Chevron Left
返回到 Sample-based Learning Methods

學生對 阿尔伯塔大学 提供的 Sample-based Learning Methods 的評價和反饋

927 個評分
192 條評論


In this course, you will learn about several algorithms that can learn near optimal policies based on trial and error interaction with the environment---learning from the agent’s own experience. Learning from actual experience is striking because it requires no prior knowledge of the environment’s dynamics, yet can still attain optimal behavior. We will cover intuitively simple but powerful Monte Carlo methods, and temporal difference learning methods including Q-learning. We will wrap up this course investigating how we can get the best of both worlds: algorithms that can combine model-based planning (similar to dynamic programming) and temporal difference updates to radically accelerate learning. By the end of this course you will be able to: - Understand Temporal-Difference learning and Monte Carlo as two strategies for estimating value functions from sampled experience - Understand the importance of exploration, when using sampled experience rather than dynamic programming sweeps within a model - Understand the connections between Monte Carlo and Dynamic Programming and TD. - Implement and apply the TD algorithm, for estimating value functions - Implement and apply Expected Sarsa and Q-learning (two TD methods for control) - Understand the difference between on-policy and off-policy control - Understand planning with simulated experience (as opposed to classic planning strategies) - Implement a model-based approach to RL, called Dyna, which uses simulated experience - Conduct an empirical study to see the improvements in sample efficiency when using Dyna...



Great course, giving it 5 stars though it deserves both because the assignments have some serious issues that shouldn't actually be a matter. All the other parts are amazing though. Good job


Really great resource to follow along the RL Book. IMP Suggestion: Do not skip the reading assignments, they are really helpful and following the videos and assignments becomes easy.


126 - Sample-based Learning Methods 的 150 個評論(共 188 個)

創建者 John H


It was good.

創建者 Oren


Fun course!

創建者 Jialong F


learn much

創建者 Sohail



創建者 LuSheng Y


Very good.

創建者 Oriol A


Very good

創建者 Pouya E



創建者 Artod d



創建者 Justin O



創建者 chao p



創建者 Alejandro S H


The course material are great. You will learn a lot from the assignments and from the book. The videos are a good refresher of what you'll read in the book, sometimes with improved animated visuals. However, I've a few nitpicks that prevent me from giving it 5 stars. (1) The instructors do not interact much with the students in the forum (if at all). (2) There's an inaccuracy in one of the videos that (as of the instant I'm doing this review) hasn't been fixed yet. (3) The quizzes sometime ask for questions that are NOT in the assigned homework materials (I'm thinking now about a question about prioritized sweeping in the planning section, but there are others). This is not a big deal, the questions will ring a bell immediately and you will find the section of the book where the answer lies (or you will answer out of common sense). (4) There's a video about applying RL in continuous tasks in robotics (purely motivational, not part of the syllabus) that is missing the second part. I'm guessing it's in the next course?

創建者 Oscar R R M


I have to admit that the practical programming tasks are excellent. The discussions sections are well maintained by professors. The recommended readings are good. The exams are fine but some questions can be a bit confusing or not clear enough.

I consider the lectures to be the weakest part of this course. In lectures, they only provide a short summary of what is already written in the book, which may be useful for some people, but I prefer the old way of long lectures in a blackboard full of mathematical proofs and historical notes providing true understanding of the topic. The invited lecturers are pretty good, since they show some historical notes or cutting edge projects.

So far this is the best course I have found on this topic.

創建者 Neil S


This is THE course to go with Sutton & Barto's Reinforcement Learning: An Introduction.

It's great to be able to repeat the examples from the book and end up writing code that outputs the same diagrams for e.g. Dyna-Q comparisons for planning. The notebooks strike a good balance between hand-holding for new topics and letting you make your own msitakes and learn from them.

I would rate five stars, but decided to drop one for now as there are still some glitches in the coding of Notebook assignments, requiring work-arounds communicated in the course forums. I hope these will be worked on and the course materials polished to perfection in future.

創建者 Jorge A C


Excellent course, it complements very well the textbook by Sutton and Barto. The quizzes focus on conceptual issues, some of them not covered in the video lectures but presented in the textbook, which should be read carefully and in-depth. The programming assignments are based on the textbook examples and they are very effective in reinforcing what the course teaches despite not being that difficult nor time consuming. Although I have been able to navigate the course on my own I am taking one star off because there has been almost no feedback from the instructors in the discussion forums when I took the course in August 2020.

創建者 Arsham M


The course content is of high quality and it would not be trevial even for people with a background in computer science and machine learning. The course works are very well defined, structured and clear. The only thing I guess could be improved is the way that content is delivered by the instructors. Overall, I do recommend taking this course to people who want to start exploring RL or who wants to gain better proficiency in Python programming for RL.

創建者 Stefano P


The course is overall very good: lectures are very clear, quizzes are challenging and the course relies on a text book, provided when you enroll. The only weak point, but not a serious issue, is that most of the lectures do not add content to what is in the book. Since studying the book is in fact mandatory, they could have used the lectures to better explain some concepts, assuming people read the book. Sometimes they do, but not so often.

創建者 Michael R


Lectures were good, but not as intuition building as in the first course. The biggest strength of this course is that it follows a good textbook and expects you to read it. Quizzes and programming assignments are good for learning, but all the programming assignments are very scripted/guided. As a result, I think that it would be very easy to finish this course and still not be able to set up a sample-based learning model on your own.

創建者 Qianbo Y


A very thorough and well-designed course. It covers almost all important topics of tabular methods of Reinforcement Learning and follows the RL textbook very well. The only imperfectness of this course is the way instructors explaining the concepts. It is obvious that the instructors are reading off the scripts and not particularly explaining with their own words, which makes the lecture part less comprehensible.

創建者 Scott L


This course series is an incredible introduction to the basics of reinforcement learning, full stop. The course ... style, if you will, is a bit weird at first, but it seems to have been done on purpose with the aim of making the course somewhat timeless; they are presenting maths that will not change, in a format that will (hilariously) be no more slightly corny and weird in 2030 as it is in 2019.

創建者 Jimson T


The course includes very clear explanation of sample_based learning methods. However, if the program assignment can include more detail explanation per steps, it will help students quickly realize what they are going to implement. In addition, an clear content page that shows the purpose of each cell at the beginning of the assignment can help students understand whole structure of the code.

創建者 David C


A very good course. The lectures are brief and provide a quick overview of the topics. The quizzes require more in-depth reading to pass (covering material not discussed in the lectures) and the projects are difficult but rewarding and really help to cement the information. My only suggestion would be to lengthen the lectures to provide more discussion on the topics.

創建者 Marius L


Overall, I found the course well made, inspiring and balanced. The videos really helped me to understand the rather austere textbook. I give 4 stars because some of the coding exercises felt more like work in progress, without the help of other students I would not have been able to overcome these issues.



I like the structure of the course and it is in general well done. However, the programming assignment system is occasionally difficult to work with. I would suggest that the lectures be a little longer and a little more details, going into the nuts and bolts of how the algorithms work.

創建者 Amit J


Course material is good but I'm slightly disappointed by the quality of lectures. They sound just short monologues based on snippets from the book. I wish they were slightly longer and more original going into various aspects and examples of the very interesting subject material.

創建者 Yicong H


Jump for here to there, it's nice to have all these algorithms. My gut tells me something is not correct. Too much focus on experience, which means a lot of data. The model part is touched very little, and main focus is on when model is wrong.....