Artificial Intelligence Neural Networks and Deep Learning 2

Question 1

What is batch normalization designed to do in deep neural networks?

Accepted Answer

Normalize activations within a layer to stabilize and accelerate training

Answer

Batch normalization normalizes layer inputs over a mini-batch, reducing internal covariate shift and allowing higher learning rates.

Question 2

Which architecture introduced residual (skip) connections to train very deep networks?

Accepted Answer

ResNet

Answer

ResNet introduced skip connections that add the input of a block to its output, enabling gradients to flow more easily through hundreds of layers.

Question 3

What distinguishes a recurrent neural network (RNN) from a feedforward network?

Accepted Answer

RNNs have feedback connections that allow information to persist across time steps

Answer

RNNs have recurrent connections that pass hidden states from one time step to the next, enabling them to model sequential and temporal data.

Question 4

What is the primary advantage of using Long Short-Term Memory (LSTM) over a simple RNN?

Accepted Answer

LSTMs can capture long-range dependencies by controlling information flow with gates

Answer

LSTM gates (input, forget, output) regulate what information is stored, forgotten, or passed on, mitigating the vanishing gradient problem in long sequences.

Question 5

In a softmax output layer for multi-class classification, what do the output values represent?

Accepted Answer

A probability distribution over all classes summing to 1

Answer

Softmax converts raw logits into a probability distribution where all values are positive and sum to 1, one per class.

Artificial Intelligence Practice Test

Artificial Intelligence Practice Test

Artificial Intelligence Neural Networks and Deep Learning 2