Organizers: Domenico Lembo, Valerio Santarelli, Domenico Fabio Savo - Affiliation: Sapienza Università di Roma & OBDA Systems s.r.l.
The tutorial illustrates methodologies for developing ontology-based data management (OBDM) solutions, which aim at coupling conceptual views of the information domain, expressed as OWL 2 ontologies, with the actual and possibly pre-existing data sources. In the tutorial we will introduce the basics of OBDM, present Graphol, a graphical model for quick development of OWL 2 ontologies, and illustrate how to link ontologies with data. We will conduct a hands-on-session in which participants will develop (small) OBDM applications and will experiment OBDM functionalities, such as answering SPARQL queries, by exploiting state-of-the-art OBDM tools.
Note: The hands-on session will make use of a virtual machine. We recommend attendees to install VirtualBox prior to the tutorial.
Ontology-based data management (OBDM) is a recent paradigm for accessing and managing data sources through an ontology that acts as a conceptual, integrated view of the data, and declarative mappings that connect the ontology to the data sources. The OBDM framework has garnered widespread interest in recent years, which has led to the development of advanced instruments based on this technology. These tools have been recently adopted for the development of OBDM applications for enterprise data management in various domains. One important aspect in OBDM concerns the construction of a system specification, i.e., the definition of both the ontology and the mapping. This is a complex activity, which needs various competences and adequate methodologies and tools supporting it, and this is specifically the topic addressed by the present tutorial. The tutorial leverages on the experience of the Sapienza research group on semantic technologies, which has been involved in the last years in several industrial and academic projects and has recently founded OBDA Systems, an innovative startup of Sapienza University of Rome.
After recalling the basics of OBDM, in the tutorial we will first concentrate on the ontology design task. To this aim, we will introduce a graphical language, called Graphol (http://www.obdasystems.com/graphol), which allows the designer to specify full-fledged OWL 2 ontologies in a diagrammatic form. Specifically, we will start by using familiar ER/UML class diagrams. We will translate such diagrams in Graphol, showing that these diagrams admit an essentially isomorphic representation in Graphol, which however has a precise correspondence to OWL 2. Then, we will move to more complex constructs and axioms that cannot be captured in ER/UML but that can be specified in OWL 2 and hence in Graphol. Source data analysis and mapping design will be the next topic of the tutorial. We will give an overview of the formalisms for mapping data to an ontology and will highlight the main issues arising in mapping design, as well as possible approaches to this task. Finally, we will present tools that support the designer for the development of an OBDM specification and for the deployment of the specification over an actual source instance database. Participants will be invited to use such tools in the final hands-on-session of the tutorial.
Previous edition of the event: https://sites.google.com/a/dis.uniroma1.it/mt4obda/
We will introduce the basic concepts related to OBDM, including the possibility of checking the ontology through intensional reasoning and performing query answering of conjunctive queries (i.e., SPARQL basic graph patterns).
We will illustrate how to build a global ontology. We will start using familiar conceptual formalisms like Entity-Relationship Diagrams used in Databases and UML Class Diagrams used in Software Engineering. Then we will show how their semantics can be made precise through a translation in OWL 2. Such translation will be based on the graphical ontology language Graphol, which is in fact able to capture full OWL 2. We will also single out its fragment corresponding to the OWL 2 QL profile.
Then we will move to data source analysis and will show how to map the data therein to elements of the ontology. In particular we will discuss how to formulate queries over the sources and how to generate objects for the ontology through suitable constructors. We will also give some indications, coming from experience, on how to write effective mappings.
The tutorial will be complemented by a hands-on session in which participants will have the possibility of trying what learned through the use of state-of-the-art OBDM tools. In particular, to avoid the overhead of installing and configuring such tools, we plan to distribute usb pendrives with a minimal OS and all tools ready to run.
Tutorial Material: Slides will be made available (through the web site of the tutorial) only to attendees approximately one month before the tutorial. On site we will distribute usb pendrives containing a virtual machine with all tools needed for the hands-on session.
Learning Outcomes: OBDM is a major novelty in dealing with diverse data sources in a semantically rich way. It requires multidisciplinary competences, crossing different Semantic Web subareas and themes. The tutorial will mainly serve the following objectives:
Prerequisites: Basic understanding of languages for conceptual modeling and databases as those acquired in university courses. Familiarity with description logics will be of help but it is not required to follow the tutorial.
Domenico Lembo (Ph.D. in Engineering in Computer Science) is associate professor at Sapienza Università di Roma. His research interests include artificial intelligence, databases, knowledge representation, Semantic Web, data integration, ontologies, inconsistency management in information systems, Description Logics, from both the theoretical and application point of view. He has been co-Chair of the program committee of RW 2017, SEBD 2015, RR 2013, and DL 2007, and general chair of DL 2012. He is member of the steering committee of the International Conference on Web Reasoning and Rule Systems (RR). He has been lecturer in tutorials on topics related to this proposal presented at major international conferences and in courses for Ph.D. programs in computer science.
Domenico Fabio Savo (Ph.D. in Engineering in Computer Science) is a research associate at DIAG, the Department of Computer, Control, and Management Engineering at Sapienza University of Rome, and co-founder and leading member of the OBDA Systems's research and project development team. His main research interests focus on Information and Data Management, in particular the application of reasoning techniques over ontologies for data management (Ontology-based data management and Ontology-based data access), Knowledge representation and modeling, Semantic Web, and Inconsistency Management in ontologies. On these topics he has co-authored more than 40 publications in conference and workshop proceedings and scientific journals. He has participated in several projects in collaboration with private companies and public administration bureaus with the aim of applying semantic technologies to real-world scenarios. He regularly serves as program committee member of the major artificial intelligence conferences. In 2017, he (co)founded OBDA Systems, an innovative startup of Sapienza University, which proposes state-of-the-art solutions based on the most recent breakthroughs in the field of semantic technologies. Finally, he regularly teaches courses at academic level.
Valerio Santarelli (Ph.D. in Engineering in Computer Science) is a research associate at the Sapienza Università di Roma, Dipartimento di Informatica Informatica, Automatica e Gestionale, a member of the Data and Service Integration Laboratory (DASI Lab), and C.E.O. of OBDA Systems. His research interests focus on Knowledge Representation and Reasoning, Artificial Intelligence, Ontology-based Data Management and Access, Visual Languages, and Semantic Web. He is author of several scientific publications in international conferences and journals in these fields. During his doctorate studies, he has taken part in national and European projects centered on applying Ontology-based data access (OBDA) technologies. Such collaborations with private and public companies have allowed him to acquire significant experience in the field in developing OBDA solutions for real-world scenarios. In 2017, he (co)founded OBDA Systems, an innovative startup of Sapienza University, which proposes state-of-the-art solutions based on the most recent breakthroughs in the field of semantic technologies.