KU Leuven researchers have shown machines can learn to respond to unfamiliar objects like a human would.
Current state-of-the-art image-recognition technologies are designed to recognize a fixed set of objects using deep neural networks. The KU Leuven researchers found deep neural networks are good at making objective decisions, as well as developing human-level sensitivities to object shapes. For example, they say their research shows a machine can learn to describe what a new shape reminds it of.
"This means we're on the right track in developing machines with a visual system and vocabulary as flexible and versatile as ours," says KU Leuven researcher Jonas Kubilius.
However, even if machines will some day be equipped with visual systems as powerful as that of humans, Kubilius says self-driving cars and other types of technology would still make mistakes. Nevertheless, he says even in those instances when a self-driving car misidentifies an object, its decision would be at least as reasonable as that of a human driver.
From KU Leuven
View Full Article
Abstracts Copyright © 2016 Information Inc., Bethesda, Maryland, USA
No entries found