Deep Learning Convolutional Neural Networks 2

Question 1

What is the effect of adding zero-padding to the input of a convolutional layer?

Accepted Answer

Preserves spatial dimensions by adding zeros around the input border

Answer

Increases the number of filters

Answer

Removes the need for pooling

Answer

Increases the stride value

Question 2

In transfer learning with CNNs, what are the early layers of a pretrained model typically found to detect?

Accepted Answer

Low-level features such as edges and textures

Answer

High-level semantic concepts like faces

Answer

Only domain-specific patterns

Answer

Class probability distributions

Question 3

What is a depthwise separable convolution, as used in MobileNet?

Accepted Answer

A factored convolution splitting spatial and channel mixing into separate steps to reduce computation

Answer

A convolution applied only to depth channels

Answer

A 1×1 convolution reducing channel dimensions

Answer

A strided convolution replacing pooling layers

Question 4

What problem does the Inception module in GoogLeNet address?

Accepted Answer

Choosing filter size by applying multiple filter sizes in parallel and concatenating outputs

Answer

Vanishing gradients in shallow networks

Answer

Eliminating fully connected layers entirely

Answer

Replacing activation functions with normalization

Question 5

What is the main purpose of 1×1 convolutions in deep CNN architectures?

Accepted Answer

Perform channel-wise linear combinations to reduce or expand depth cheaply

Answer

Increase spatial resolution

Answer

Apply spatial pooling

Answer

Add non-linearity without changing spatial dimensions

Question 6

Which layer type is typically used at the end of a CNN before the final classification head?

Accepted Answer

Global average pooling or flatten layer

Answer

Convolutional layer

Answer

Batch normalization layer

Answer

Depthwise convolution layer