Vision and Goals
Whole-Brain Models for Biologically Inspired Learning
From the essay
Despite advances in performance of deep learning artificial neural networks on difficult control problems, most algorithms use abstract neurons and rely on massive amounts of data and energy. Neurobiological learning mechanisms differ from methods used to train artificial neural networks. The brain only utilizes a small fraction of their energy, and yet generates widely generalizable intelligence. Bridging performance between these two types of systems is a major challenge that requires a paradigm shift in understanding how neural systems learn and adapt.
Adapted from PDF pages 1.
Spacing defects in the PDF text layer have been repaired in the introduction; some full-text searches may still be affected.