UNIT 4 QUIZ

Consider the two neural networks (NNs) shown in Figures 1 and 2, with 𝑅𝑒𝐿𝑈 activation (𝑅𝑒𝐿𝑈(𝑧) = max{0, 𝑧}, ∀𝑧 ∈ ℝ). ℝ denotes the set of real numbers. The connections and their corresponding weights are shown in the Figures. The biases at every neuron are set to 0. For what values of 𝑝, 𝑞, 𝑟 in Figure 2 are the two NNs equivalent, when 𝑥1 ,𝑥2 , 𝑥3 are positive? (GATE 2024)

What are the updated values of weights OF W13, W23 after first back propagation cycle, where learning rate is 1 where target value of y is 0.5 using MLP (use Sigmodal Activation Function, round of 4 decimal points).

What do you mean by backpropagation ______________ in neural networks?

• Forward pass of data

• Training process for adjusting weights

• Activation of output layer

• None of the above

In a feedforward neural network, information flows ___________ direction.

There are M neurons in k-1 hidden layer and N neurons in K-2 Hidden layer with a biased input to the K-1 layer then how many number of connected paths to the K-1 th layer?

In MLP K-2th layer is_______________ , K-1th layer is__________ and Kth layer is___________ layers.

Which of the following is the drawback of mlp

• Saturation

• Noise of the dataset

• Cannot work with huge data

Back propagation algorithm can use the following techniques

• Stochastic gradient descent

• Mini batch gradient descent

• Batch gradient descent

• Weight updation rule

The _______ activation function used in perceptron algorithm

Which of the following are the activation functions

• Linear

• Sigmoidal

• Softmax

• Hyperbolic tangent