Tree Transducers in Machine Translation
Abstract:
We review several model of tree transducers with respect to properties
that are relevant in machine translation. These properties include:
suitable expressiveness, symmetry, preservation of regularity, and
closure under composition. For three tree transducer models, top-down
tree transducers, extended top-down tree transducers, and extended
multi bottom-up tree transducers, the relevant results are presented
in an informal and illustrative manner. The aim is provide a synopsis
that illustrates how theory (tree transducers) and practice (machine
translation) interact on this particular example.
For scheduling information, please see the Stuttgart reading group page.