Towards Semantic Interpretation of Structured Data Sources in Privacy-Preserving Environments

Christina Karalka, Georgios Meditskos, Nick Bassiliades

Keywords: knowledge graphs, ontologies, data sources, semantic interpretation, privacy preservation

Abstract: As the use of sensitive data becomes increasingly prevalent, it is essential to ensure that privacy preserving technologies are effectively utilized to protect such data. Relational databases are commonly used for data storage, but they may not provide sufficient insights for identifying privacy vulnerabilities. Moreover, due to the various legal and technical terms, as well as the various actors involved, it is difficult to decide the privacy-preserving technology and the type of the configuration needed for a specific dataset. This short paper presents work in progress towards adding a semantic layer on top of structured data sources for efficient and intelligent use of data in privacy-preserving scenarios. More specifically, we present key research directions for the development of SemCrypt, a novel framework for schema-enrichment through semantic annotations and mappings to Knowledge Bases and domain ontologies so as to: a) interlink and contextually enrich schemata and data in an interoperable manner; b) use the underlying semantics to assist data owners in assessing privacy preserving technologies depending on the sensitivity of data in different use cases, such as in health, finance and cyber threat intelligence.