Deep Learning Convolutional Neural Networks 3

Question 1

What is data augmentation used for when training CNNs?

Accepted Answer

Artificially expanding training data with transformations to improve generalization

Answer

Increasing inference speed

Answer

Reducing the number of model parameters

Answer

Speeding up gradient computation

Question 2

What does fine-tuning mean in the context of transfer learning with pretrained CNNs?

Accepted Answer

Continuing to train some or all pretrained layers on the new task with a low learning rate

Answer

Training only the classification head from scratch

Answer

Freezing all pretrained weights permanently

Answer

Replacing all pretrained weights with random initialization

Question 3

What is feature map in the context of a CNN?

Accepted Answer

The output activation volume produced by applying a set of filters to the previous layer

Answer

The input image itself

Answer

The weight matrix of a convolutional layer

Answer

The final class probability vector

Question 4

Which CNN architecture was the first to demonstrate the power of deep learning on ImageNet, winning in 2012?

Accepted Answer

AlexNet

Answer

LeNet-5

Answer

VGGNet

Answer

ResNet

Question 5

What is the vanishing gradient problem's equivalent for very deep CNNs without residual connections?

Accepted Answer

Gradient signal becoming too weak to update early layer weights effectively

Answer

Activation collapse

Answer

Feature maps growing unboundedly

Answer

Filter weights converging to zero in the first layer

Question 6

What does an object detection CNN like YOLO do differently from a standard image classification CNN?

Accepted Answer

It predicts both class labels and bounding box coordinates simultaneously

Answer

It uses recurrent layers instead of convolutional layers

Answer

It only processes images in grayscale

Answer

It does not use any convolutional layers