A Distinguishing supervised learning from reinforcement learning in a feedforward model { 1, 1} and t = 1,, T, are projected onto a hiddenlayer h