Paper: Combining EBMT, SMT, TM and IR Technologies for Quality and Scale

ACL ID W12-0106
Title Combining EBMT, SMT, TM and IR Technologies for Quality and Scale
Venue Joint Workshop on Exploiting Synergies between Information Retrieval and Machine Translation (ESIRMT) and Hybrid Approaches to Machine Translation (HyTra)
Session
Year 2012
Authors

In this paper we present a hybrid statisti- cal machine translation (SMT)-example-based MT (EBMT) system that shows significant improvement over both SMT and EBMT base- line systems. First we present a runtime EBMT system using a subsentential transla- tion memory (TM). The EBMT system is fur- ther combined with an SMT system for effec- tive hybridization of the pair of systems. The hybrid system shows significant improvement in translation quality (0.82 and 2.75 abso- lute BLEU points) for two different language pairs (English?Turkish (En?Tr) and English? French (En?Fr)) over the baseline SMT sys- tem. However, the EBMT approach suffers from significant time complexity issues for a runtime approach. We explore two methods to make the system scalable at runtime. First, we use an heuristic-...