Adaptive ML Training Data Sanitization for Privacy-Accuracy Trade-offs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge lies in utilizing sensitive data for training machine learning models without exposing personally identifiable information (PII) to vulnerabilities associated with processing and storage in insecure environments, such as public clouds.
Innovation Solution
A system comprising a data sanitization subsystem, a machine learning (ML) subsystem, and an adaptive retraining subsystem, which implements data sanitization algorithms with adjustable parameters to generate sanitized datasets, trains ML models, and iteratively adjusts sanitization parameters to ensure model accuracy and data privacy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data sanitization parameters are increased to enhance PII protection, then data privacy is improved, but model training accuracy deteriorates
Solution Approach 1:
The system dynamically adjusts data sanitization parameters based on feedback from model performance evaluation. The adaptive retraining subsystem modifies sanitization intensity iteratively, transforming the static sanitization process into a dynamic system that responds to performance metrics, thereby resolving the contradiction between privacy protection and model accuracy
Solution Approach 2:
The system changes sanitization parameters (such as noise level, masking ratio, or perturbation intensity) to find the optimal balance point. By adjusting these parameters iteratively based on model performance feedback, the system achieves both adequate PII protection and acceptable model training accuracy
2Measurement precision
If iterative re-training with parameter adjustments is implemented to improve model accuracy, then model performance is improved, but computational time and resources increase
Solution Approach 1:
The system implements a feedback loop where model performance is evaluated against accuracy thresholds, and this feedback drives automatic parameter adjustments for subsequent training iterations. This closed-loop control ensures that retraining continues only until performance targets are met, preventing unnecessary computational waste while achieving required accuracy levels
Solution Approach 2:
The system performs preliminary evaluation of model performance after each training iteration with adjusted sanitization parameters. By assessing accuracy early and deciding whether further retraining is necessary, the system avoids unnecessary computational time while ensuring performance requirements are met
Data Source
AI summary
Systems, computer program products, and methods are described herein for secure training of machine learning models. The present disclosure is configured to receive a training dataset, wherein the training dataset comprises personally identifiable information (PII); implement a data sanitization algorithm using adjustable data sanitization parameters on the training dataset to generate a corresponding sanitized training dataset; train an ML model using the sanitized training dataset; deploy the trained ML model on a live dataset; evaluate a performance accuracy of the ML model against an accuracy threshold; and iteratively adjust the data sanitization parameters for subsequent iterative re-training of the ML model until the performance accuracy of the ML model meets the accuracy threshold.