O'Reilly logo

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

Deep Reinforcement Learning and GANs: Advanced Topics in Deep Learning

Video Description

6+ Hours of Video Instruction

An intuitive introduction to the latest developments in Deep Learning.


Deep Reinforcement Learning and GANs LiveLessons is an introduction to two of the most exciting topics in Deep Learning today. Generative Adversarial Networks cast two Deep Learning networks against each other in a “forger-detective” relationship, enabling the fabrication of stunning, photorealistic images with flexible, user-specifiable elements. Deep Reinforcement Learning has produced equally surprising advances, including the bulk of the most widely-publicized “artificial intelligence” breakthroughs. Deep RL involves training an “agent” to become adept in given “environments,” enabling algorithms to meet or surpass human-level performance on a diverse range of complex challenges, including Atari video games, the board game Go, and subtle hand-manipulation tasks. Throughout these lessons, essential theory is brought to life with intuitive explanations and interactive, hands-on Jupyter notebook demos. Examples feature Python and Keras, the high-level API for TensorFlow, the most popular Deep Learning library.

The companion materials for this LiveLesson can be found at https://github.com/the-deep-learners/TensorFlow-LiveLessons/.

About the Instructor

Jon Krohn is Chief Data Scientist at untapt, a machine-learning startup in New York. He presents an acclaimed series of tutorials on artificial neural networks, includingDeep Learning with TensorFlow LiveLessonsandDeep Learning for Natural Language Processing LiveLessons. He also teaches his curriculum in-classroom at the NYC Data Science Academy. Jon holds a doctorate in neuroscience from Oxford University and has been publishing on machine learning in leading academic journals since 2010.

Skill Level

  • Intermediate
Learn How To
  • Understand the high-level theory and key language around deep reinforcement learning and generative adversarial networks
  • Architect GANs that create convincing images in the style of human-drawn illustrations
  • Build deep RL agents that become adept at performing in a wide variety of environments, such as those provided by OpenAI Gym
  • Run automated experiments for optimizing deep reinforcement learning agent parameters, such as its artificial-neural-network configuration
  • Appreciate what the current limitations of “artificial intelligence” are and how they may be overcome in the near future
Who Should Take This Course
  • Perfectly suited to software engineers, data scientists, analysts, and statisticians who want to further their understanding of deep learning.
  • Code examples are provided in Python, so familiarity with it or another object-oriented programming language would be helpful.
  • Previous experience with statistics or machine learning is not necessary.
Course Requirements
  • The author’s earlier Deep Learning with TensorFlow LiveLessons, or equivalent foundational Deep Learning knowledge, are a prerequisite.
Lesson Descriptions

Lesson 1: The Foundations of Artificial Intelligence
This lesson starts off by examining what the term “AI” means and how it relates to deep learning‚Äîparticularly the advanced Deep Learning topics we’re covering in these LiveLessons. It continues by discussing cutting-edge applications of generative adversarial networks and deep reinforcement learning algorithms that have recently revolutionized the field of machine learning. We then quickly review how to run the code in these LiveLessons on your own machine as well as the foundational deep learning theory that is essential for building these advanced-topics specializations upon.

Lesson 2: Generative Adversarial Networks (GANs)
Lesson 2 begins by covering the high-level theory of what GANs are and how they are able to generate realistic-looking images. Next is the “Quick, Draw!” game, which is used as the source of hundreds of thousands of hand-drawn images from a single class‚Äîsuch as apples, rhinoceroses, or rainbows‚Äîfor a GAN to learn to imitate. The bulk of the lesson is spent developing the intricate code for the three primary components of a GAN: the discriminator network, the generator network, and the adversarial network that pits them against each other.

Lesson 3: Deep Q-Learning Networks (DQNs)
Lesson 3 is the first of three lessons that explore deep reinforcement learning algorithms. It introduces a simple game called the Cartpole Game that is used throughout the rest of the lessons to train your deep reinforcement learning algorithms. Next, the lesson delves a bit into the essential theory of deep reinforcement learning as well as DQNs, a popular type of deep reinforcement learning agent. With that theory under your belt, you’ll be able to understand at an intuitive level the code that you subsequently develop when you define your own DQN Agent and have it interact with The Cartpole Game within a handy library called OpenAI Gym.

Lesson 4: OpenAI Lab
In the previous lesson, a deep Q-learning network was used to master the Cartpole Game. This lesson builds upon those deep reinforcement learning foundations by using the OpenAI Lab both to visualize your DQN agent’s performance in real-time and to straightforwardly modify its hyperparameters. You learn how to automate the search through hyperparameters to optimize your agent’s performance and to gauge your agent’s overall fitness.

Lesson 5: Advanced Deep Reinforcement Learning Agents
The previous two lessons covered deep reinforcement learning largely through the lens of the deep Q-Learning network. In this lesson, your arsenal of deep reinforcement learning algorithms expands. The lesson begins with coverage of policy gradient algorithms, and then these are combined with Q-learning to discover the so-called actor-critic algorithms. The lesson closes out by covering why deep learning is reshaping software in general and by returning to the discussion of artificial intelligence—specifically addressing the limitations of contemporary deep learning approaches.

About Pearson Video Training

Pearson publishes expert-led video tutorials covering a wide selection of technology topics designed to teach you the skills you need to succeed. These professional and personal technology videos feature world-leading author instructors published by your trusted technology brands: Addison-Wesley, Cisco Press, Pearson IT Certification, Prentice Hall, Sams, and Que Topics include: IT Certification, Network Security, Cisco Technology, Programming, Web Development, Mobile Development, and more. Learn more about Pearson Video training at http://www.informit.com/video.

Table of Contents

  1. Introduction
    1. Deep Reinforcement Learning and GANs: Introduction 00:02:53
  2. Lesson 1: The Foundations of Artificial Intelligence
    1. Learning objectives 00:00:47
    2. 1.1 The Contemporary State of AI 00:25:40
    3. 1.2 Applications of Generative Adversarial Networks 00:16:27
    4. 1.3 Applications of Deep Reinforcement Learning 00:19:38
    5. 1.4 Running the Code in these LiveLessons 00:06:03
    6. 1.5 Review of Prerequisite Deep Learning Theory 00:14:10
  3. Lesson 2: Generative Adversarial Networks (GANs)
    1. Learning objectives 00:00:49
    2. 2.1 Essential GAN Theory 00:08:59
    3. 2.2 The “Quick, Draw!” Game Dataset 00:06:36
    4. 2.3 A Discriminator Network 00:28:07
    5. 2.4 A Generator Network 00:18:49
    6. 2.5 Training an Adversarial Network 00:26:51
  4. Lesson 3: Deep Q-Learning Networks (DQNs)
    1. Learning objectives 00:00:55
    2. 3.1 The Cartpole Game 00:06:37
    3. 3.2 Essential Deep RL Theory 00:09:45
    4. 3.3 Essential DQN Theory 00:07:16
    5. 3.4 Defining a DQN Agent 00:29:30
    6. 3.5 Interacting with an OpenAI Gym Environment 00:18:12
  5. Lesson 4: OpenAI Lab
    1. Learning objectives 00:00:40
    2. 4.1 Visualizing Agent Performance 00:11:50
    3. 4.2 Modifying Agent Hyperparameters 00:06:21
    4. 4.3 Automated Hyperparameter Experimentation and Optimization 00:06:59
    5. 4.4 Fitness 00:03:43
  6. Lesson 5: Advanced Deep Reinforcement Learning Agents
    1. Learning objectives 00:00:48
    2. 5.1 Policy Gradients and the REINFORCE Algorithm 00:08:19
    3. 5.2 The Actor-Critic Algorithm 00:02:35
    4. 5.3 Software 2.0 00:06:43
    5. 5.4 Approaching Artificial General Intelligence 00:06:33
  7. Summary
    1. Deep Reinforcement Learning and GANs: Summary 00:01:26