Discovering Meaningful Units with Visually Grounded Semantics from Image Captions

Open in new window