Ilknur Durgar El-Kahlout

Initial Explorations in English to Turkish Statistical Machine Translation

This paper presents some very preliminary results for and problems in developing a statistical machine translation system from English to Turkish. Starting with a baseline word model trained from about 20K aligned sentences, we explore various ways of exploiting morphological structure to improve upon the baseline system. As Turkish is a language with complex agglutinative word structures, we experiment with morphologically segmented and disambiguated versions of the parallel texts in order to also uncover relations between morphemes and function words in one language with morphemes and functions words in the other, in addition to relations between open class content words. Morphological segmentation on the Turkish side also conflates the statistics from allomorphs so that sparseness can be alleviated to a certain extent. We find that this approach coupled with a simple grouping of most frequent morphemes and function words on both sides improve the BLEU score from the baseline of 0.0752 to 0.0913 with the small training data. We close with a discussion on why one should not expect distortion parameters to model word-local morpheme ordering and that a new approach to handling complex morphotactics is needed.

TÜBİTAK SMT System Submission for WMT2016.

We describe the TÜBİTAK TurkishEnglish machine translation systems submissions in both directions for the WMT 2016: News Translation Task. We experiment with phrase-based and hierarchical phrase-based systems for both directions using word-level and morpheme-level representations for the Turkish side. Finally we perform system combination which results in 0.5 BLEU increase for Turkishto-English and 0.3 BLEU increase for English-to-Turkish.

Aligning turkish and english parallel texts for statistical machine translation

This paper presents a preliminary work on aligning Turkish and English parallel texts towards developing a statistical machine translation system for English and Turkish. To avoid the data sparseness problem and to uncover relations between sublexical components of words such as morphemes, we have converted our parallel texts to a morphemic representation and then used standard word alignment algorithms. Results from a mere 3K sentences of parallel English–Turkish texts show that we are able to link Turkish morphemes with English morphemes and function words quite successfully. We have also used the Turkish WordNet which is linked with the English WordNet, as a bootstrapping dictionary to constrain root word alignments.