The Role of Deep Learning in Artificial Intelligence: Separating Fact from Fiction.

Researchers at Bar-Ilan University have shown that efficient learning on an artificial shallow architecture can achieve the same classification success rates as deep learning architectures, but with less computational complexity. This raises the question of whether deep learning is necessary for artificial intelligence. The efficient realization of shallow architectures requires a shift in the properties of advanced GPU technology and future dedicated hardware developments. Brain-inspired shallow learning has advanced computational capability with reduced complexity and energy consumption.