Ad Astra Awards
Ad Astra Journal
Science library
White book
University rankings
Who's who
Theses and dissertations
Ad Astra association
Press releases
Funding opportunities
>> Românã

Gabriel Istrate. Grammatical Inference and Symbolic Dynamics. In 5th European Conference on Complex Systems (ECCS' 08), Jerusalem, Israel, September 14-19, 2008. 2008.

Abstract: [Published only as an abstract on the conference website. Full-length paper to follow]

We show that Lindenmayer systems can be learned in polynomial time from membership and equivalence queries (that is queries are syntax trees of the corresponding grammar).

The result is directly motivated by (and can be applied to) the problem of
learning the symbolic representations associated to dynamical systems induced by unimodal (and several classes of multimodal) maps. Particular cases
of interest include the ones when they arise via operators on kneading sequences, such as renormalization and
composition rules. Learning syntax trees amounts to learning these structural mechanisms.

Keywords: grammatical inference, ET0L languages, applied symbolic dynamics


Posted by Gabriel Istrate


© Ad Astra 2001-2013