THE USE OF RECURRENT NEURAL NETWORKS LANGUAGE MODEL IN TURKISH-ENGLISH MACHINE TRANSLATION

Turkish with its agglutinative structure needs speacial attention in machine translation research. In this work, a Turkish-English machine translation system is developed for TED1 talks with hierarchical phrase-based machine translation by employing morphological representation for Turkish. Then it is aimed to improve the translation performance by utilizing n-best rescoring with recurrent neural network language model and system combination together.