On the Exact Learnability of Automata with Small Cover Time

We present algorithms for exactly learning unknown environments that can be described by deterministic nite automata. The learner performs a walk on the target automaton, where at each step it observes the output of the state it is at, and chooses a labeled edge to traverse to the next state. We assume that the learner has no means of a reset, and we also assume that the learner does not have access to a teacher that gives it counterexamples to its hypotheses. We present two algorithms, one assumes that the outputs observed by the learner are always correct and the other assumes that the outputs might be erroneous. The running times of both algorithms are polynomial in the cover time of the underlying graph of the target automaton.