Inference of residual finite-state tree automata from membership queries and finite positive data

Abstract

The area of Grammatical Inference centers on learning algorithms: Algorithms that infer a description (e.g., a grammar or an automaton) for an unknown formal language from given information in finitely many steps. Various conceivable learning settings have been outlined, and based on those a range of algorithms have been developed. One of the language classes studied most extensively with respect to its algorithmical learnability is the class of regular string languages.