Learning k-edge Deterministic Finite Automata in the Framework of Active Learning

摘要

One of most attractive topics in grammatical inference is theoretically study on learnability of some classes of automata corresponding with defined formal languages. For last two decades a number of theoretical results have been reported and played as essential knowledge for applications in other fileds such as speech recognition and music-style recognition. In this paper, we consider the problem of learning k-edge determinisitc finite automata in the framework of active learning. The results show that the class of k-edge deterministic finite automata identification in the limit with membership queries and equivalence queries.