A given Deep Neural Network (DNN) applies inverted dropout. The DNN comprises of...
✅ The verified answer to this question is available below. Our community-reviewed solutions help you understand the material better.
A given Deep Neural Network (DNN) applies inverted dropout. The DNN comprises of 3 sensing neurons (N1 - N3), 4 processing neurons (N4 - N7) in hidden layer 1, 3 processing neurons (N8 - N10) in hidden layer 2, and 1 processing neuron (N11) in the output layer. All processing neurons use the sigmoid activation function. During a feedforward phase, N6 is expected to have a net function (z) output of 0, the actual activation is however scaled to 0.625 due to the implementation of inverted dropout. How many neurons were dropped during this feedforward phase?
Want instant access to all verified answers on stemlearn.sun.ac.za?
Get Unlimited Answers To Exam Questions - Install Crowdly Extension Now!