Nadir Durrani's dissertation has been selected as the best language technology / computational linguistics dissertation for the years 2011 to 2013 by the German Society for Computational Linguistics.
The dissertation proposes a new statistical model that learns translation and reordering patterns by representing them in a sequence of operations. The model was primarily designed to improve long distance reordering in German-to-English SMT; but it addresses numerous other drawbacks in Phrase-based SMT such as phrasal independence and spurious segmentation problems.
The model has been integrated into Moses, a state-of-the-art SMT system and has shown to significantly improve competition grade systems across 10 language pairs. The model is acknowledged as one of the prominent approaches that have led to actual improvements in systems in the large scale evaluation campaigns such as WMT and IWSLT.