what are the steps for using a gradient descent algorithm? calculate error between the actual value and the predicted value reiterate until you find the best weights of network pass an input through the network and get values from output layer initialize random weight and bias go to each neurons which contributes to the error and change its respective values to reduce the error