De-identified Computing Circuit Using Analog PUF Offset Errors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional privacy protection methods in data collection centers fail to ensure secure data transfer and are prone to information leakage, necessitating enhanced privacy protection measures for both data collection centers and terminal apparatuses.
Innovation Solution
A computing circuit with a de-identified architecture utilizing an analog physical unclonable function (PUF) to introduce unpredictable sensing drift errors, which are used for data de-identification through a de-identification circuit that operates on accumulated data to generate de-identification data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional privacy protection methods are used through a data collection center, then data management is simplified, but data security and privacy protection are insufficient
Solution Approach 1:
The patent introduces a de-identification circuit as an intermediary component between the arithmetic array and external systems. This circuit contains analog offset errors that act as a mediator to transform identifiable computation results into de-identified data, thereby enhancing privacy protection without requiring complex centralized management architectures
Solution Approach 2:
The de-identification circuit utilizes its own inherent analog offset errors (natural variations in transistor characteristics) to perform de-identification operations. The system leverages its own manufacturing imperfections as a resource for privacy protection, eliminating the need for external key management infrastructure and achieving self-contained privacy protection
2Productivity
If data is transferred through traditional channels, then data processing is efficient, but information leakage risks increase
Solution Approach 1:
The patent changes the parameter space of data representation by performing computations in the analog domain and utilizing analog offset errors to transform the data into a de-identified form. This parameter transformation allows efficient processing while inherently protecting against information leakage, as the de-identified data cannot be easily reverse-engineered
Solution Approach 2:
The patent converts the harmful effect of analog offset errors (which traditionally represent manufacturing imperfections and sources of computational inaccuracy) into a beneficial privacy protection mechanism. These same manufacturing variations that cause computational drift are leveraged to de-identify data, turning a source of error into a security feature
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach effectively enhances privacy protection by ensuring secure data processing and transfer by generating de-identification data with sufficient privacy levels, mitigating risks of information leakage during data transfer.
Implementation Method 1
an analog physical unclonable function (PUF) to introduce unpredictable sensing drift errors
Data Source
AI summary
A computing circuit with a de-identified architecture, a data computing method, a data processing system, and a data de-identification method are provided. The computing circuit includes an arithmetic array and a de-identification circuit. The computing circuit may perform an accumulation operation on input data to generate accumulated data by the arithmetic array. The de-identification circuit has an analog offset error determined based on an analog physical unclonable function. The computing circuit may operate the accumulated data according to the analog offset error to generate de-identification data by the de-identification circuit. It can not only provide the analog offset error through the transistors in the de-identification circuit, but also be combined with obfuscated code settings to dynamically adjusting the degree of de-identification of data.