An induction system that learns programs in different programming languages using genetic programming and logic grammars
Man Leung Wong, Kwong‐Sak Leung · 2002
Genetic programming (GP) and inductive logic programming (ILP) have received increasing interest. Since their formalisms are so different these two approaches cannot be integrated easily though they share many common goals and functionalities. A unification will greatly enhance their problem solving power. Moreover, they are restricted in the computer languages in which programs can be induced. We present a flexible system called LOGENPRO (The logic grammar based genetic programming system) that combines GP and ILP. It is based on a formalism of logic grammars. The system can learn programs in various programming languages and represent context-sensitive information and domain-dependent knowledge. The performance of LOGENPRO in inducing logic programs from noisy examples is evaluated. A detailed comparison with FOIL has been conducted. This experiment demonstrates that LOGENPRO is a promising alternative to other inductive logic programming systems and sometimes is superior for handling noisy data. Moreover, a series of examples are used to illustrate that LOGENPRO is so flexible that programs in different programming languages including LISP, Prolog and Fuzzy Prolog can be induced.