Pictures and words placed in one shared space. Train the space, then click a word or a picture to see its nearest neighbours.