The RWTH Aachen Machine Translation System for WMT 2011

This paper describes the statistical machine translation (SMT) systems developed by RWTH Aachen University for the translation task of the EMNLP 2011 Sixth Workshop on Statistical Machine Translation. Both phrasebased and hierarchical SMT systems were trained for the constrained German-English and French-English tasks in all directions. Experiments were conducted to compare different training data sets, training methods and optimization criteria, as well as additional models on dependency structure and phrase reordering. Further, we applied a system combination technique to create a consensus hypothesis from several different systems.