Work Area: Machine Learning
Keywords induction, machine learning, inductive logic programming, logic programming, learning from example, knowledge-intensive learning
Start Date: 1 September 92 / Duration: 36 months / Status: running
[ participants / contact ]
Abstract Inductive logic programming (ILP) is the intersection of inductive learning and logic programming. The project focuses on the following research topics: theory of ILP, theory revision and multiple predicate learning, handling imperfect data, predicate invention, and declarative bias. The project will address theoretical issues and implementation of prototype learners and will carry out empirical evaluations.
The main long term technical goal of the ILP project is to upgrade the techniques of the classical empirical learning paradigm to a logic programming framework. In this way ILP aims to overcome the two main limitations of classical empirical or similarity based learning algorithms, such as the TDIDT-family: the use of a limited knowledge representation formalism (essentially a propositional logic), and the inability to use substantial background knowledge in the learning process.
The project focuses on the following research topics:
Theoretical results obtained address:
Results in Theory revision includes:
A better understanding of predicate invention was obtained:
Results on handling imperfect data include:
Results on declarative bias include:
The expected outcome of the project is a sound basis for the development of systems that are able to induce logic programs from examples in real-life applications that involve substantial amounts of background knowledge.
Members of the project have given or will give invited talks and/or tutorials on inductive logic programming at major conferences and summerschools devoted to machine learning and/or logic programming (including ECML, ISMIS, ILPS, LLI, SCAI).
The third (Bled, 93) and fourth (Germany, 94) workshops on ILP are organised by partners in the project, and an IJCAI 93 workshop is devoted to ILP.
At the past European Conference on Machine Learning, about one third of the papers was related to ILP.
Further publications by the consortium will appear at major conferences (such as IJCAI, ICML) and major journals (AIJournal, JETAI, IEEE Transactions on Knowledge and Data Engineering, etc.)
Several ILP tools, technical reports and overviews of ILP can be obtained from the consortium.
Further information about ILP is available from the ILP home page <URL:http://www.cs.kuleuven.ac.be/~ml/esprit.ilp.6020.html>.
Katholieke Universiteit Leuven - B
Department of Computing Science
B - 3001 Heverlee
Gesellschaft für Mathematik und Datenverarbeitung - D
Universität Stuttgart - D
CNRS - LRI - F
Università di Torino - I
University of Oxford - UK
University of Stockholm - S
Prof. Dr. Ir. M. Bruynooghe
tel +32/16 20 10 15
fax +32/16 20 53 08
ILP - 6020, August 1994
please address enquiries to the ESPRIT Information Desk
html version of synopsis by Nick Cook