1 paper
Haihan Li, Haihao Li, Zhenfei Xu +2
Many multimodal tasks depend on how visual elements are ordered and composed, not only on recognizing them in isolation. Internet memes are a compact case of this problem: their pu…