We started off the chapter by understanding CNNs. We learned about the different layers of a CNN, such as convolution and pooling; where the important features from the image will be extracted and are fed to the fully collected layer; and where the extracted feature will be classified. We also visualized the features extracted from the convolutional layer using TensorFlow by classifying handwritten digits.

Later, we learned about several architectures of CNN, including LeNet, AlexNet, VGGNet, and GoogleNet. At the end of the chapter, we studied Capsule networks, which overcome the shortcomings of a convolutional network. We learned that Capsule networks use a dynamic routing algorithm for classifying the image.

In the next chapter, we will study the various algorithms used for learning text representations.