A Consider a fully connected autoencoder (each hidden node is connected to all inputs and all outputs) with 2 dimensional binary input and one hidden layer with tanh activation function. At iteration t, the weights are shown in the following autoencoder architecture along with the input vector (x1=1, x2=0). All bias values are 0.

w3=1, w4=0, w1=0, w2=1

What type of activation function is needed at the output node? What type of loss function will you use for training this autoencoder?

What will be the value of loss function at iteration t?