Synthetic Data Generation with Client-Side Privacy Validation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing artificial intelligence models require large datasets that include personal information, posing a risk of data leakage during training, which is undesirable.
Innovation Solution
A method involving a client apparatus that generates synthetic data by receiving original data, creating seed data, transmitting it to a server for candidate synthetic data generation, validating the similarity with the original data, and storing valid synthetic data to form a dataset, ensuring no personal information leakage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If original training data including personal information is used to train AI models, then model performance is improved, but personal information leakage risk increases
Solution Approach 1:
The patent creates synthetic data that copies the statistical properties and patterns of original personal data without reproducing actual personal information. A generator network learns the distribution characteristics of original data and generates artificial samples that maintain similar statistical features while being completely synthetic, thus preserving model training effectiveness while eliminating privacy risks
Solution Approach 2:
The patent introduces an intermediary validation mechanism that checks generated synthetic data against original data distributions. The validator network acts as a mediator to ensure synthetic data maintains appropriate similarity to original data patterns while filtering out any potential personal information leakage, bridging the gap between data utility and privacy protection
2Object-affected harmful factors
If synthetic data is generated to replace original personal data, then personal information leakage risk is reduced, but data similarity and training effectiveness may deteriorate
Solution Approach 1:
The patent implements a feedback loop where the validator network continuously evaluates generated synthetic data against original data distributions and provides feedback to the generator. This feedback mechanism ensures that synthetic data maintains high similarity to original data patterns while preserving privacy, allowing iterative refinement of data quality
Solution Approach 2:
The patent transforms the data generation problem by changing parameters from direct data copying to statistical property reproduction. The generator network learns and reproduces distribution parameters, correlation structures, and statistical features of original data rather than copying actual values, maintaining data utility while ensuring privacy
3Manufacturing precision
If a validation process is implemented to check similarity between synthetic and original data, then data quality is improved, but processing time and complexity increase
Solution Approach 1:
The patent implements self-service validation where the system automatically evaluates its own generated synthetic data using the validator network. The validation process is integrated into the generation workflow, with the system autonomously checking data quality metrics and filtering synthetic samples without external intervention, maintaining high data quality while managing complexity through automation
Data Source
AI summary
Proposed is a method of generating synthetic data, which includes receiving, by a client apparatus, original data including personal information, acquiring, by the client apparatus, seed data based on the original data, transmitting, by the client apparatus, the seed data to a server, receiving, by the client apparatus, first candidate synthetic data which is generated based on the seed data from the server, validating, by the client apparatus, the first candidate synthetic data based on a similarity between the first candidate synthetic data and the original data, and storing, by the client apparatus, the first candidate synthetic data as a member of candidate synthetic dataset if the first candidate synthetic data is valid synthetic data from the validation result.