Even the smartest visual recognition system requires even a small human intervention in order to be effective. But not the system developed by the scientists at Brigham Young University. The new algorithm is able to decide on its own whether what it sees in an image is important or not, and each time it sees a new object, no adjustment is needed to recognize it. In addition, it is extremely accurate, while it can distinguish even the smallest differences between similar ones objects – different already fish for example.
In essence, the way it works resembles the way the human mind works: He's learning on his own. So we do not need to teach him the differences between two objects. After seeing them once, they are then able to distinguish them.
The question is where one could be exploited smart algorithm. The answer is “everywhere”, it is enough of course to prove in practice that it does not Mistakes. Το πρώτο πράγμα που μας έρχεται στο μυαλό είναι ο στρατός. Ειδικά ο αμερικανικός στρατός uses τέτοια συστήματα εδώ και πολλά χρόνια προκειμένου τα έξυπνα συστήματα άμυνας να μπορούν να αναγνωρίζουν τους εχθρικούς στόχους. Στα συστήματα αυτά, πάντα υπάρχει ένας άνθρωπος που παίρνει την τελική απόφαση. Όχι πια. Τώρα, ο υπολογιστής θα αποφασίζει μόνος του και θα εκτοξεύει πυραύλους ή σφαίρες στον (υποτιθέμενο) εχθρό.