This chapter focuses on a gentle introduction to the domain of generative deep learning, which has been one of the core ideas at the forefront of true artificial intelligence (AI). We will be focusing on how Convolutional Neural Networks (CNNs) think or visualize patterns in images by leveraging transfer learning. They can generate image patterns never seen before depicting the way these convnets think or even dream! First released by Google in 2015, DeepDream became a viral sensation due to the interesting patterns deep networks started to generate from images. We will be covering the following major topics in this chapter:

• Motivation—psychological pareidolia
• Algorithmic pareidolia in computer vision
• Understanding what CNNs have learned by visualizing internal layers of CNN
• DeepDream algorithm and how to create your own dream

Just like the previous chapters, we will use a combination of conceptual knowledge powered with intuitive hands-on examples. The code for this chapter is available for quick reference in the Chapter 9 folder in the GitHub repository at https://github.com/dipanjanS/hands-on-transfer-learning-with-python which you can refer to as needed to follow along with the chapter.