A quantitative analysis of semantic information in deep representations of text and images

Open in new window