Full Description
This book constitutes the refereed proceedings of the International Conference on Privacy in Statistical Databases, PSD 2022, held in Paris, France, during September 21-23, 2022.
The 25 papers presented in this volume were carefully reviewed and selected from 45 submissions. They were organized in topical sections as follows: Privacy models; tabular data; disclosure risk assessment and record linkage; privacy-preserving protocols; unstructured and mobility data; synthetic data; machine learning and privacy; and case studies.
Contents
Privacy models.- An optimization-based decomposition heuristic for the microaggregation problem.- Privacy Analysis with a Distributed Transition System and a data-wise metric.- Multivariate Mean Comparison under Differential Privacy.- Asking The Proper Question: Adjusting Queries To Statistical Procedures UnderDifferential Privacy.- Towards integrally private clustering: overlapping clusters for high privacy guarantees.- Tabular data.- Perspectives for Tabular Data Protection - How About Synthetic Data?.- On Privacy of Multidimensional Data Against Aggregate Knowledge Attacks.- Synthetic Decimal Numbers as a Flexible Tool for Suppression of Post-published Tabular Data.- Disclosure risk assessment and record linkage.- The risk of disclosure when reporting commonly used univariate statistics.- Privacy-Preserving protocols.- Tit-for-Tat Disclosure of a Binding Sequence of User Analysesin Safe Data Access Centers.- Secure and non-interactive k-NN classifier using symmetric fully homomorphic encryption.- Unstructured and mobility data.- Automatic evaluation of disclosure risks of text anonymization methods.- Generation of Synthetic Trajectory Microdata from Language Models.- Synthetic data.- Synthetic Individual Income Tax Data: Methodology, Utility, and Privacy Implications.- On integrating the number of synthetic data sets m into the a priori synthesis approach .- Challenges in Measuring Utility for Fully Synthetic Data.- Comparing the Utility and Disclosure Risk of Synthetic Data with Samples of Microdata.- Utility and Disclosure Risk for Differentially Private Synthetic Categorical Data.- Machine learning and privacy.- Membership Inference Attack Against Principal Component Analysis.- When Machine Learning Models Leak: An Exploration of Synthetic Training Data.- Case studies.- A Note on the Misinterpretation of the US Census Re-identification Attack.- A Re-examination of the Census Bureau Reconstruction and Reidentification Attack.- Quality Assessment of the 2014 to 2019 National Survey on Drug Use and Health (NSDUH) Public Use Files.- Privacy in Practice: Latest Achievements of the EUSTAT SDC group.- How Adversarial Assumptions Influence Re- identification Risk Measures: A COVID-19 Case Study.
