Analytic Workload Partitioning for Privacy and Performance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Access to large, high-quality data sets for machine learning and data analytics is limited due to confidentiality and privacy concerns, as well as regulatory restrictions, leading to challenges in fields like healthcare, finance, and manufacturing.
Innovation Solution
A privacy preservation system that splits the analytic workload and data into sub-workflows, processed by privacy-preserving engines, maintaining end-to-end encryption and optimizing mathematical operations, thereby reducing computational and communicational overhead while ensuring data privacy and confidentiality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data is shared for machine learning and analytics, then data quality and sample size improve, but data privacy and confidentiality are compromised
Solution Approach 1:
The patent segments the analytic workload into multiple sub-workflows that can be distributed across different processing environments. This allows data to be processed in smaller, controlled portions rather than requiring complete data sharing, thereby maintaining privacy while enabling sufficient analytic processing.
Solution Approach 2:
The patent introduces trusted execution environments and privacy-preserving processing systems as intermediaries between data owners and analytics users. These intermediaries enable data processing without direct data sharing, using mechanisms like secure enclaves and homomorphic encryption to protect confidentiality while allowing analytics.
2Object-affected harmful factors
If homomorphic encryption is used for privacy preservation, then data confidentiality is maintained, but computational overhead increases significantly
Solution Approach 1:
The patent divides the analytic workflow into segments that can be processed using different privacy-preserving techniques. By segmenting the workload, the system can apply computationally intensive homomorphic encryption only where necessary while using lighter-weight protection mechanisms elsewhere, reducing overall computational overhead.
Solution Approach 2:
The patent optimizes homomorphic encryption parameters and selects appropriate encryption schemes based on the specific analytic requirements. By adjusting encryption parameters and choosing the right level of cryptographic protection for each sub-workflow, the system balances confidentiality with computational efficiency.
3Reliability
If data is encrypted for security, then privacy is preserved, but data processing efficiency decreases
Solution Approach 1:
The patent segments processing into encrypted and unencrypted portions, allowing efficient processing of non-sensitive data while maintaining security for sensitive operations. This segmentation enables the system to maintain security requirements without encrypting entire data sets, thus preserving processing efficiency.
Solution Approach 2:
The patent performs data preprocessing and feature engineering before encryption where possible, and uses optimized encryption schemes that minimize processing overhead. By preparing data in advance and selecting efficient cryptographic operations, the system reduces the impact of encryption on overall processing efficiency.
Data Source
AI summary
The present disclosure provides privacy preservation of analytic workflows based on splitting the workflow into sub-workflows each with different privacy-preserving characteristics. Libraries are generated that provide for formatting and/or encrypting data for use in the sub-workflows and also for compiling a machine learning algorithm for the sub-workflows. Subsequently, the sub-workflows can be executed using the compiled algorithm and formatted data.