Distilling Translations with Visual Awareness
This addresses the challenge of effectively integrating visual information in machine translation for ambiguous cases, though it appears incremental as it builds on prior work.
The paper tackles the problem of multimodal machine translation by proposing a translate-and-refine approach that uses images in a second stage decoder to improve translations, achieving state-of-the-art results and demonstrating recovery from source language errors.
Previous work on multimodal machine translation has shown that visual information is only needed in very specific cases, for example in the presence of ambiguous words where the textual context is not sufficient. As a consequence, models tend to learn to ignore this information. We propose a translate-and-refine approach to this problem where images are only used by a second stage decoder. This approach is trained jointly to generate a good first draft translation and to improve over this draft by (i) making better use of the target language textual context (both left and right-side contexts) and (ii) making use of visual context. This approach leads to the state of the art results. Additionally, we show that it has the ability to recover from erroneous or missing words in the source language.