Learning Probabilistic Automata and Markov Chains via Queries

摘要

We investigate the problem of learning probabilistic automata and Markov chains via queries in the teacher-student learning model. Probabilistic automata and Markov chains are probabilistic extensions of finite state automata and have similar structures. We discuss some natural oracles associated with probabilistic automata and Markov chains. We present polynomial-time algorithms for learning probabilistic automata and Markov Chains using these oracles.