Towards Personalised and Document-level Machine Translation of Dialogue

TL;DR

This paper discusses advancing personalized and document-level neural machine translation for dialogue, focusing on integrating contextual information, improving cohesion translation, and establishing reliable evaluation metrics across multiple languages.

Contribution

It proposes methods to incorporate extra-textual context into NMT, enhances translation of cohesion devices, and develops evaluation metrics for PersNMT and DocNMT in dialogue domains.

Findings

Improved translation accuracy with context integration

Enhanced cohesion device translation quality

Established evaluation metrics for personalized and document-level NMT

Abstract

State-of-the-art (SOTA) neural machine translation (NMT) systems translate texts at sentence level, ignoring context: intra-textual information, like the previous sentence, and extra-textual information, like the gender of the speaker. Because of that, some sentences are translated incorrectly. Personalised NMT (PersNMT) and document-level NMT (DocNMT) incorporate this information into the translation process. Both fields are relatively new and previous work within them is limited. Moreover, there are no readily available robust evaluation metrics for them, which makes it difficult to develop better systems, as well as track global progress and compare different methods. This thesis proposal focuses on PersNMT and DocNMT for the domain of dialogue extracted from TV subtitles in five languages: English, Brazilian Portuguese, German, French and Polish. Three main challenges are addressed: (1) incorporating extra-textual information directly into NMT systems; (2) improving the machine translation of cohesion devices; (3) reliable evaluation for PersNMT and DocNMT.