论文信息 - Induction of Prolog Programs with Markus

Induction of Prolog Programs with Markus

This paper describes empirical Inductive Logic Programming system Markus, which was designed as a shell for experimenting in different real-life domains. The current version of the system uses as its basis ’covering’ paradigm (also used by some other systems, e.g. Quinlan’s Foil). Within this paradigm, the development of single program clauses is performed by iterative deepening search of optimally generated refinement graphs (also used in Shapiro’s Model Inference System). The system was successfully applied to several nontrivial learning tasks: construction of qualitative models, program verification, formula invention and construction of some typical Prolog programs, the most difficult being the quick-sort.

Marko Grobelnik | M. Grobelnik

[1] Ehud Shapiro,et al. Algorithmic Program Debugging , 1983 .

[2] Yves Deville,et al. Logic Program Synthesis from Incomplete Specifications , 1993, J. Symb. Comput..

[3] Alan W. Biermann,et al. Induction of schemata for program synthesis , 1990 .

[4] Ivan Bratko,et al. Inductive Learning Applied to Program Construction and Verification , 1992, AIFIPP.

[5] Ehud Shapiro,et al. Third International Conference on Logic Programming , 1986 .

[6] Ivan Bratko,et al. Learning Qualitative Models of Dynamic Systems , 1994, ML.

[7] Ehud Shapiro,et al. Inductive Inference of Theories from Facts , 1991, Computational Logic - Essays in Honor of Alan Robinson.

[8] J. Ross Quinlan,et al. Determinate Literals in Inductive Logic Programming , 1991, IJCAI.

[9] Leon Sterling,et al. Refinement Strategies for Inductive Leaming of Simple Prolog Programs , 1991, IJCAI.

[10] Matthew M. Huntbach. An Improvised Version of Shapiro's Model Inference System , 1986, ICLP.

[11] Stephen Muggleton,et al. Efficient Induction of Logic Programs , 1990, ALT.

[12] Saso Dzeroski,et al. Learning Nonrecursive Definitions of Relations with LINUS , 1991, EWSL.