Proving Soundness of SPARQL Query Results Using Selective Disclosure of RDF Datasets and Zero-Knowledge Proofs

Our work connects the W3C Verifiable Credentials (VC) data model and zero-knowledge proofs (ZKPs) to allow for minimised information disclosure. More generally, as VCs are Resource Description Framework (RDF) datasets, our work enables the application of the following ZKPs on such an RDF dataset: selective disclosure, proof of numeric bounds, and proof of set non-membership – all on the level of RDF terms. To pre-process RDF datasets for such ZKP applications, we introduce schema-free and schema-based approaches, and highlight their differences. We then present a data model for credential presentation and its semantics, where we show that selective disclosure is equivalent to RDF’s simple entailment, and that verifications of the presented proofs are validity checks on the processed RDF dataset.