Focused Concatenation for Context-Aware Neural Machine Translation
This work addresses the need for better context handling in machine translation, offering incremental improvements for translation systems.
The paper tackled the problem of context-aware neural machine translation by proposing an improved concatenation method that focuses on translating the current sentence while discounting context loss, and it demonstrated superiority over baseline and other systems in translation quality and discourse phenomena.
A straightforward approach to context-aware neural machine translation consists in feeding the standard encoder-decoder architecture with a window of consecutive sentences, formed by the current sentence and a number of sentences from its context concatenated to it. In this work, we propose an improved concatenation approach that encourages the model to focus on the translation of the current sentence, discounting the loss generated by target context. We also propose an additional improvement that strengthen the notion of sentence boundaries and of relative sentence distance, facilitating model compliance to the context-discounted objective. We evaluate our approach with both average-translation quality metrics and contrastive test sets for the translation of inter-sentential discourse phenomena, proving its superiority to the vanilla concatenation approach and other sophisticated context-aware systems.