Search icon CANCEL
Subscription
0
Cart icon
Your Cart (0 item)
Close icon
You have no products in your basket yet
Arrow left icon
Explore Products
Best Sellers
New Releases
Books
Videos
Audiobooks
Learning Hub
Conferences
Free Learning
Arrow right icon
Arrow up icon
GO TO TOP
Deep Reinforcement Learning Hands-On

You're reading from   Deep Reinforcement Learning Hands-On Apply modern RL methods, with deep Q-networks, value iteration, policy gradients, TRPO, AlphaGo Zero and more

Arrow left icon
Product type Paperback
Published in Jun 2018
Publisher Packt
ISBN-13 9781788834247
Length 546 pages
Edition 1st Edition
Languages
Arrow right icon
Author (1):
Arrow left icon
Maxim Lapan Maxim Lapan
Author Profile Icon Maxim Lapan
Maxim Lapan
Arrow right icon
View More author details
Toc

Table of Contents (21) Chapters Close

Preface 1. What is Reinforcement Learning? 2. OpenAI Gym FREE CHAPTER 3. Deep Learning with PyTorch 4. The Cross-Entropy Method 5. Tabular Learning and the Bellman Equation 6. Deep Q-Networks 7. DQN Extensions 8. Stocks Trading Using RL 9. Policy Gradients – An Alternative 10. The Actor-Critic Method 11. Asynchronous Advantage Actor-Critic 12. Chatbots Training with RL 13. Web Navigation 14. Continuous Action Space 15. Trust Regions – TRPO, PPO, and ACKTR 16. Black-Box Optimization in RL 17. Beyond Model-Free – Imagination 18. AlphaGo Zero Other Books You May Enjoy Index

PG on Pong

As covered in the previous section, the vanilla PG method works well on a simple CartPole environment, but surprisingly badly on more complicated environments. Even in the relatively simple Atari game Pong, our DQN was able to completely solve it in 1M frames and showed positive reward dynamics in just 100k frames, whereas PG failed to converge. Due to the instability of PG training, it became very hard to find good hyperparameters, which is still very sensitive to initialization.

This doesn’t mean that the PGs are bad, because, as we’ll see in the next chapter, just one tweak of the network architecture to get the better baseline in the gradients will turn PG into one of the best methods (Asynchronous Advantage Actor-Critic (A3C) method). Of course, there is a good chance that my hyperparameters are completely wrong or the code has some hidden bugs or whatever. Regardless, unsuccessful results still have value, at least as a demonstration of bad convergence...

lock icon The rest of the chapter is locked
Register for a free Packt account to unlock a world of extra content!
A free Packt account unlocks extra newsletters, articles, discounted offers, and much more. Start advancing your knowledge today.
Unlock this book and the full library FREE for 7 days
Get unlimited access to 7000+ expert-authored eBooks and videos courses covering every tech area you can think of
Renews at $19.99/month. Cancel anytime
Banner background image