Reinforcement Learning with Pytorch

Learn to apply Reinforcement Learning and Artificial Intelligence algorithms using Python, Pytorch and OpenAI Gym

3.65 (394 reviews)
Udemy
platform
English
language
Data Science
category
instructor
2,685
students
7 hours
content
Aug 2020
last update
$59.99
regular price

What you will learn

Reinforcement Learning basics

Tabular methods

Bellman equation

Q Learning

Deep Reinforcement Learning

Learning from video input

Description

UPDATE:

All the code and installation instructions have been updated and verified to work with Pytorch 1.6 !!


Artificial Intelligence is dynamically edging its way into our lives. It is already broadly available and we use it - sometimes even not knowing it  - on daily basis. Soon it will be our permanent, every day companion.

And where can we place Reinforcement Learning in AI world? Definitely this is one of the most promising and fastest growing technologies that can eventually lead us to General Artificial Intelligence! We can see multiple examples where AI can achieve amazing results - from reaching super human level while playing games to solving real life problems (robotics, healthcare, etc).

Without a doubt it's worth to know and understand it!

And that's why this course has been created.

We will go through multiple topics, focusing on most important and practical details. We will start from very basic information, gradually building our understanding, and finally reaching the point where we will make our agent learn in human-like way - only from video input!

What's important - of course we need to cover some theory - but we will mainly focus on practical part. Goal is to understand WHY and HOW.

In order to evaluate our algorithms we will use environments from - very popular - OpenAI Gym. We will start from basic text games, through more complex ones, up to challenging Atari games

What will be covered during the course ? 

- Introduction to Reinforcement Learning

- Markov Decision Process

- Deterministic and stochastic environments

- Bellman Equation

- Q Learning

- Exploration vs Exploitation

- Scaling up

- Neural Networks as function approximators

- Deep Reinforcement Learning

- DQN

- Improvements to DQN

- Learning from video input

- Reproducing some of most popular RL solutions

- Tuning parameters and  general recommendations

See you in the class!

Content

Welcome to the course

Welcome!
Before you start - Videos quality!
Resources

Introduction

Introduction #1
Introduction #2
Introduction #3
Introduction #4
Environment setup / Installation
Lab. OpenAI Gym #1
Lab. OpenAI Gym #2
Lab. OpenAI Gym #3
Lab. OpenAI Gym #4

Tabular methods

Deterministic & Stochastic environments
Rewards
Bellman equation #1
Bellman equation #2
Resource - code
Lab. Algorithm for deterministic environments #1
Lab. Algorithm for deterministic environments #2
Lab. Algorithm for deterministic environments #3
Lab. Algorithm for deterministic environments #4
Lab. Test with stochastic environment
Q-Learning
Lab. Algorithm for stochastic environments
Exploration vs Exploitation
Lab. Egreedy
Lab. Adaptive egreedy
Bonus Lab. Value iteration
Homework
Homework. Solution
Homework. Tuning

Scaling up

Scaling up
Neural Networks review
Lab. Neural Networks review #1
Lab. Neural Networks review #2
Lab. Random CartPole
Lab. Epsilon egreedy revisited
Lab. Pytorch updated ( version 0.4.0 )
Article. Pytorch updated!
Lab. OpenAI Gym + Neural Network #1
Lab. OpenAI Gym + Neural Network #2
Lab. OpenAI Gym + Neural Network #3
Lab. Extended logging

DQN

Deep Reinforcement Learning
Lab. Deep Reinforcement Learning
Lab. Tuning challenge
Experience Replay
Lab. Experience Replay #1
Lab. Experience Replay #2
Lab. Experience Replay #3
DQN
Lab. DQN

DQN Improvements

Double DQN
Lab. Double DQN
Dueling DQN
Lab. Dueling DQN
Lab. Dueling DQN Challenge

DQN with video output

CNN Review
Lab. Random Pong
Saving & Loading the Model
Lab. Pong from video output #1
Lab. Pong from video output #2
Lab. Pong from video output #3
Lab. Pong from video output #4
Lab. Pong from video output #5
Lab. Pong from video output #6
Potential improvements
Article. Stacking 4 images together

Final notes

What's next?

Screenshots

Reinforcement Learning with Pytorch - Screenshot_01Reinforcement Learning with Pytorch - Screenshot_02Reinforcement Learning with Pytorch - Screenshot_03Reinforcement Learning with Pytorch - Screenshot_04

Reviews

Ariadna
March 21, 2023
El curso estuvo muy bien, las explicaciones detalladas, el código paso por paso. Me encantó y aprendí muchísimo
Jeremy
July 22, 2022
One important missing part is how to define our own environments / transform some example issue into a RL training environment. This course focuses mainly on existing algorithms and their implementation, which is good. I wouldn't expect more than one lesson on how to create a custom environment (or maybe 2 depending on the nature of the data: vision vs tabular for instance). Toy problems are fine to understand how things work but when we want to push things further, datasets are also a central point.
Charles
July 25, 2021
Thank you so much for this wonderfull course. The instructor was very clear. I particularly appreciated that he coded everything from scratch and show the papers.
Maurice
July 4, 2021
great hands-on introduction to reinforcement learning on open ai gym environments. The course starts with simple games, introducing principles gently but builds up to a more complicated game and introducing further principles required to solve it. Finally, Pong is solved using reinforcement learning. What I like esspecially about this course is the gradual build up of complexity. You truely get to appreciate the advanced techniques (introduced by deepmind only a couple of years ago!) in simple terms. Simple terms is of course relative, you still need to have your brain active and it does take some effort to really understand what is going on. I can recommend this course to anyone who learns from an hands-on approach.
Peter
February 28, 2021
I like how he walks through the programs and builds up from the previous programs! Some theory but then it's translated into code
Reinhard
January 31, 2021
Fast jedes Kapitel beginnt mit einem link zu einem Dokument, das man unbedingt gelesen haben sollte. Ohne diesen Schritt bleibt vieles unklar.
L
December 24, 2020
Course is a good survey and gives practical coverage of basic RL topics. There is a lot of detailed information on PyTorch issues, especially when working with batching and CNN. It would be very helpful to have clearer high-level explanations of bigger concepts. E.g. value iteration, Q learning, and deep Q network learning.
Matei
November 7, 2020
I really like the step-by-step tutorial-style app and the fact that the speaker takes the time to introduce even basic things like torch.zeros or torch.rand.
Cem
November 11, 2018
It was a nice introduction to RL. Fundamental concepts like Bellman Equation and Q-Learning are given in a simple and instructive way. Then the DQN, a popular Deep RL method, is introduced in a detailed way with the support of several exercises and the brief discussion of the original publications in chronological order.
SP
November 4, 2018
this course is one of the best rl in practice I've been around the web he will teach you from very beginning with very basic code very basic concept step by step to more advance concept and very most of the time he explain it clearly!!! other thing that i like about the course is this course doesn't like some of ML course on udemy that skip the math equation the instructor really teach you the math when it needed not to much not too little short and sweet course!!!
Anthony
September 11, 2018
Initially I wasn't sure if this course was above my head as I have had no experience with Machine learning, but I took the plunge and I was very pleasantly surprised, the tutor explained things very clearly and if you followed the videos you would 100% get the same code result as provided. The tutor was also actually funny, even though he might not have always intended to be. I went through all the content in a couple days, really exciting stuff, I'll definitely go through the content again. Overall a perfect start to my Machine Learning education. Minor Note, if you have windows, you will need to do some extra research to get a couple of the tools to work. This Course only really provides the installation steps for Mac and Linux. Which is fair enough as the full toolset only officially support these two OS'
Steven
September 8, 2018
Great course on RL, and particuraly DQN. For a beginner like me, I really appreciated that a lot of PyTorch details are very well explained. I hope that other courses on RL will be release because it's very captivating. Thanks for this one.
Massimiliano
August 8, 2018
excellent course; simple and precise explanations; it also deals with topics that are not often encountered; congratulations to the author!
Gleb
August 2, 2018
This is the best RL course on Udemy covering state-of-the-art deep q-learning approaches except actor-critic algorithm.
Joey
June 12, 2018
The course seems good for me so far but the audio quality is very poor and the video could be better. Edit* the audio quality improves after a few videos, definitely a good course for anyone with some idea about reinforcement learning and an interest in making ai for games. If you don't know anything and you want to understand in depth how things work you might end up reading a good few articles.

Charts

Price

Reinforcement Learning with Pytorch - Price chart

Rating

Reinforcement Learning with Pytorch - Ratings chart

Enrollment distribution

Reinforcement Learning with Pytorch - Distribution chart
1678738
udemy ID
5/6/2018
course created date
7/3/2019
course indexed date
Bot
course submited by