Analyzing Optimization for Statistical Machine Translation: MERT Learns Verbosity, PRO Learns Length.

Francisco Guzmn, Preslav Nakov, Stephan Vogel

Browse the full CoNLL paper archive.