Deep learning NATO document labels: A preliminary investigation
Marc Richter, Michael Street, Peter Lenk · 2018
This paper provides an introductory, mostly nontechnical exposition of recent Deep Learning concepts, as relevant to NATO information processing requirements. A baseline case of learning confidentiality classification rules for NATO documents is described, originally performed by applying classical machine learning methods. In the course of the investigation described in this paper, the classical approaches have been complemented by Deep Learning experiments. These experiments shed some light on the potential advantages of using Deep Learning approaches for NATO document processing and beyond.