The Machine Learning research group is a subgroup of the DTAI research group which is part of the Department of Computer Science at the KU Leuven. It is led by Profs. Maurice Bruynooghe, Hendrik Blockeel, Jesse Davis, Bettina Berendt and Luc De Raedt and counts about 12 post-docs and 30 PhD students representing virtually all areas of machine learning and data mining. The group focuses on machine learning and data mining research involving structured data, symbolic, logical and probabilistic representations, background knowledge and applies it's techniques to challenging domains in the life sciences and action- and activity learning.
For an overview of research topics, take a look at the DTAI research overview.
History of the ML Research Group
Machine learning research at the KU Leuven was initiated in 1986, when Luc De Raedt started his PhD studies under the direction of Maurice Bruynooghe. In the 90s, the group rapidly gained recognition for its seminal contributions to inductive logic programming. During this period, members of the group developed a number of well-known systems such as Clint, Claudien, Tilde, Warmr and ICL, many of which are now gathered in our wide-spread ACE-tool, and the activities of the group rapidly expanded into domains such as reinforcement learning and distance-based learning. Since then, the group focuses on applications in the life sciences (especially chemo- and bio-informatics), in constraint-based data mining and inductive databases, and statistical relational learning (combining probabilistic models with logic). In 2006, with the return of Luc De Raedt to Leuven, the group of Hendrik Blockeel and Maurice Bruynooghe essentially merged with the former machine learning lab at the University of Freiburg. It is now one of the largest machine learning lab groups in Europe.
Members of the group have been involved as coordinators or principle investigators in many European projects, including the ESPRIT III and IV projects ILP (Inductive Logic Programming I and II) and Aladin, the EU FP5 and FP6 IST FET projects APRIL (Application of Probabilistic Inductive Logic Programming I and II), IQ (Inductive Querying), SolEuNet, and cInQ (Consortium on Inductive Querying), the EU FP7 IST FET project BISON, and the EU FP7 FET project ICON.
The group's faculty are action or area editors of journals such as Artificial Intelligence, Machine Learning, Journal of Machine Learning Research, Theory and Practice of Logic Programming and AI Communications.
Additionally, they have acted as organizers and program-co-chairs of key events such as:
- ECML 1993
- ILP 1995, 2014
- BNAIC 2002
- ECML/PKDD 2001/2003/2013
- ICML 2005
- ACAI 2007
- ILP 2007
- SRL/ILP/MLG 2009
- Benelearn 2010
- ECAI 2012
- IDA 2014.
- StarAI 2014 / 2015
The group is also the key organiser of a bi-annual workshop, now known as the Spring meeting on Machine Learning.
- ECCAI Dissertation award 2006/2009/2012/2013
- ACP Dissertation Award 2012
- IBM Belgium Dissertation award 2009/2014
- Best paper awards at ECML/PKDD, SIGKDD, ICTAI, ICPRAM, DS, ...
Machine Learning in a nutshell
Machine learning is the subfield of Artificial Intelligence and Computer Science that studies how machines can learn. A machine learns when it improves its performance on specific tasks with experience. In order to learn, machine learning methods analyze their past experience in order to find useful regularities, which explains why machine learning is closely related to data mining. The machine learning group is investigating all types of machine learning and data mining problems and techniques, though it focuses on dealing with structured data (such as graphs, trees and sequences), symbolic, logical and relational representations, and the use of knowledge and constraints. The group is well-known for its work on inductive logic programming, (statistical) relational learning, relational reinforcement learning, decision tree learning, graph mining, and inductive databases and constraint-based mining. It also studies applications in the life sciences and action- and activity learning.