DL-Activation Functions

What is the primary purpose of an activation function in a neural network?

• To initialize weights

• To introduce non-linearity

• To adjust learning rate

• To increase the number of neurons

Which activation function is commonly used in output layers for binary classification?

• ReLU

• Tanh

• Sigmoid

• Softmax

What is the range of the sigmoid activation function?

• (-∞, ∞)

• (-1, 1)

• (0, 1)

• (0, ∞)

Which activation function is prone to the vanishing gradient problem?

• ReLU

• Leaky ReLU

• Softmax

• Sigmoid

What is the output range of the Tanh activation function?

• (0, 1)

• (-∞, ∞)

• (-1, 1)

• (0, ∞)

Which activation function is commonly used in hidden layers of deep networks?

• ReLU

• Sigmoid

• Tanh

• Step function

What does ReLU stand for?

• Residual Learning Unit

• Regularized Linear Update

• Recursive Linear Unit

• Rectified Linear Unit

Which of the following is an advantage of the ReLU activation function?

• It avoids vanishing gradient problem

• It is linear in nature

• It has a range between -1 and 1

• It works only for positive inputs

What is the main advantage of the Tanh activation function over Sigmoid?

• Tanh is always positive

• Tanh centers the data around 0

• Tanh is computationally faster

• Tanh can classify multiple categories

The step function is not used in deep learning because:

• It is non-linear

• It has a range from -1 to 1

• It is too complex to compute

• It does not allow backpropagation