1 paper · 1 filter
Yiming Huang, Aozhe Jia, Xiaodan Zhang +1
Attention-based transformer models have achieved remarkable progress in multi-modal tasks, such as visual question answering. The explainability of attention-based methods has rece…