While data anonymization has been a well-explored field for several decades, and Semantic Web technologies and standards being more and more used by institutions and organizations for modeling Linked Open Data, privacy-preserving data publishing is still obscure on this field. It appears important to define such a framework, notably using Semantic Web standards such as RDF graphs and triples, or SPARQL queries.
We introduce and develop a declarative framework for privacy-preserving Linked Data publishing in which privacy and utility policies, modeling the data provider's disclosing and usage constraints are specified as SPARQL queries. This approach is data-independent and leads to inspect only the privacy and utility policies in order to determine the sequence of anonymization operations applicable to any graph instance for satisfying the policies. The soundness of our approach is proved and we gauge their performance through experiments.
This is talk relates to an accepted paper of ISWC 2018 which will be formally presented next October.
