What goes into a word: generating image descriptions with top-down spatial knowledge.

Mehdi Ghanimifard, Simon Dobnik

Browse the full INLG paper archive.