Book Image

Python Deep Learning - Second Edition

By : Ivan Vasilev, Daniel Slater, Gianmario Spacagna, Peter Roelants, Valentino Zocca
Book Image

Python Deep Learning - Second Edition

By: Ivan Vasilev, Daniel Slater, Gianmario Spacagna, Peter Roelants, Valentino Zocca

Overview of this book

With the surge in artificial intelligence in applications catering to both business and consumer needs, deep learning is more important than ever for meeting current and future market demands. With this book, you’ll explore deep learning, and learn how to put machine learning to use in your projects. This second edition of Python Deep Learning will get you up to speed with deep learning, deep neural networks, and how to train them with high-performance algorithms and popular Python frameworks. You’ll uncover different neural network architectures, such as convolutional networks, recurrent neural networks, long short-term memory (LSTM) networks, and capsule networks. You’ll also learn how to solve problems in the fields of computer vision, natural language processing (NLP), and speech recognition. You'll study generative model approaches such as variational autoencoders and Generative Adversarial Networks (GANs) to generate images. As you delve into newly evolved areas of reinforcement learning, you’ll gain an understanding of state-of-the-art algorithms that are the main components behind popular games Go, Atari, and Dota. By the end of the book, you will be well-versed with the theory of deep learning along with its real-world applications.
Table of Contents (12 chapters)

Driving policy with ChauffeurNet

In this section, we'll discuss a recent paper called ChauffeurNet: Learning to Drive by Imitating the Best and Synthesizing the Worst (https://arxiv.org/abs/1812.03079). It was released in December 2018 by Waymo, one of the leaders in the AV space. The following are some of the properties of the ChauffeurNet model:

  • It is a combination of two interconnected networks. The first is a CNN called FeatureNet, which extracts features from the environment. These features are fed as inputs to a second, recurrent network called AgentRNN, which them to determine the driving policy.
  • It uses imitation supervised learning similarly to the algorithms we described in the Imitation driving policy section. The training set is generated based on records of real-world driving episodes. ChauffeurNet can handle complex driving situations such as lane changes,...