Tree-based Hybrid Machine Translation
Andreas Søeborg Kirkedal · CBS Research Portal (Copenhagen Business School) · 2012
I present an automatic post-editing ap-proach that combines translation systems which produce syntactic trees as output. The nodes in the generation tree and target-side SCFG tree are aligned and form the basis for computing structural similar-ity. Structural similarity computation aligns subtrees and based on this alignment, sub-trees are substituted to create more accu-rate translations. Two different techniques have been implemented to compute struc-tural similarity: leaves and tree-edit dis-tance. I report on the translation quality of a machine translation (MT) system where both techniques are implemented. The ap-proach shows significant improvement over the baseline for MT systems with limited training data and structural improvement for MT systems trained on Europarl. 1