A novel estimator based learning automata algorithm

Abstract

In this paper, we propose a novel algorithm to learn a Büchi automaton from a teacher who knows an ω-regular language. The algorithm is based on learning a formalism named family of DFAs (FDFAs) recently proposed by An-gluin and Fisman [10]. The main catch is that we use a classification tree structure instead of the standard observation table structure. The worst case storage space required by our algorithm is quadratically better than the table-based algorithm proposed in [10]. We implement the first publicly available library ROLL (Reg-ular Omega Language Learning), which consists of all ω-regular learning algorithms available in the literature and the new algorithms proposed in this paper. Experimental results show that our tree-based algorithms have the best performance among others regarding the number of solved learning tasks.

A Parameter-Free Gradient Bayesian Two-Action Learning Automaton Scheme

“…During the last decade, SE ri has been considered as the state-of-art algorithm for a long time, however, some recently proposed algorithms [8], [9] claim a faster convergence than SE ri . To make a comprehensive comparison among currently available techniques, as well as to verify the effectiveness of the proposed parameter-free scheme, in this section, PFLA is compared with several classic parameter-based learning automata schemes, including DP ri [20], DGPA [4], DBPA [6], DGCPA * [8], SE ri [5], GBSE [9] and LELA R [7].…”

“…MLE-based LA [4], [20] are proved to be a great success, achieving a tremendous improvement on the rate of convergence comparing with traditional variable structure stochastic automata. However, as we revealed in [8], MLE suffers from one principle weakness, i.e., MLE is unreliable when the quantity of samples is small.…”

“…The concept of stochastic estimator was employed in [5] so that the influence of lacking of samples can be reduced by introducing controlled randomnesses to MLE. In [8], we proposed an interval estimator based learning automata DGCPA, in which the upper bound of a 99% confidence interval of c i is used as estimates of reward probabilities. Both of these two LA schemes broke the records of convergence rate when proposed, which confirmed the defect of traditional MLE.…”

“…For deterministic estimator based learning automata, the smallest value of the resolution parameter that yielded a hundred percent accuracy in a certain number of experiments is selected. The situation is more sophisticated when concerning the stochastic estimator based schemes [5], [8], [9], because extra configurable parameters should be set to control the perturbation added. Parameter tuning is intend to balance the trade-off between speed and accuracy.…”

A novel parallel framework for pursuit learning schemes

“…Deterministic estimator based LA, such as DP ri [18], DGPA [19] and the newly presented LELA [20], DGCPA [21], are the major family of LA. SE ri [22], a very fast LA scheme, has an extra tunable parameter to control the randomness imposed to the deterministic estimates.…”