Proceedings of the Third Workshop on Statistical Machine Translation.

Chris Callison-Burch, Philipp Koehn, Christof Monz, Josh E. Schroeder, Cameron Shaw Fordyce

2008 · 36 citations · 0 references

Concepts

Abstract

The ACL 2008 Workshop on Statistical Machine Translation (WMT-08) took place on Thursday, June 19 in Columbus, Ohio, United States, immediately following the annual meeting of the Association for Computational Linguistics, which was hosted by the Ohio State University. This is the third time this workshop has been held. It has its root in the ACL 2005 Workshop on Building and Using Parallel Texts In the following years the Workshop on Statistical Machine Translation was held at HLT-NAACL 2006 in New York City, US, and at ACL 2007 in Prague, Czech Republic. The focus of our workshop was to use parallel corpora for machine translation. Recent experimentation has shown that the performance of SMT systems varies greatly with the source language. In this workshop we encouraged researchers to investigate ways to improve the performance of SMT systems for diverse languages, including morphologically more complex languages and languages with partial free word order. Prior to the workshop, in addition to soliciting relevant papers for review and possible presentation we conducted a shared task that brought together machine translation systems for an evaluation on previously unseen data. This year's task resembled the shared tasks of previous years in many ways, but also included Hungarian-English and Spanish-German as new language pairs. In addition, we evaluated submitted systems against new test sets from the newswire domain. The results of the shared task were announced at the workshop, and these proceedings also include an overview paper for the shared task that summarizes the results, as well as provides information about the data used and any procedures that were followed in conducting or scoring the task. In addition, there are short papers from each participating team that describe their underlying system in some detail. Due to the large number of high quality submission for the full paper track, shared task submissions were presented as posters. The poster session was held in the afternoon and gave participants of the shared task the opportunity to present their approaches. The rest of the day was devoted to oral paper presentations and Daniel Marcu's invited talk in the afternoon. Like in previous years, we have received a far larger number of submission than we could accept for presentation. This year we have received 18 full paper submissions and 26 shared task submissions. In total WMT-08 featured 12 full paper oral presentations and 25 shared task poster presentations. The invited talk was given by Daniel Marcu of the Information Sciences Institute at the University of Southern California.