1 paper
Chen Wei, Chenxi Liu, Siyuan Qiao +3
We demonstrate text as a strong cross-modal interface. Rather than relying on deep embeddings to connect image and language as the interface representation, our approach represents…