Computer Science
隐私保护机器学习
Privacy-Preserving Machine LearningA set of techniques and procedures designed to perform machine learning tasks on sensitive data (such as medical records) without exposing the underlying private information. This involves balancing the analytical utility of the data with the security requirements of privacy. / 一组旨在在敏感数据（如医疗记录）上执行机器学习任务而不暴露底层私人信息的技术和程序。这涉及平衡数据的分析效用与隐私的安全要求。
- Knowledge type
- 实际应用
- First proposed or discovered
- Proposer
- Turkey Farm member
- Creator
- 农场主
- Last updated
- 2026-11-06 23:05:52
Sub-knowledge points
Anonymization often preserves machine learning valueSynthetic data shows significant performance drop in classification due to lower discriminative feature boundariesRegression accuracy drops significantly for synthetic data (R² reduction to 0.431)Anonymized data preserves data structure better than synthetic data
Higher-order knowledge points
Data-driven decision making in healthcareClinical analysis using tabular medical data
Main parameters and performance indicators
Data Anonymization: Pseudonymization, generalization, attribute-based perturbationSynthetic Data Generation: Conditional Tabular Generative Adversarial Network (CTGAN)Analytical Tasks: Classification, regression, clusteringPrivacy Metric: Utility preservation vs. privacy leakage
Open problems
Preserving complex numerical relationships and extreme values in synthetic dataOptimizing the balance between privacy and utility across different dataset characteristicsMaintaining cluster distributions and inherent variability in synthetic data
Related literature
ranked by community starsOn the Effectiveness of Data Anonymization and Synthetic Data Generation for Privacy-Preserving Machine Learning in HealthcareOpen source
Related products and links
Anonymized tabular datasetsSynthetic tabular datasets generated by CTGAN
Creation and modification trail
1 recordsRecord title: Privacy-Preserving Machine Learning.
Node update feedback
review required