Search Results for author: Yutaro Omote

Found 3 papers, 0 papers with code

Supervised Visual Attention for Multimodal Neural Machine Translation

no code implementations COLING 2020 Tetsuro Nishihara, Akihiro Tamura, Takashi Ninomiya, Yutaro Omote, Hideki Nakayama

This paper proposed a supervised visual attention mechanism for multimodal neural machine translation (MNMT), trained with constraints based on manual alignments between words in a sentence and their corresponding regions of an image.

Machine Translation Sentence +1

Dependency-Based Relative Positional Encoding for Transformer NMT

no code implementations RANLP 2019 Yutaro Omote, Akihiro Tamura, Takashi Ninomiya

This paper proposes a new Transformer neural machine translation model that incorporates syntactic distances between two source words into the relative position representations of the self-attention mechanism.

Machine Translation NMT +2

Cannot find the paper you are looking for? You can Submit a new open access paper.