Privacy-Preserving Machine Learning Version 8

Privacy-Preserving Machine Learning MEAP V08
Copyright
Welcome
Brief contents
Chapter 1: Privacy considerations in machine learning
	1.1 The Privacy Complications in the AI Era
	1.2 The Threat of Learning Beyond the Intended Purpose
		1.2.1 The Problem of Private Data in the Clear
		1.2.2 Reconstruction Attacks
		1.2.3 Model Inversion Attacks
		1.2.4 Membership Inference Attacks
		1.2.5 De-Anonymization or Re-Identification Attacks
		1.2.6 Challenges of Privacy Protection in Big Data Analytics
	1.3 Securing Privacy while Learning from Data: Privacy-Preserving Machine Learning
		1.3.1 Use of Differential Privacy
		1.3.2 Local Differential Privacy
		1.3.3 Privacy-preserving Synthetic Data Generation
		1.3.4 Privacy-preserving Data Mining Techniques
		1.3.5 Compressive Privacy
	1.4 How is This Book Structured?
	1.5 Summary
Chapter 2: Differential privacy for machine learning
	2.1 What is Differential Privacy (DP)?
		2.1.1 The Concept of Differential Privacy
		2.1.2 How Differential Privacy Works?
	2.2 Mechanisms of Differential Privacy
		2.2.1 Binary Mechanism (Randomized Response)
		2.2.2 Laplace Mechanism
		2.2.3 Exponential Mechanism
	2.3 Properties of Differential Privacy
		2.3.1 Post-Processing Property of Differential Privacy
		2.3.2 Group Privacy Property of Differential Privacy
		2.3.3 Composition Properties of Differential Privacy
	2.4 Summary
Chapter 3: Advanced concepts of differential privacy for machine learning
	3.1 How to Apply Differential Privacy in Machine Learning?
	3.2 Differentially Private Supervised Learning Algorithms
		3.2.1 Differentially Private Naive Bayes Classification
		3.2.2 Differentially Private Logistic Regression
		3.2.3 Differentially Private Linear Regression
	3.3 Differentially Private Unsupervised Learning Algorithms
		3.3.1 Differentially Private K-means Clustering
	3.4 Case Study: Differentially Private Principal Component Analysis
		3.4.1 The Privacy of PCA Over Horizontally Partitioned Data
		3.4.2 The Design of Differentially Private PCA Over Horizontally Partitioned Data
		3.4.3 Experimentally Evaluating the Performance of the Protocol
	3.5 Summary
Chapter 4: Local differential privacy for machine learning
	4.1 What is Local Differential Privacy?
		4.1.1 The Concept of Local Differential Privacy
		4.1.2 Randomized Response for Local Differential Privacy
	4.2 The Mechanisms of Local Differential Privacy
		4.2.1 Direct Encoding
		4.2.2 Histogram Encoding
		4.2.3 Unary Encoding
	4.3 Summary
Chapter 5: Advanced mechanisms of local differential privacy for machine learning
	5.1 A Quick Recap on Local Differential Privacy
	5.2 The Advanced Mechanisms of Local Differential Privacy
		5.2.1 The Laplace Mechanism for LDP
		5.2.2 Duchi’s Mechanism
		5.2.3 Piecewise Mechanism
	5.3 A Case Study and the Implementation of Locally Differentially Private Naïve Bayes Classification
		5.3.1 The Use of Naïve Bayes with Machine Learning Classification
		5.3.2 LDP Naïve Bayes with Discrete Features
		5.3.3 LDP Naïve Bayes with Continuous Features
		5.3.4 Evaluating the Performance of Different LDP Protocols
	5.4 Summary
Chapter 6: Privacy-preserving synthetic data generation
	6.1 Overview of Synthetic Data Generation
		6.1.1 What is Synthetic Data? Why is it Important?
		6.1.2 Application Aspects of Using Synthetic Data for Privacy Preservation
		6.1.3 How to Generate Synthetic Data?
	6.2 Assuring Privacy via Data Anonymization
		6.2.1 The Issue of Private Information Sharing vs Privacy Concerns
		6.2.2 Use of k-anonymity Against Re-identification Attacks
		6.2.3 Anonymization Beyond k-anonymity
	6.3 Differential Privacy for Privacy-preserving Synthetic Data Generation
		6.3.1 Differentially Private Synthetic Histogram Representation Generation
		6.3.2 Differentially Private Synthetic Tabular Data Generation
		6.3.3 Differentially Private Synthetic Multi-Marginal Data Generation
	6.4 Case study on Private Synthetic Data Release via Feature-level Micro-aggregation
		6.4.1 Generating Synthetic Data
		6.4.2 Evaluating the Performance of the Generated Synthetic Data
	6.5 Summary
Chapter 7: Privacy-preserving data mining techniques
	7.1 The Importance of Privacy Preservation in Data Mining and Management
	7.2 Privacy Protection in Data Processing and Mining
		7.2.1 What is Data Mining and How it can Help?
		7.2.2 Impact of Privacy Regulatory Requirements
	7.3 Protecting Privacy by Modifying the Input
		7.3.1 Applications and the Limitations
	7.4 Protecting Privacy when Publishing Data
		7.4.1 Implementing Data Sanitization Operations in Python
		7.4.2 k-anonymity
		7.4.3 Implementing k-anonymity in Python
	7.5 Summary
Chapter 8: Privacy-preserving data management and operations
	8.1 A Quick Recap on Privacy Protection in Data Processing and Mining
	8.2 Privacy Protection beyond k-anonymity
		8.2.1 l-diversity
		8.2.2 t-closeness
	8.3 Protecting Privacy by Modifying the Data Mining Output
	8.4 Privacy Protection in Data Management Systems
		8.4.1 Database Security and Privacy: Threats and Vulnerabilities
		8.4.2 How Probable is a Modern Database System to Leak Private Information?
		8.4.3 Attacks on Database Systems
		8.4.4 Privacy Preserving Techniques in Statistical Database Systems
		8.4.5 Toward Designing Tailor-made Privacy Preserving Database System
	8.5 Summary
Chapter 9: Compressive privacy for machine learning
	9.1 Introduction to Compressive Privacy
	9.2 The Mechanisms of Compressive Privacy
		9.2.1 Principal Component Analysis (PCA)
		9.2.2 Other Dimensionality Reduction (DR) Methods
	9.3 Implementing Compressive Privacy for Machine Learning Applications
		9.3.1 The Accuracy of the Utility Task
		9.3.2 The Effect of ρ\' in DCA for Privacy and Utility
	9.4 Case Study: Privacy-Preserving PCA/DCA on Horizontally Partitioned Data
		9.4.1 Recap on Different Dimensionality Reduction Approaches
		9.4.2 Use of Additive Homomorphic Encryption
		9.4.3 Overview of the Proposed Approach
		9.4.4 How Privacy-Preserving Computation Works
		9.4.5 Evaluating the Efficiency and Accuracy of the Privacy Preserving PCA/DCA
	9.5 Summary
Chapter 10: Putting it all together: designing a privacy-enhanced platform for research data protection and sharing (DataHub)
	10.1 Overview and the Significance of Having a Research Data Protection and Sharing Platform
		10.1.1 Motivation behind the DataHub
		10.1.2 What are the important features that we are looking at?
	10.2 Understanding the Research Collaboration Workspace
		10.2.1 The Architectural Design
		10.2.2 Blend of Different Trust Models
		10.2.3 Configuring Access Control Mechanisms
	10.3 Integrating Privacy and Security Technologies into DataHub
		10.3.1 Data Storage with Cloud-based Secure NoSQL Database Solution
		10.3.2 Privacy-Preserving Data Collection with Local Differential Privacy
		10.3.3 Privacy-Preserving Machine Learning (PPML)
		10.3.4 Privacy-Preserving Query Processing
		10.3.5 Use of Synthetic Data Generation in the DataHub Platform
	10.4 Summary
Appendix A: More details about Differential Privacy
	A.1 What is the formal definition of differential privacy?
	A.2 Other differential privacy mechanisms
	A.3 The formal definition of sequential composition DP
	A.4 The formal definition of parallel composition DP
	A.5 References