VICTR: Visual Information Captured Text Representation for Text-to-Vision Multimodal Tasks.

Soyeon Caren Han, Siqu Long, Siwen Luo, Kunze Wang, Josiah Poon

Browse the full COLING paper archive.