Description

The network should implement the XOR function. Perform one epoch of backpropagation as introduced in the lecture on multilayer perceptrons.
Notes:
• The activation function f for a perceptron is the sigmoid function:
1
𝑓(𝑥)=1+𝑒!”
• The thresholds are not shown in the network. The threshold nodes are set to -1.
• Use the following initial parameter values:
𝜃#$($) =2 𝜃#’($) =−1 𝜃#$(’) =−2
𝜃$$($) =6 𝜃$’($) =8 𝜃$$(’) =6
𝜃’$($) =−6 𝜃’’($) =−8 𝜃’$(’) =−6
• The learning rate is set to η = 0.7
i. Compute the activations of the hidden and output neurons.
ii. Compute the error of the network.
iii. Backpropagate the error to determine ∆θij for all weights θij and updates the weight
θij.
2. What is the difference between “model bias” and “model variance”?
i. Why is a high bias, low variance classifier undesirable? ii. Why is a low bias, high variance classifier (usually) undesirable?
3. Describe how validation set, and cross-validation can help reduce overfitting?