The AMU-UEdin Submission to the WMT 2017 Shared Task on Automatic Post-Editing

Proceedings of the Conference on Machine Translation (WMT), Volume 2: Shared Task Papers, pages 639–646 Copenhagen, Denmark, September 711, 2017. c⃝2017 Association for Computational Linguistics The AMU-UEdin Submission to the WMT 2017 Shared Task on Automatic Post-Editing Marcin Junczys-Dowmunt Information Systems Laboratory Adam Mickiewicz University in Pozna´n junczys@amu.edu.pl Roman Grundkiewicz School of Informatics University of Edinburgh rgrundki@exseed.ed.ac.uk Abstract This work describes the AMU-UEdin sub- mission to the WMT 2017 shared task on Automatic Post-Editing. We explore mul- tiple neural architectures adapted for the task of automatic post-editing of machine translation output. We focus on neural end-to-end models that combine both in- puts mt and src in a single neural architec- ture, modeling {mt, src} → pe directly. Apart from that, we investigate the influ- ence of hard-attention models which seem to be well-suited for monolingual tasks, as well as combinations of both ideas. 1 Introduction During the WMT 2016 APE two systems relied on neural models, the CUNI system (Libovický et al., 2016) and the shared task winner, the sys- tem submitted by the Adam Mickiewicz Univer- sity (AMU) team (Junczys-Dowmunt and Grund- kiewicz, 2016). This submission explored the ap- plication of neural translation models to the APE problem and achieved good results by treating different models as components in a log-linear model, allowing for multiple inputs (the source src and the translated sentence mt) that were de- coded to the same target language (post-edited translation pe). Two systems were considered, one using src as the input (src → pe) and another us- ing mt as the input (mt → pe). A simple string- matching penalty integrated within the log-linear model was used to control for higher faithfulness with regard to the raw MT output. The penalty fires if the APE system proposes a word in its out- put that has not been seen in mt. The influence of the components on the final result was tuned with Minimum Error Rate Training (Och, 2003) with regard to the task metric TER. With neural encoder-decoder models, and multi-source models in particular, the combination of mt and src can be now achieved in more natu- ral ways than for previously popular phrase-based statistical machine translation (PB-SMT) systems. Despite this, results for multi-source or double- source models in APE scenarios are incomplete or unsatisfying in terms of performance. In this work, we explore a number of single- source and double-source neural architectures which we believe to be better fits to the APE task than vanilla encoder-decoder models with soft at- tention. We focus on neural end-to-end models that combine both inputs mt and src in a single neural architecture, modeling {mt, src} → pe di- rectly. Apart from that, we investigate the influ- ence of hard-attention models which seem to be well-suited for monolingual tasks. Finally, we cre- ate combinations of both architectures. Following (Junczys-Dowmunt and Grund- kiewicz, 2016), we also attempt to generate more artificial data for the task. Instead of relying on filtering towards specific error rates, we generate text with …

Content truncated to 500 words