Umeå University's logo

Nowadays, the amount of data generated by users on the Web has increased dramatically, and managing it efficiently is getting more challenging, especially for small or medium-sized organizations. Often, the data to be managed is encoded in a low-level format, requiring domain experts to manually produce a high-level conceptual view from the raw data. A virtual Knowledge Graph (VKG) is a semantic framework that stands as a data integration paradigm aiming to provide convenient, user-friendly access to the data. VKGs, formerly known as ontology-based data access (OBDA), have emerged as an information management system that solves the complex problem of data integration by exposing the end users to an ontology that is typically expressed in some fragment of the Web Ontology Language (OWL2), standardized by the World Wide Web Consortium (W3C). The framework is a virtual approach that typically consists of three main components: an ontology, which is a high-level and conceptual representation of the domain of interest, a set of data sources, and the mapping between the two.

However, with the vision of the Semantic Web, which has consisted of enabling the "Read/Write" Web for structured data, the main focus of research in VKGs has been centered around query-answering which consists of using the ontology layer to extract information specified through a query from the underlying data sources. Yet the problem of updates in VKGs has, however, received little attention and represents an important feature that will enable VKGs to be fully-fledged and, thus, let user fully manage their data from the ontology they are exposed to. This dissertation aims to study and introduce the notions of ontology-based update in the context of VKGs and to study the foundational issues of this extension. In other words, the aim is to study how updates posed over the ontology layer are rewritten into equivalent updates over the underlying data sources.

However, due to the complex nature of VKG mappings, the translation of ontology-based updates is not always deterministic and might lead to extra unintended updates in the knowledge graph, which we refer to as side effects. Considering ontologies specified in DL-LiteR, the formal counterpart of OWL 2 QL, we study how to efficiently translate ontology-based updates by relying on the reverse of VKG mappings. Secondly, based on a given comparison metric for ontology-based update translations, we compute a set of translations with minimum side effects. We also introduce the notion of preference (provided by the user in a declarative format) that can be used to provide a deterministic translation. Finally, we demonstrate the practical feasibility of our approach by implementing the proposed techniques within the Ontop VKG system, translating SPARQL Update operations into SQL statements via R2RML mappings.

In Virtual Knowledge Graphs (VKGs), access to a relational data source is provided through an ontology that is linked to the data source via declarative mappings. While the problem of query answering in VKGs has been studied extensively over the past years, much less attention has been devoted to the problem of instance-level updates over the VKG, realized by updating the underlying data source. Due to the form of VKG mappings, translating VKG updates into a source updates might lead to side-effects in the VKG, i.e., unwanted insertions or deletions. In this paper, we build on a recent proposal for translating VKG updates into source updates, and extend it by introducing the notion of compensation, which are additional updates that aim at reverting side-effects. We provide a novel algorithm relying on multiple levels of compensation and show that it computes source updates with minimal side-effects in the VKG.

The Virtual Knowledge Graph (VKG) paradigm enables querying het-erogeneous relational databases through a unified semantic layer comprising anontology and declarative mappings (typically in R2RML). While querying iswell-supported, propagating updates from the virtual RDF graph (ABox) backto the source databases remains a challenge. This paper addresses the problemof translating updates (expressed in SPARQL Update) applied over the ABoxinto equivalent SQL updates over the underlying databases, specifically withinthe Ontop VKG system. Key difficulties arise from the non-injectivity inherentin R2RML mappings, where a single ABox update can correspond to multiplesource update possibilities, and the potential for these source updates to causeunintended side effects—additional insertions or deletions in the ABox beyondthe user’s original intent. While relying on Ontop’s query rewriting engine, ourmethod employs lineage computation to identify source tuples for deletion and astrategy for handling existential variables during insertion. It generates candidateSQL translations, analyzes their potential side effects on the virtual ABox, andselects the ones that minimize these unintended consequences. This work repre-sents a step toward closing the Linked Data Life Cycle loop, allowing changes inthe knowledge graph to be reflected in the corresponding source.

Virtual Knowledge Graph (VKG) is known as a data integration paradigm used to efficiently manage the heterogeneity of richly structured data that is common inside several organizations, in inter-organizational settings, and more openly on the Web. Although such a paradigm continues to gain importance in both foundational and applied research, updates in VKG systems remain an open challenge that has received less attention. Yet, a solution to such a problem would be of great importance, as it would allow VKG systems to be full-fledged, thus allowing end-users to fully manage source data through the lens of the ontology they are exposed to. This research aims to propose a comprehensive framework for instance-level updates in VKGs, where updates posed over the ontology have to be translated into source-level updates and, more importantly, how the side effects related to the propagation of ontology-based updates to the underlying data source can be minimized.

In Virtual Knowledge Graphs (VKGs), access to a relational data source is provided through an ontology, which is linked to the data source via declarative mappings. VKGs stand as a predominant paradigm for the access to (and integration of) heterogeneous data sources. However, little attention has been paid so far to the issue of updates in VKGs expressed over the ontology, which represents a crucial feature for fully managing data sources through the lens of an ontology. In this paper, we consider the problem of updating a VKG instance by specifying a set of insertions and deletions of ontology instances and propagating these updates to the underlying data source through the VKG mapping. We consider ontologies specified in the DL-LiteR lightweight ontology language and study the problem for the case where source queries in mappings are unions of conjunctive queries. We rely on the notion of maximum recovery of VKG mappings, borrowed from the data exchange setting, and propose methods to compute the set of source updates that translate an ontology update with a minimal side-effect, considering both insertions and deletions of multiple ABox assertions.

The field of Virtual Knowledge Graphs (VKGs) continues to grow in both academic and applied contexts. Yet, the issue of updates in VKG systems has not yet received adequate attention, although it is crucial to manage data modifications at the data source level through the lens of an ontology. In this paper, we focus on VKGs whose ontology is specified in the lightweight ontology language DL-LiteA, and we propose diverse settings and research directions we intend to explore to address the challenge of translating ontology-based updates into updates at the level of data sources. We also pay attention to the important problem of automated analysis of mappings, which plays a major role when it comes to reformulating ontology-based update requests into update requests over the data sources.