Advanced Deep Learning
Dattaraj Jagdish Rao · 2019
This chapter discusses some advanced concepts in deep learning. It focuses on various models that extract complex visual patterns. One of the major improvements that Deep Learning (DL) introduces is new types of layers that help to build special types of models. These models work well on specific types of unstructured data like images or text. The chapter looks at how DL provides specialized layers and network architectures to help in image analysis. These types of networks are specialized neural networks known as Convolutional Neural Networks (CNNs). CNNs have been universally accepted as the best models available today for analyzing images and extracting knowledge from them. The chapter also considers useful techniques that data scientists use regularly to solve problems. Using augmentation techniques, more data can be created that can be used to train the models.