Machine learningLesson 2 of 3 in the course Neural networks

Convolutional Neural Networks: Design, Training, History, and Applications

5 flashcards · 5 quiz questions · 5 match pairs · 5 fill-in sentences · English

Save course
Lesson progress
0 / 5

Your progress is kept in this browser only.Sign up to save it

Summary

This lesson introduces convolutional neural networks (CNNs), explains why they are effective for structured data such as images, and relates their design to visual processing.

A convolutional neural network (CNN) is a feedforward neural network that learns useful features by optimizing filters, also called kernels. A filter moves across an input, such as an image, and responds to patterns it has learned to detect. The resulting responses form a feature map.

All flashcards

What is a CNN?
A convolutional neural network is a feedforward network that learns filters to identify patterns in inputs such as images, audio, or text.
What is a filter or kernel?
A learnable set of weights applied across an input region to produce responses to a feature; the responses form a feature map.
What is local connectivity?
The design in which a neuron receives input from only a limited spatial region of the previous layer, called its receptive field.
Why are weights shared in a convolutional layer?
Sharing reuses the same filter at different positions, greatly reducing the number of parameters and allowing a feature to be detected at multiple locations.
How does a CNN build a hierarchy of features?
Stacked layers combine local patterns into responses based on increasingly broad regions, allowing later layers to represent more complex structures.

Quiz questions

  1. What does optimizing a CNN's filters enable the network to learn?Answer: Local patterns represented by filter responses.
  2. What is a receptive field in a convolutional layer?Answer: The region of the previous layer that can affect a neuron's output.
  3. Why does sharing weights reduce the parameter count of a convolutional layer?Answer: The same filter parameters are reused at multiple locations.
  4. How does translation equivariance differ from translation invariance?Answer: Equivariance shifts the response correspondingly; invariance leaves the response unchanged.
  5. What does a transposed convolution do?Answer: It applies the transpose of a convolution's associated linear transformation.

Based on the Wikipedia article Convolutional neural network, licensed CC BY-SA 4.0. Adapted into a summary, flashcards and quizzes.

More in Machine learning

All Machine learning lessons
Machine learningAutoencoder Architectures, Training, Regularization, and Applications5 flashcards · 5 quiz questions · 4 gamesPlay free →Machine learningSupport Vector Machines: Margins, Kernels, Optimization, and Extensions5 flashcards · 5 quiz questions · 4 gamesPlay free →Machine learningNeural-Network Activation Functions: Properties and Families5 flashcards · 5 quiz questions · 4 gamesPlay free →Machine learningHierarchical Clustering: Linkage, Dendrograms, and Scalable Methods5 flashcards · 5 quiz questions · 4 gamesPlay free →Machine learningNeural Networks: Architectures, Training, History, and Applications5 flashcards · 5 quiz questions · 4 gamesPlay free →Machine learningConfusion Matrices: Binary Outcomes, Imbalance, and Extensions5 flashcards · 5 quiz questions · 4 gamesPlay free →