Stanford University's Arabic-to-English Statistical Machine Translation System for the 2009 NIST MT Open Evaluation

This document describes Stanford Univer- sity's first entry into a NIST Arabic-English MT evaluation. We describe two main im- provements over a previous Chinese-English submission (Galley et al., 2008): a hierarchi- cal lexicalized reordering model (Galley and Manning, 2008) and a technique for perform- ing minimum error rate training (Cer et al., 2008) that outperforms the standard Powell method.