Fig. 1From: Concept recognition as a machine translation problemExample of the full translation pipeline. Each step is seen as a translation problem. The input is text and the final output is the ontology class identifiers for each detected text mentionBack to article page