Man and machines against crime: An approach based on visual learning

摘要

Law enforcement agencies in Brazil communicate with each other using documents formalized in a language predominantly written in pure text. However, postmodern criminal organizations play complex roles, making it difficult to describe their actions using only text. Visual memory is relevant to learning and thus should be applied. Learning about crime requires time, effort, and dedication, but visual resources can minimize these obstacles. Because the human period of concentration is short, long texts make it difficult to hold people’s attention. The tactic of distributed practice is recommended as a strategy for learning, but with the use of features such as diagrams. In addition, learning about criminal activities using multimedia formats is effective. The teaching of content production skills for visual learning is a necessity at police academies; however, the acquisition of these abilities is slow. The analysis of legacy collections of massive texts by computer, through natural language processing (NLP) is suggested as an alternative. The process of extracting entities and relationships is illustrated, and the main steps involved are explained. The skills for generating visual content and NLP should be strengthened, since they complement each other in the fight against criminal organizations.