Learning from a Smarter Teacher
We present a new learning algorithm introducing a helpful teacher who models the learners’ knowledge. Our algorithm, called learning from extensions (LEX), learns finite-state transducers using only one type of query called extension query. Our query was inspired by equivalence queries and counterexamples, but we show in this article that it is possible to learn efficiently finite state transducers using only extension queries (it is known that only with membership queries or only with equivalence queries is not possible). The teacher answers an extension query by connecting the new information asked by the learner with the information that the learner already knows. We prove that our new algorithm LEX discovers a target finite-state transducer in polynomial time. We also discuss briefly several complexity aspects and we give an example.
KeywordsMathematical Linguistics Automaton Learn Input Alphabet Membership Query Deterministic Finite Automaton
Unable to display preview. Download preview PDF.
- 5.Balcazar, J.L., Diaz, J., Gavalda, R.: Algorithms for learning finite automata from queries: A unified view. In: Advances in Algorithms, Languages, and Complexity, pp. 53–72 (1997)Google Scholar
- 6.Angluin, D.: Negative results for equivalence queries. Machine Learning 5(2), 121–150 (1990)Google Scholar