… Though the main concerns of this thesis are the theoretical aspects of learning finite-state
machines, most of the algorithms presented in this thesis have been implemented and tested …

… computational and statistical aspects of realizable unsupervised proper learning of PDFA. …
the theoretical aspects of learning finite-state machines, most of the algorithms presented in …

… paper, we survey these generative objects and study their … will study the relation of probabilistic
finite-state automata with … recognition, statistical pattern recognition, machine translation, …

… the statistical statement of machine translation, we overview here how modeling, learning …
More specifically, we focus on stochastic finite-state transducers (SFST), which constitute an …

… on this formulation; additionally, we describe them using a recently introduced notion of
statistical query learning for distributions, which we believe makes for a clearer presentation. …

… Given a separating family, we will use z to denote the maximum number of elements in a …
Barzdin" [TB) contains a detailed analysis of statistical properties of finite state machines. …

… state graph for two important aspects: The identification of the terminal components, and the
… algorithms. In Section IV we illustrate how steady-state probabilities of finite state machines …

… aspects. First, we want to evaluate how the estimation error of the learning algorithm behaves
as we increase the training set … to compute the statistics on the larger training set (step 1 on …

… algorithms to compute the steady-state probabilities for very large finite state machines. These
… We are also working on the numerical aspects of the problem, namely aggregation and …

… Machine learning methods applied to motion capture data provide a way of doing this. This
paper presents a method for learning the parameters of a finite-state machine … other aspect of …