DFKI-LT - Linguistically-Augmented Bulgarian-to-English Statistical Machine Translation Model
Linguistically-Augmented Bulgarian-to-English Statistical Machine Translation Model
1 Proceedings of the Joint Workshop on Exploiting Synergies between Information Retrieval and Machine Translation (ESIRMT) and Hybrid Approaches to Machine Translation (HyTra),
In this paper, we present our linguistically-augmented statistical machine translation model from Bulgarian to English, which combines a statistical machine translation (SMT) system (as backbone) with deep linguistic features (as factors). The motivation is to take advantages of the robustness of the SMT system and the linguistic knowledge of morphological analysis and the hand-crafted grammar through system combination approach. The preliminary evaluation has shown very promising results in terms of BLEU scores (38.85) and the manual analysis also confirms the high quality of the translation the system delivers.
Files: BibTeX, W12-0116, W12-0116.pdf