Data Science Lab – IDLab – Ghent University – imec

Data Science Lab

DSLab is a research group of Ghent University with a focus on machine learning, information retrieval, multimedia computing, exploratory data mining, data modelling and visualization, knowledge and reasoning, coding and information theory, robotics and control, and the use of high performance computing for these tasks. The group has world leading expertise in the entire data value chain, from data acquisition, storage, representation and coding, to mining and learning from data, and finally valorization. Reflecting this, we are engaged in a wide range of activities, including fundamental/basic research, applied research, and contract-based research as well as consultancy for industrial partners.

DSLab is now part of IDLab, one of the core research groups of IMEC, performing fundamental and applied research on internet technology and data science. At a national level IDLab is collaborating with about 175 industrial partners in many multidisciplinary projects. Internationally there is strong involvement in European research projects, especially in the Future Internet area, with about 70 finished and running EU projects over the past 10 years. IDLab counts about 300 researchers whose research resulted in 2350 publications in international journals and conference proceedings, 97 PhD’s, 45 international awards, and 8 spin-off companies.

The main DSLab competencies linked to DiSSeCt are to be found in the following fundamental research tracks:

  • Self-learning Intelligent Agents with Reasoning Capabilities: Today’s intelligent agents are limited in their abilities. By annotating the available interfaces on the web semantically an agent can reason autonomously  and discover how these services can be combined to reach a goal set by the end user (Restdesc). IDLab has been active in both the standardization of these annotations and has been working on the development of autonomous reasoning-agents for a wide variety of use cases.
  • Semantic Knowledge Generation and Publication at Scale: Creating semantically annotated Linked Data is often perceived as challenging by the human end-user. IDLab therefore has been actively developing tools to automate this data lifting process (RML) as far as possible and is researching annotation interfaces which make linked data generation an attainable goal even for the nontechnical end user. IDLab is one of the front-runners in publishing information on the web in a cost effective way and has pioneered a system (LDF)  which increases the availability of web data by adding the web clients to the querying process. These tools are continuously improved to better scale out, to allow query federation (i.e. query multiple sources at once) and also  to enable the interaction with streaming data.

Role in the project

The research of DSLab within DiSSeCt focusses on two main components:

  • Scalable and decentralized functional workflow engine: Central to the distributed semantic DiSSeCt services is the workflow engine, which will power the complex service composition. Nowadays, service providers face a management problem for their developed services and workflows using those developed services. An update of a service can slightly change its functionality and, moreover, can break workflows that make use of it. DSLab will target this problem by designing a workflow engine that can automatically compose workflows, based on service functionalities. This way, services with equal functionality will become interchangeable and updates of services that alter their functionality will not break existing workflows. This kind of workflow engine can help managing services developed by service providers. At the same time the workflows composed by the engine must be able to follow specific company policies to ease the integration of the engine into the companies’ infrastructure. A last feature of the workflow engine is that it must be able to easily combine different external resources. These resources can be information resources (metadata) or REST services. Key ingredient for our knowledge extraction is the combination of these disparate resources. Next to these innovative features in the workflow engine, this research will also focus on realizing a horizontally scalable decentralized workflow engine. As the number of resources to combine can be very high and higly federated (e.g. eHealth / Transport case, smart cities), the engine must be highly scalable.
  • Semantic service/data exposure: A key enabler to develop a workflow infrastructure that can compose workflows automatically and supports management of services and workflows is functional service descriptions. Hence methodologies to support automated generation of these data and service descriptions will be incepted by DSLab.

Researchers

Afbeeldingsresultaat voor erik mannens

Prof. Dr. Erik Mannens is professor at DSLab and business developer at imec. He has successfully managed +50 projects. He received his PhD degree in Computer Science Engineering (2011) at UGent, his Master’s degree in Computer Science (1995) at K.U. Leuven University, and his Master’s degree in Electro-Mechanical Engineering (1992) at KAHO Ghent.

Before joining UGent-imec in 2005 as research manager, he was a software engineering consultant and Java architect for over a decade. His major expertise is centered around big data analytics, metadata modeling, Semantic Web technologies, broadcasting workflows, iDTV and web development in general. He is involved in several projects as senior researcher and finished up his PhD on Semantic News Production; he was co-chair of the W3C Media Fragments Working Group and actively participating in other W3C’s semantic web standardization activities (Media Annotations, Provenance, Hydra, Linked Data Platform, and eGovernment).

Since 2008 Erik is paving the Open Data path in Flanders. He stood at the cradle of the first Hackatons and is a founding member of the Open Knowledge Foundation (Belgian Chapter). Since then, he is frequently invited as Open Data evangelist at national and international events. He currently actively participates in W3C’s eGov and Data On The Web working groups. Furthermore his team is owner of the Open Sourced Linked Open Data Publishing frameworks TheDataTank, R&Wbase, RML, and Linked Data Fragments.

On all of these subjects he has published +180 papers and book chapters. He is also member of the technical committee and/or organizing committee of several high level journals and conferences.

linkedin twitter website googlescholar

Afbeeldingsresultaat voor ruben verborgh

Prof. Dr. Ruben Verborgh is a professor of Semantic Web technology at IDLab, Ghent University – imec, and a research affiliate at the Decentralized Information Group of CSAIL at MIT. He is also a technology advocate for Inrupt, supporting the Solid ecosystem that gives you back control and choice—online and offline. He loves discussing about the Web, Linked Data, decentralization, Web APIs, hypermedia clients, and much more.

He aims to build a more intelligent generation of clients for a decentralized Web at the intersection of Linked Data and hypermedia-driven Web APIs. Through the creation of Linked Data Fragments, he introduced a new paradigm for query execution at Web-scale. He has co-authored two books on Linked Data, and contributed to more than 200 publications for international conferences and journals on Web-related topics.

linkedin twitter website googlescholar

Afbeeldingsresultaat voor miel vander sande Dr. Ing. Miel Vander Sande graduated as Bachelor in Multimedia and Communication technology and in 2010 as Master in Industrial Engineering: ICT. He wrote his Master thesis in collaboration with the University of Valencia, Spain about analysing and visualising RFID tracking data. After that, he concluded his education with a teachers degree in Informatics. Since september 2011, Miel joined DSLab as a researcher. His main interest and expertise are (linked open) data publishing (a.o. in the context of Open Knowledge Foundation), versioning, and querying on a Read/Write Web. He is active as Open Data activist in the Belgian and European community, supporting policy making and the organisation of events (such as App contests). Furthermore, he participates in the Semantic Web research community by organising several international workshops (WaSABi, SemDev, and NoISE) and as member of the W3C Linked Data Platform Working Group. Currently, Miel is active in multiple Flemish government projects for creating Linked Open Data ecosystems.

linkedin twitter website googlescholar
Afbeeldingsresultaat voor anastasia dimou Dr. Anastasia Dimou is a Post-Doc researcher at the Internet Technology & Data Science Lab at Gent University, Belgium. Anastasia joined the IDLab research group in February 2013. Her research expertise lies in the area of the Semantic Web, Linked Data Generation and Publication, Data Quality and Integration, Knowledge Representation and Management. She has broad experience on Semantic Wikis and Classification.  As part of her research, she investigated a uniform language for describing the mapping rules for generating high-quality Linked Data from multiple heterogeneous data formats and access interfaces and she also conducted research on Linked Data generation and publishing workflows. Her research activities led to the development of the RML tool chain (RMLProcessor, RMLEditor, RMLValidator, and RMLWorkbench). Anastasia has been involved in different national and l research projects and publications
linkedin twitter website googlescholar
doerthe Doerthe Arndt
linkedin twitter website googlescholar
Afbeeldingsresultaat voor joachim van herwegen Joachim Van Herwegen graduated as a master in Mathematical Computer Science. He joined IDLab in 2013 as a researcher. His research focuses on improving semantic querying efficiency and integrating semantic reasoning, and has presented his work several times at international conferences. Recently he helped create the semantic meta-querying engine Comunica to further support semantic querying on the Web.

twitter website googlescholar