Reinforcement Learning is a subfield of Machine Learning, but is also a general purpose formalism for automated decision-making and AI. This course introduces you to statistical learning techniques where an agent explicitly takes actions and interacts with the world. Understanding the importance and challenges of learning agents that make decisions is of vital importance today, with more and more companies interested in interactive agents and intelligent decision-making. This course introduces you to the fundamentals of Reinforcement Learning. When you finish this course, you will: - Formalize problems as Markov Decision Processes - Understand basic exploration methods and the exploration/exploitation tradeoff - Understand value functions, as a general-purpose tool for optimal decision-making - Know how to implement dynamic programming as an efficient solution approach to an industrial control problem This course teaches you the key concepts of Reinforcement Learning, underlying classic and modern algorithms in RL. After completing this course, you will be able to start using RL for real problems, where you have or can specify the MDP. This is the first course of the Reinforcement Learning Specialization.
SM
Excellent course, with a very nice presentation style, both the professors are excellent in their presentations and the material is well researched and delivered. A very valuable course.
U
The book is essential reading. It took me longer than the estimates to do the reading and the programming assignments. I would have liked more gridworld examples to get a faster hang of it.
MN
The concepts may sound confusing in the beginning, but as you go forward you find it interesting and understanding. I suggest you completely read the reading assignments before watching the videos.
KS
All the concepts were well explained and this course was perhaps the best I have found for RL.Great efforts have been put into making the course and It goes well in line with the suggested textbook.
AA
This course is the best course for anyone who needs to enter into the field of RL. Content within the course is excellent and instructors have explained each and every topic very well.
AB
Concepts are bit hard, but it is nice if you undersand it well, espically the bellman and dynamic programming.Sometimes, visualizing the problem is hard, so need to thoroghly get prepared.
HS
One of the best courses I finished on Coursera, I really like the structure of the course. Textbook is also provided which really helps. Looking forward to next course in the series.
YW
Clear instruction and insightful exercises! Enjoy this course! Also, please read the book if you want to understand better about the course materials and rationales behind the exercises.
RD
I was so confused about the fundamental concepts, but doing this course has given me a solid foundation of RL.This is a must-do course if you are starting with Reinforcement Learning.
AM
This course is great for people who are just starting out. The programming assignments are really great and practically introduce you to the basic concepts of reinforcement learning.
GJ
The concepts are explained in a simple and illustrative manner which helps in getting a better understanding of the concepts. The assignments and quizzes are also really well designed.
HT
This course is one of the best I've learned so far in coursera. The explanations are clear and concise enough. It took a while for me to understand Bellman equation but when I did, it felt amazing!
Showing: 20 of 689
I found the explanations of theory of RL to replicate what was written in the book. Without examples the videos were no value add.
I had to go through the RL course by David Silver in youtube to understand the concepts.
Course material is standard and mostly follows Sutton and Barto textbook. Unfortunately, most contents overlap with the existing reinforcement learning course on Coursera and David Silver's youtube videos. The course will be much more useful if it covers more practical stuff instead.
I was very disappointed that the free trial period ends before my assignments are graded by peers. I would suggest that the course should be arranged so that students can finish it during the free period.
I am not sure if RLglue is an appropriate package to use in the exercise, as it is not as a standard tool as all practitioners are familiar with. If the instructors believe it is something useful in the future, they should explain it more in detail in the lecture.
I give the course a low rating for several reasons, the first being the most important one: The instructors basically completely absent. Having issues or problems? They don't bother. Not a single reply from either instructor in the forums for months or years. Second: Flawed and inprecise notebooks. Well known issues with random numbers, but no updates. Incorrect book references which will let you implement formulas other than intended. Third: Tons of short videos with 30% summary and "what you will learn", which is ridiculous for 3 minute videos. Fourth reason: Mathematical depth missing after the first subcourse. Suggestion: Watch the David Silver and Stanford youtube lessons instead. For free and better explained. Compared to, for instance, Andrew NGs specialization, this one is really bad mostly thanks to the complete disinterest of the instructors.
Fantastic Course. That's the RL MOOC I have been waiting for so long. No surprise it is from Students of RL guru R. Sutton at Uni of Alberta. Very clearly and simply explained. Exercise and Test difficulty spot on. Wouldn't change a yota from this Course. Can't wait to access the rest of this specialization
Too much history and talks about who we are. Not efficient time spent.
Poor explanations with count on book. Not suitable for listening or on the go study.
Easy things made so complicated. (First you forced to get into math and other roots in the book and then video with some explanations when already not needed.) And could be explained better, not in 3 minutes. This is red, this is green and here we go - Malevich.
Got some insights but not happy about time spent.
The main reason I enrolled in this course was to have an opportunity to have my questions answered. I had already gone through videos of RL lectures from different universities before this. Hence, the value of the course diminished for me when some of my questions were not always answered by the TAs or the Staff
Not much help available on forums
Is a very good introduction to Reinforcement Learning. It also gives a very nice foundation of the basics of this area without being shy of showing some math. Could use more examples about modeling real world problems as MDPs but otherwise is a very complete course.
I understood all the necessary concepts of RL. I've been working on RL for some time now, but thanks to this course, now I have more basic knowledge about RL and can't wait to watch other courses
The lectures are not indicative of the problem sets. Both are very interesting and cover the materials well, but as a beginner with Dynamic Programming the bugs in the Notebooks are hard to distinguish from a lack of knowledge. The locked cells also make it hard to iterate slowly, to see the sensitivity of the algorithms to certain variables. Overall it is a great learning experience and the staff/mentors step in for support.
I've just finished this course, it is really wonderful and I learnt a lot, as a professional Backend Developer without a formal background in Machine Learning. It has a lot of mathematical theory and exercises, derivations, really good explanations, and even some coding tasks to apply this knowledge.
At first I was doubtful I would make it to the end as I was feeling rusty on my maths since I didn't practice them much after university, but with effort and patience I was able to see how everything is built from the ground up and got a really good picture of how the fundamentals of RL work.
The course is based on the famous "Reinforcement Learning: An Introduction" by Sutton and Barto, the 2nd edition of which was only released recently, and which the Data Scientists I work with say is the go-to book for RL. The book is a magnificent resource available digitally for free, but I have enjoyed this course so much that I got the physical version, and after auditing the course for a week decided to jump in to do my best in the whole specialization.
Very clear and engaging presentation, well thought out and typical Coursera-style programming assignments. Definitely looking forward to taking the rest of the sequence.
The content was pretty good. However, the final requirement on the final programing assignment was vague and required a very specific implimentation to match test cases. It was frustrating to have to search the forums for the exact sequence used to recreate a very specific dataset.
This course was super helpful. I had tried a couple other online introductions to RL, but this was the only one where I could really engage and learn the material effectively. Would recommend!
Solid introduction, but materials could be better prepared, e.g., overview of important concepts / formulas. Furthermore I would have liked to have more programming assignments and also more quizzes to practice the theory.
It was a good course, but I feel like there could have been programming assignments for week 2 and 3 to really help understand the bellman equations. Also, the jupyter notebook was pretty buggy sometimes.
Course instructors should improve their teaching style by writing equations in hand and explaining point-by-point. There is no need to show their faces in the video while teaching. They sounded like 'radio' throughout the course.
Not recommended
Concepts are bit hard, but it is nice if you undersand it well, espically the bellman and dynamic programming.
Sometimes, visualizing the problem is hard, so need to thoroghly get prepared.
To be honest I didn't like videos in the course. Lectors read prepared text as robots. No pauses in places that are hard to understand. I had to do lots of replays to understand vids. Without reading the book I wouldn't be able to understand the material. Having read the book it's questionable if there is a value in watching videos. Also there are only 2 programming assignments and in each assignment it's required to write only a couple of functions while the rest of the code is already written. Programming assignments were like puzzles where you need to understand the code written and plug missing part. It's not creating my own program.