> In its feed-forward architecture, layers of perceptrons—also referred to as neurons—first perform weighted averaging of their inputs, followed by nonlineariti

> In its feed-forward architecture, layers of perceptrons—also referred to as neurons—first perform weighted averaging of their inputs, followed by nonlinearities such as a sigmoid or rectified-linear curves. I have a really hard time keeping reading after this passage.