m3P: Towards Multimodal Multilingual Translation with Multimodal Prompt

arXiv:2403.17556v1 Announce Type: new
Abstract: Multilingual translation supports multiple translation directions by projecting all languages in a shared space, but the translation quality is undermined by the difference between languages in the text-only modality, especially when the number of languages is large. To bridge this gap, we introduce visual context as the universal language-independent representation to facilitate multilingual translation. In this paper, we propose a framework to leverage the multimodal prompt to guide the Multimodal Multilingual neural Machine Translation (m3P), …

abstract arxiv bridge context cs.ai cs.cl difference gap languages multilingual multimodal multiple prompt quality space text translation type visual