Learning in Rich Representations: Inductive Logic Programming and Computational Scientific Discovery

• Encourage more research on learning in rich representations, such as relational representations and differential equations, which can be used for modeling a variety of real world problems. Inductive logic programming (ILP) is concerned with learning from data and domain knowledge in relational representations. ILP started off by addressing the task of learning logic programs from examples and background knowledge (Muggleton 1992; Lavrač and Džeroski 1994; De Raedt 1996). Recent developments, however, have broadened its scope to address a variety of learning tasks in relational representations. A significant part of ILP research now goes under the heading (Multi)Relational Data Mining – (M)RDM (Džeroski and Lavrač 2001) – and is concerned with finding patterns such as relational association rules and relational decision trees from multi-table relational databases. As another example, ILP has been also used in a reinforcement learning context (Džeroski et al. 1998; 2001; Driessens and Džeroski 2002).