The functional form for a single perceptron is given by y = sign( + b), where x is the data point and y is the predicted label. Suppose

your data is 5-dimensional (i.e., x = (x1, x2, x3, x4, x5)) and real-valued. Find a simple 2-layer network of perceptrons that implements the Decreasing-Order function, i.e., it return +1 if x1 > x2 > x3 > x4 > x5 and -1 otherwise.

Your network should have 2 layers: the input nodes, feeding into 4 hidden perceptrons, which in turn feed into 1 output perceptron. Clearly indicate all the weights and

biases of all the 5 perceptrons in your network.