Attribute-Value Learning Versus Inductive Logic Programming: The Missing Links (Extended Abstract)

Two contributions are sketched. A first contribution shows that a special case of relational learning can be transformed into attribute-value learning. However, it is much more tractable to stick to the relational representation than to apply the sketched transformation. This provides a sound theoretical justification for inductive logic programming. In a second contribution, we show how existing attribute-value learning techniques and systems can be upgraded towards inductive logic programming using the ‘Leuven’ methodology and illustrate it using the Claudien, Tilde, ICL, Warmr, TIC, MacCent and RRL systems.