Chapter 1, Recognizing traffic signs using Convnets, shows how to extract the proper features from images with all the necessary preprocessing steps. For our convolutional neural network, we will use simple shapes generated with matplotlib. For our image preprocessing exercises, we will use the Yale Face Database.
Chapter 2, Annotating Images with Object Detection API, details a the building of a real-time object detection application that can annotate images, videos, and webcam captures using TensorFlow's new object detection API (with its selection of pretrained convolutional networks, the so-called TensorFlow detection model zoo) and OpenCV.
Chapter 3, Caption Generation for Images, enables readers to learn caption generation with or without pretrained models.
Chapter 4, Building GANs for Conditional Image Creation, guides you step by step through building a selective GAN to reproduce new images of the favored kind. The used datasets that GANs will reproduce will be of handwritten characters (both numbers and letters in Chars74K).
Chapter 5, Stock Price Prediction with LSTM, explores how to predict the future of a mono-dimensional signal, a stock price. Given its past, we will learn how to forecast its future with an LSTM architecture, and how we can make our prediction's more and more accurate.
Chapter 6, Create and Train Machine Translation Systems, shows how to create and train a bleeding-edge machine translation system with TensorFlow.
Chapter 7, Train and Set up a Chatbot, Able to Discuss Like a Human, tells you how to build an intelligent chatbot from scratch and how to discuss with it.
Chapter 8, Detecting Duplicate Quora Questions, discusses methods that can be used to detect duplicate questions using the Quora dataset. Of course, these methods can be used for other similar datasets.
Chapter 9, Building a TensorFlow Recommender System, covers large-scale applications with practical examples. We'll learn how to implement cloud GPU computing capabilities on AWS with very clear instructions. We'll also utilize H2O's wonderful API for deep networks on a large scale.
Chapter 10, Video Games by Reinforcement Learning, details a project where you build an AI capable of playing Lunar Lander by itself. The project revolves around the existing OpenAI Gym project and integrates it using TensorFlow. OpenAI Gym is a project that provides different gaming environments to explore how to use AI agents that can be powered by, among other algorithms, TensorFlow neural models.