abstract = "The process of learning often consists of Inductive
Inference, making generalisations from samples. The
problem here is finding generalizations (Grammars) for
Formal Languages from finite sets of positive and
negative sample sentences. The focus of this paper is
on Context-Free Languages (CFLs) as defined
Context-Free Grammars (CFGs), some of which are
accepted by Deterministic Push-Down Automata (D-PDA).
This paper describes a meta-language for constructing
D-PDAs. This language is then combined with Genetic
Programming to evolve D-PDAs which accept languages.
The technique is illustrated with two favourite CFLs.",