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

4.8
206 個評分
45 個審閱

課程概述

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...

熱門審閱

KN

Oct 03, 2019

Great course! The notebooks are a perfect level of difficulty for someone learning RL for the first time. Thanks Martha and Adam for all your work on this!! Great content!!

UZ

Nov 23, 2019

Good balance of theory and programming assignments. I really like the weekly bonus videos with professors and developers. Recommend to everyone.

篩選依據:

26 - Sample-based Learning Methods 的 43 個評論(共 43 個)

創建者 Mark J

Sep 23, 2019

In my opinion, this course strikes a comfortable balance between theory and practice. It is, essentially, a walk-through of the textbook by Sutton and Barto entitled, appropriately enough, 'Reinforcement Learning'. Sutton's appearances in some of the videos are an added treat.

創建者 David R

Dec 10, 2019

Course is not easy, videos presentation is a bit dull - but the material is cool and interesting, and the additional quizzes, videos and especially notebooks make it a great course - you learn a lot and see progress. Highly recommended.

創建者 Shashidhara K

Dec 12, 2019

This course required more work than the 1st in the series, (may be i took it lightly as the first was not that difficult). Request : Please include some worked examples (calculations) or include in graded/ungraded quiz, will be nice.

創建者 Ashish S

Sep 16, 2019

A good course with proper Mathematical insights

創建者 David P

Nov 03, 2019

Really a wonderful course! Very professional and high level.

創建者 Umut Z

Nov 23, 2019

Good balance of theory and programming assignments. I really like the weekly bonus videos with professors and developers. Recommend to everyone.

創建者 AhmadrezaSheibanirad

Nov 10, 2019

This course doesn't cover all concept of Sutton book. like n-step TD (chapter7) or some Planning and Learning with Tabular Methods (8-5, 8-6, 8-7, 8-8, 8-9, 8-10, 8-11), but what they teach you and cover are so practical, complete and clear.

創建者 Pachi C

Dec 08, 2019

Great and fantastic course!!!

創建者 Antonio P

Dec 13, 2019

Great Course

創建者 Max C

Oct 24, 2019

Some of the programming homeworks were difficult to debug due to the feedback from autograder being unhelpful.

創建者 David C

Oct 10, 2019

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

Sep 20, 2019

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.

創建者 Scott L

Sep 26, 2019

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 JDH

Sep 23, 2019

Rating 4.3 stars – so far (first two classes combined)

Lectures: 4.0stars

Quizes: 4.0stars

Programming assignments: 4.5stars

Book (Sutton and Barto): 4.5stars

In the spectrum from the theoretical to practical where you have, very roughly,...

(1) “Why”: Why you are doing what you are doing

(2) “What”: What you are doing

(3) “How”: How to implement it (eg programming)…

...this is a “what-how” class.

To cover the “why-what” I strongly recommend augmenting this class with David Silver’s lectures (on Youtube) and notes from a class he gave at UCL. This covers more of the theory/math behind RL but covers less on the coding. Combined together with this class it probably comprises the best RL education you can get *anywhere*, creating a 5-star combo.

http://www0.cs.ucl.ac.uk/staff/d.silver/web/Teaching.html

創建者 Yicong H

Dec 05, 2019

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.....

創建者 Navid H

Oct 16, 2019

definitely interesting subjects, but I do not like the teaching method. Very mechanic and dull, with not enough connection to the real world

創建者 Neil S

Sep 12, 2019

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 Chan Y F

Nov 04, 2019

The video content is not elaborated enough, need to read the book and search on the web to understand the idea