CNN Visualizing CNN Feature Maps 3

Question 1

What is the purpose of 'maximally activating patches' in CNN visualization?

Accepted Answer

To find input image regions that produce the highest activation for a specific filter

Answer

Maximally activating patches are the input image crops that cause a particular filter or neuron to fire most strongly, revealing what it has learned to detect.

Question 2

Which method synthesizes an input image from scratch to maximize a neuron's activation, rather than searching the training set?

Accepted Answer

Activation maximization (deep dream / feature visualization)

Answer

Activation maximization iteratively updates a random input image via gradient ascent to maximize a target neuron's response, revealing the ideal stimulus.

Question 3

Why are deeper convolutional layers in a CNN said to have larger 'receptive fields'?

Accepted Answer

Each neuron integrates information from an increasingly larger region of the original input

Answer

As feature maps are stacked, each subsequent neuron indirectly covers a larger spatial region of the original input, giving it a larger effective receptive field.

Question 4

When visualizing CNN feature maps for debugging, dead filters (all-zero activations) often indicate what problem?

Accepted Answer

Dying ReLU problem where filters never activate

Answer

All-zero feature maps often result from the dying ReLU phenomenon, where neurons get stuck outputting zero for all inputs after large negative pre-activations.

Question 5

In PyTorch, what is a common approach to capture feature maps from an intermediate layer during a forward pass?

Accepted Answer

Registering a forward hook on the target layer

Answer

PyTorch's register_forward_hook() attaches a callback to a layer that captures its output tensor whenever data passes through it.

CNN - Convolutional Neural Networks Practice Test

CNN - Convolutional Neural Networks Practice Test

CNN Visualizing CNN Feature Maps 3