Neural Network Convolutional Neural Networks

Question 1

What is a convolutional layer and what operation does it perform?

Accepted Answer

A layer that slides learned filters across the input to produce feature maps

Answer

A layer that fully connects all neurons

Answer

A layer that reduces dimensions via pooling

Answer

A layer that applies dropout to feature maps

Question 2

What does max pooling accomplish in a CNN?

Accepted Answer

It downsamples feature maps by taking the maximum value in each pooling window

Answer

It learns to detect edges in the input

Answer

It normalizes feature map values between 0 and 1

Answer

It applies a learned transformation to reduce dimensions

Question 3

What is the receptive field of a neuron in a CNN?

Accepted Answer

The region of the input image that influences that neuron's activation

Answer

The set of neurons it is connected to in the same layer

Answer

The size of the filter applied to the neuron

Answer

The number of feature maps the neuron belongs to

Question 4

What is the purpose of padding in convolutional layers?

Accepted Answer

To preserve spatial dimensions and allow filters to be applied at border regions

Answer

To add extra training examples to the dataset

Answer

To prevent overfitting by adding noise

Answer

To increase the number of learned filters

Question 5

What does the stride parameter control in a convolutional layer?

Accepted Answer

The step size with which the filter slides across the input

Answer

The number of filters in the layer

Answer

The depth of the feature maps produced

Answer

The learning rate for that layer's weights

Question 6

What is a 1×1 convolution and why is it useful?

Accepted Answer

A convolution that mixes channel information without affecting spatial dimensions

Answer

A convolution that applies no transformation

Answer

A convolution used only for the final classification layer

Answer

A convolution that doubles the spatial dimensions