What is the purpose of gradient descent in neural networks: (a) To find the value of a parameter that reduces the difference between output and the target (b) To figure out how many nodes should be in a neural network (c) To move the value of a parameter to zero. (d) To find the steepest slope of a curve