and generating masking signatures for secure data transmission and model syncing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If vehicle data is transmitted to remote computing systems for model training, then model training capability is improved, but data privacy is compromised
Solution Approach 1:
The patent introduces isometric transformation as an intermediary process that converts raw vehicle data into transformed data with preserved spatial relationships but obscured original meanings. This intermediary transformation layer enables remote model training while protecting user privacy, as the transformed data cannot be directly traced back to original personal information
Solution Approach 2:
The patent creates a transformed copy of the original vehicle data through isometric transformations. This copy maintains the structural and spatial relationships necessary for model training but removes personally identifiable information. The model trains on this transformed copy rather than the original sensitive data
2Measurement precision
If datasets from multiple vehicles are combined for training, then model accuracy is improved, but data compatibility challenges increase
Solution Approach 1:
The patent applies isometric transformations that change the parameter representation of data from multiple vehicles into a unified transformed space. By transforming spatial coordinates, temporal stamps, and contextual features through consistent isometric operations, data from different vehicles becomes compatible while preserving their unique characteristics for accurate model training
3Object-affected harmful factors
If isometric transformations are applied to mask data, then data privacy is improved, but processing complexity increases
Solution Approach 1:
The patent performs isometric transformations as a preliminary action before data transmission or storage. By pre-transforming the data into masked form with preserved spatial relationships, the system eliminates the need for complex real-time processing during model training, as the transformation is completed upfront
Data Source
AI summary
Methods, computing systems, and technology for remotely training a machine-learned model using masked inputs is presented. For example, a computing system may be configured to obtain vehicle operations data associated with a user and a vehicle. The computing system may be configured to transform the vehicle operations data based on one or more isometric transformations to obtain a masked input, the one or more isometric transformations associated with a feature masking key. The computing system may be configured to input the masked input to a machine-learned model to obtain a masked output, wherein the machine-learned model was trained using previous masked inputs. The computing system may be configured to determine, based on an action masking key, an action associated with the masked output. The computing system may be configured to control a component of the vehicle based on the action.