Apply the backpropagation algorithm to find the new weights of a feedforward neural network after applying the input-output pattern ({1,1},0), where {1,1} are two inputs, and 0 is the output. Assume the network has a single hidden layer with 2 neurons and all weights (including bias in hidden and output layers) are initially zero. The learning rate is 0.1 and all neurons use the sigmoid activation function. [20 MARKS]