UAlacant word-level and phrase-level machine translation quality estimation systems at WMT 2016

Abstract

This paper describes the Universitat d'Alacant submissions (labeled as UAlacant) to the machine translation quality estimation (MTQE) shared task at WMT 2016, where we have participated in the word-level and phrase-level MTQE subtasks. Our systems use external sources of bilingual information as a black box to spot sub-segment correspondences between the source segment and the translation hypothesis. For our submissions, two sources of bilingual information have been used: machine translation (Lucy LT KWIK Translator and Google Translate) and the bilingual concordancer Reverso Context. Building upon the word-level approach implemented for WMT 2015, a method for phrase-based MTQE is proposed which builds on the probabilities obtained for word-level MTQE. For each sub-task we have submitted two systems: one using the features produced exclusively based on online sources of bilingual information, and one combining them with the baseline features provided by the organisers of the task.

UAlacant machine translation quality estimation at WMT 2018: a simple approach using phrase tables and feed-forward neural networks

“…The approach presented here builds on previous work by the same authors (Esplà-Gomis et al, 2015a,b;Esplà-Gomis et al, 2016) in which insertion positions were not yet predicted and a slightly different feature set was used. As in the original papers, here we use black-box bilingual resources from the Internet.…”

“…• Finally, Esplà-Gomis et al (2015a,b), andEsplà-Gomis et al (2016) perform word-level MT QE by using other MT systems to translate sub-segments of S and T and extracting features describing the way in which these translated sub-segments match sub-segments of T . This is the work most related to the one presented in this paper.…”

“…This paper describes the Universitat d'Alacant submissions to the word-and sentence-level machine translation (MT) quality estimation (QE) shared task at WMT 2018 (Specia et al, 2018). Our approach is an extension of a previous approach (Esplà-Gomis et al, 2015a,b;Esplà-Gomis et al, 2016) in which we simply marked the words t j of a machine-translated segment T as OK (no changes are needed) or as BAD (needing editing). Now we also mark the gaps γ j after each word t j as OK (no insertions are needed) or as BAD (needing the insertion of one or more words).…”