Neural Machine Translation with Phrase-Level Universal Visual Representations
This work addresses data sparsity in multimodal machine translation, which is a bottleneck for researchers and practitioners in AI and NLP, though it is incremental as it builds on existing retrieval and representation learning techniques.
The paper tackles the problem of multimodal machine translation (MMT) suffering from a shortage of paired sentence-image data by proposing a phrase-level retrieval-based method that learns visual representations from existing datasets, allowing MMT to break this limitation. Experiments show the method significantly outperforms strong baselines on multiple MMT datasets, particularly when textual context is limited.
Multimodal machine translation (MMT) aims to improve neural machine translation (NMT) with additional visual information, but most existing MMT methods require paired input of source sentence and image, which makes them suffer from shortage of sentence-image pairs. In this paper, we propose a phrase-level retrieval-based method for MMT to get visual information for the source input from existing sentence-image data sets so that MMT can break the limitation of paired sentence-image input. Our method performs retrieval at the phrase level and hence learns visual information from pairs of source phrase and grounded region, which can mitigate data sparsity. Furthermore, our method employs the conditional variational auto-encoder to learn visual representations which can filter redundant visual information and only retain visual information related to the phrase. Experiments show that the proposed method significantly outperforms strong baselines on multiple MMT datasets, especially when the textual context is limited.