Deep Visual-Semantic Alignments for Generating Image Descriptions #1 Post by nh » Tue, Jan 06, 2015, 3:01 PM UTC Deep Visual-Semantic Alignments for Generating Image Descriptionscs.stanford.edu