Survey: Weighted Extended Top-Down Tree Transducers Part III - Composition

Aurélie Lagoutte (1) , Andreas Maletti (2)

1 ENS Cachan - École normale supérieure - Cachan
2 Institute for Natural Language Processing [Stuttgart]

In this survey (functional) compositions of weighted tree transformations computable by weighted extended top-down tree transducers are investigated. The existing results in the literature are explained and illustrated. It is argued, why certain compositions are not possible in the general case, and 3 informed conjectures provide an insight into potentially 3 new composition results that extend and complement the existing results. In particular, if all were true, then the beautiful symmetry in the composition results for weighted top-down and bottom-up tree transducers would be recovered.

Mots clés: weighted tree transducer; top-down tree transducer; composition; deletion; copying

Domaines: Théorie et langage formel [cs.FL]

Aurélie Lagoutte, Andreas Maletti. Survey: Weighted Extended Top-Down Tree Transducers Part III - Composition. Algebraic Foundations in Computer Science, 7020, pp.272 - 308, 2011, 978-3-642-24896-2. ⟨10.1007/978-3-642-24897-9_13⟩. ⟨hal-01508560⟩