NN UBS

Find the ideal choice of fitness function for output layer for predicting multiple classes

• Softmax

• Relu

• Sigmoid

• Tanh

How many outcomes can be produced by artificial neural networks?

• 1

• 2

• 3

• Many

What are the various layers present in ANN?

• Input Layer

• Hidden Layer

• Output Layer

• All Mentioned Above

After receiving the outcome, it is compared with the original one, and the weights are updated is referred to as ____________?

• Signal Propagation

• Forward Propagation

• Backward Propagation

• Channel Propagation

A 4-input neuron has weights 1, 2, 3 and 4. The transfer function is linear. The inputs are 4, 10, 5 and 20 respectively. The output will be:

• 238

• 76

• 119

• 178

Will your answer change if the activation function is Sigmoid?

• Yes

• No

What are the steps for using a gradient descent algorithm?

1. 1. Calculate error between the actual value and the predicted value

2. 2. Reiterate until you find the best weights of network

3. 3. Pass an input through the network and get values from output layer

4. 4. Initialize random weight and bias

5. 5. Go to each neurons which contributes to the error and change its respective values to reduce the error

• 1, 2, 3, 4, 5

• 4, 3, 1, 5, 2

• 5, 4, 3, 2, 1

• 3, 2, 1, 5, 4

1. In training a neural network, you notice that the loss does not decrease in the few starting epochs.

2. 1. The learning is rate is low

3. 2. Regularization parameter is high

4. 3. Stuck at local minima

What according to you are the probable reasons?

• 1 ,2

• 2,3

• 1. 1,3

• Any of these

First Order Gradient descent would not work correctly (i.e. may get stuck) in which of the following graphs?

• A

• B

• C

• D

Suppose we have one hidden layer neural network as shown above. The hidden layer in this network works as a dimensionality reductor. Now instead of using this hidden layer, we replace it with a dimensionality reduction technique such as PCA.

Would the network that uses a dimensionality reduction technique always give same output as network with hidden layer?

• Yes

• No

Answer explanation

Because PCA works on correlated features, whereas hidden layers work on predictive capacity of features.