Summary: Effective Algorithmic Debugging for
Inductive Logic Programming
Jukka Paakki
Department of Computer Science and Information Systems
University of Jyv¨askyl¨a
P.O.Box 35, FIN­40351 Jyv¨askyl¨a, Finland
email: paakki@cs.jyu.fi
Tibor Gyim'othy Tam'as Horv'ath
Research Group on the Theory of Automata
Hungarian Academy of Sciences
Aradi v'ertan'uk tere 1, H­6720 Szeged, Hungary
email: fh42gyi,h158horg@ella.hu
Abstract
As suggested by Shapiro, an inductive learning process can be assisted by an algorithmic debug­
ging phase. Algorithmic debugging is a semi­automatic technique where the user acts as an oracle
who can answer queries concerning the intended meaning of the program on the abstract level of
procedures. The central problem with the basic algorithm is the potentially large number of queries
to the oracle. The method presented in this paper improves Shapiro's original scheme by integrating
testing and (static) slicing with algorithmic debugging. This integrated technique is more effective
than the basic one, in that the number of questions to the oracle is reduced. First, the oracle is not

Source: Alexin, Zoltán - Institute of Informatics, University of Szeged

Collections: Computer Technologies and Information Sciences