Projects

Ongoing

Declarative problem solving is on the brink of a breakthrough. By applying novel language and inference technologies from knowledge representation and reasoning (KRR), a formal theory expressing application knowledge can drive many functionalities in a software system. In this project, the KRR subgroup of DTAI (Prof. Denecker and Prof. Janssens) intends to push this technology substantially beyond its current limitations by working on theory, languages and algorithms. The current technology is best suited for knowledge intense applications such as business applications. So far the novel technologies of KRR have not or only partially been studied in this field. Thereofore, KRR joins forces with the LIRIS group (Prof. Vanthienen) to develop this approach in business problems. To warrant the most impact in the economically important world of business software, an important subgoal will be to blend the new technologies with the recent influential OMG DMN (Decision Modeling & Notation) standard.

Companies can obtain significant business value from the automation of routine decision making and advice in standard, day-to-day operating procedures, laws and regulations. Knowledge-based AI techniques allow to offer automated, correct and reliable online advice/decisions for numerous routine cases in areas such as insurance, energy premiums, tax and social security benefits, pricing and discounts, eligibility decisions, etc. The knowledge builds on explicit rules and regulations that can be modeled and exploited in multiple ways. This knowledge is not just to be learned from data, as it has to be compliant to existing rules and regulations, and has to be fully explainable, which often is a challenge in data-based AI. The modelling approach is enabled in practice by the DMN (Decision Model and Notation) standard, co-created by the co-promoter. DMN is already heavily used in banking, insurance, social security and standard procedures in general. The research groups LIRIS and EAVISE have significant expertise when it comes to building these models in a correct and consistent manner, and in executing them in different ways using state-of-the-art Knowledge Representation technology. In the project we want to combine our expertise into a full-track integration from modelling to deployment and apply it to real cases, by producing an operational decision modelling methodology.

  • FWO Flanders project (2019-2023) – promotor, “PRODIGY: Process-Decision Integration for Knowledge-intensive Process Modeling and Discovery” (01/01/2019 – 31/12/2022)

Modeling and automating business processes is important for improving efficiency and effectiveness in businesses. During the execution of a business process, a plethora of business decisions need to be made in order to ensure the correct enactment of the business process. With increasing automation of business processes, the need for the automation of routine business decisions grows, as business processes can contain a lot of decisions or can even be the enactment of one major decision.

Despite extensive research on how to design process models, the question of integrated process and decision design has received less attention in literature. However, the integration of decisions with processes is of paramount importance for knowledge-intensive processes. The knowledge that is encapsulated in the decisions is strongly needed for automation, since decision-making (in e.g. online and mobile applications) needs to be faster and more efficient. Therefore, it is important that the decision logic can be formulated in a correct, flexible and maintainable way. Sound decision modeling and mining techniques are required that allow to unambiguously describe and implement the decisions deployed in business processes and systems. Although a lot of research has already been devoted to process modeling and mining quality criteria, there is still a lot of research to be done on the decision side, and especially on the integration between decisions and processes, bothin modeling and mining.

The aim of this project is to develop an integrated decision-driven process modeling and mining approach, and a set of new complexity metrics for this approach.

Past Projects

Research Projects:

  • TETRA (Technology Transfer) project (2017-2019) – co-promotor, “Decision Analytics”

Enterprises often experience difficulties coping with complex decision-making processes. The new Decision Modeling and Notation (DMN) standard tries to improve this by also allowing business users without ICT background to model decisions and decision making processes. This allows them to analyze them better, and to automate them (partly). The overall goal of this project is to help companies to more and better use DMN.

  • Fund for Scientific Research – Flanders (F.W.O.-Vlaanderen), Gotcha: fraud detection through network analysis, 2015-2018

Fraud is a widespread phenomenon, present in many economic sectors and applications. Nowadays, fraudulent structures are getting more complex, and fraudsters often operate in a strongly interconnected community. They are part of a web with fraud spreading in a viral like manner. As such, new approaches and methodologies are required to introduce and define the right network representation and analyze fraud as a contagious virus infecting related entities. The main aim of this doctoral research proposal is to develop a new, scalable and integrated algorithm to detect fraud more efficiently by exploiting the interrelated structure between these entities. The developed technique should not only evaluate the direct influences of individual entities on each other, but also take into account the entire network structure. Apart from individual fraudulent behavior, fraudulent community detection addresses the curtailment of further expansion of fraudulent groups into the network. Moreover, fraudulent behavior changes over time and is highly adaptive to a changing environment. A thorough understanding of the mechanisms that govern the processes in networks will allow to control the dynamics of fraud, and face the time-evolving character of fraud. Starting from a theoretical perspective, the acquired knowledge and tools developed in this project will be implemented in practice, using an extensive fraud data set of the Belgian governmental NSSO (National Social Security Office).

Promotor: Baesens, B.
Co-Promotors: Vanthienen, J., Snoeck, M.

  • Fund for Scientific Research – Flanders (F.W.O.-Vlaanderen), Project FWO G0804 13N, Improving Process Modeling and Control using Advanced Business Process Analytics and Enriched Event Logs, 2013-2016.

The fields of business process analytics (including process mining) and business process management have become essential in supporting business operations. With the rise of process-aware information systems, a vast amount of design, analysis and visualization tools has become available in order to model business processes, monitor key performance indicators, and perform deeper analysis tasks (i.e. process mining). The main goal of this research proposal is to obtain fitter process models, by bringing current business process analytics to a more robust level via the enrichment of event logs, which serve as the input for such techniques, thereby leading to better insights and to the improvement of process modeling and process control.

Promotor: Vanthienen, J.
Co-Promotor: Poels, G. (UGent)
Researcher: N

  • Research Fund K.U.Leuven, OT/10/010: Business Process Mining: new techniques and evaluationmetrics, 2010-2014

A new and promising way of acquiring insightsinto business processes is the analysis of the event logs of information and/orenterprise resource planning (ERP) systems, hereby verifying process compliance(expected behavior versus portrayed behavior). Building upon previous research describing techniques that representprocess mining as a first-order classification problem on event logssupplemented with artificial negative events, it is examined how rule inductiontechniques can be applied to various process mining tasks. The novelty of theproposed research consists of a process mining approach in which rules andgenerated negative events have a prominent place, by considering a processinstance as a trajectory in a state space that is spanned by the domains of thedifferent activities, events and business concepts belonging to a specificprocess. The proposed research also tries to establish a general framework forthe evaluation of induced process models with negative events and attempts todefine a set of new metrics. The theoretical contributions will be empiricallyvalidated for applications in audit compliance, risk management and healthcare,and also for semi-structured processes.

Promotor: Vanthienen, J.
Co-Promotor: Baesens, B.
Researchers: Caron, F.; vanden Broucke, S.

  • (2009) Bilateral scientific cooperation project K.U.Leuven – Tsinghua University, Project 3H051154, Intelligent Enterprise Resource Planning systems, 2009 – 2010, Promotor: J. Vanthienen
  • (2006) Nationale Bank van België – NB/06/08, Het ontwikkelen en valideren van risicomodellen in de context van Basel II, 2007, Promotor: J. Vanthienen, B. Baesens; Research Fellow: J. Huysmans
  • (2005) Fund for Scientific Research – Flanders (F.W.O.-Vlaanderen), Project FWO G.0615.05, Using business intelligence techniques for risk profiling of economic entities, 2005 – 2008, Promotor: J. Vanthienen; Research Fellow: D. Martens
  • (2004) Belgian Science Policy Office (AGORA project) – AG/01/06, INFO-NS: Intelligent exploitation tools for nonstructured information for the Belgian federal police, 2004-2005, Co-promotor: J. Vanthienen; Research Fellow: N. Kumar
  • (2003) Research Fund K.U.Leuven OT/03/12, Knowledge discovery for customer scoring using neural networks and support vector machines, 01/10/2003-30/09/2007, Promotor: J. Vanthienen; Research Fellow: J. Huysmans
  • (2003) Project DWTC, Agora, AG/031/082, Oprichting van een centrale databank voor de coördinatie van de controles van het personen- en goederenvervoer (Directie Controle & Transportorganisatoren van het Directoraat generaal Vervoer te Land van de FOD Mobiliteit en Vervoer), 01/01/2003-31/12/2003 FEDERAL OFFICE FOR SCIENTIFIC, TECHNICAL AND CULTURAL AFFAIRS, Promotor: J. Vanthienen; Research Fellow: Ch. Mues
  • (1997) Project BWTS97/01 Bilaterale Wetenschappelijke en Technologische Samenwerking met Polen (Wroclaw University of Economics) “Knowledge acquisition and intelligent distributed learning in resolving managerial issues”, Vlaamse regering, administratie wetenschap en innovatie, 10/12/97-10/3/2001, hoofdpromotor: L.U.C., Co-promotor: J. Vanthienen
  • (1995) Research Project F.W.O. nr. G.O135.95 “Ontwerp van methoden voor ontwikkeling, validatie en implementatie van kennisgebaseerde informatiesystemen”, 1/1/95-31/12/00, Promotor: J. Vanthienen, co-promotor: M. Verhelst; Research Fellow: Ch. Mues, G. Wets
  • (1994) Research Fund K.U.Leuven OT/94/2: Development of methods for the design, validation and implementation of knowledge based information systems, 1994, Promotor: J. Vanthienen, co-promotor: M. Verhelst; Research Fellow: A. Aerts, Ch. Mues

Educational Projects:

  • (2007) OOF project Onderwijsontwikkelingsfonds Associatie K.U.Leuven – OOF2007/29, Persoonlijk Informatiebeheer, 2007-2009, Co-promotor: J. Vanthienen
  • (2005) Research Grants K.U.Leuven Educational Council, Project JVT/DV/OOI/2005, MIRO: Modeloplossingen van bestuurlijke Informatiesystemen in een inteRactief geïntegreerd Onderwijsplatform, (2003) 2005 – 2007, Promotor: J. Vanthienen; Research Fellow: B. Weynants
  • K.U.Leuven convenantproject: CHEOPS, 2003-2005, Promotor: J. Vanthienen; Research Fellow: J. Tisaun
  • (2000) OOI-project K.U.Leuven: Integrated implementation of self-test environments in the ETEW curricula (GIZEH), 2000–2002, RESEARCH GRANT K.U.LEUVEN EDUCATIONAL COUNCIL, Promotor: J. Vanthienen; Research Fellow: D. Vanderbist
  • (1995) Educational project Impulsprogramma Onderwijs Departement TEW, 1995 “Bevorderen van zelfstudie door middel van uitwisseling via het Internet van extra studiemateriaal, Promotors: J. Vanthienen; F. Put
  • (1994) Educational project ‘onderwijsvernieuwing en zelfstudie’ Onderwijsraad K.U.Leuven (GeBu 479/l/1117)
  • Computergesteund onderwijs van informatica-toepassingen, (1 februari 1994-31 januari 1995), Promotor: J. Vanthienen; Research Fellow: G. Wets