Learning representations by back-propagating errors https://www.nature.com/articles/323533a0.pdf David Rumelhart, Geoffrey Hinton, Ronald Williams Graph embedding that predates Word embedding?