A neural network implementation with Feed Forward and Back-Propagation to be trained to draw a curve. The Neural Network contains 2 hidden layer of 6 and 7 neurons respectivly, one input(a value along the curve from 0 to one ), and 2 outputs representing the curve coordiates
We will use the logistic activation function
Calculate the jacobian of cost function c (average function ) with respect to weights and biases in each layer.