Chevron Left
Retour à A Complete Reinforcement Learning System (Capstone)

Avis et commentaires pour d'étudiants pour A Complete Reinforcement Learning System (Capstone) par Université de l'Alberta

4.7
étoiles
554 évaluations
116 avis

À propos du cours

In this final course, you will put together your knowledge from Courses 1, 2 and 3 to implement a complete RL solution to a problem. This capstone will let you see how each component---problem formulation, algorithm selection, parameter selection and representation design---fits together into a complete solution, and how to make appropriate choices when deploying RL in the real world. This project will require you to implement both the environment to stimulate your problem, and a control agent with Neural Network function approximation. In addition, you will conduct a scientific study of your learning system to develop your ability to assess the robustness of RL agents. To use RL in the real world, it is critical to (a) appropriately formalize the problem as an MDP, (b) select appropriate algorithms, (c ) identify what choices in your implementation will have large impacts on performance and (d) validate the expected behaviour of your algorithms. This capstone is valuable for anyone who is planning on using RL to solve real problems. To be successful in this course, you will need to have completed Courses 1, 2, and 3 of this Specialization or the equivalent. By the end of this course, you will be able to: Complete an RL solution to a problem, starting from problem formulation, appropriate algorithm selection and implementation and empirical study into the effectiveness of the solution....

Meilleurs avis

JJ

27 avr. 2020

This is the final chapter. It is one of the easiest and it was fun doing that lunar landing project. This specialisation is the best for a person taking baby steps in the reinforcement learning.

CR

26 févr. 2020

Great course for learning the fundamentals. I liked that it tied into function approximation for deep reinforcement learning. The text book made the fundamental concepts more clear.

Filtrer par :

76 - 100 sur 117 Avis pour A Complete Reinforcement Learning System (Capstone)

par Ryan Y

4 mars 2021

Thank you very much!

par dariojavo

18 oct. 2020

Excellent material!

par Jose

29 juin 2020

excellent course

par BC

6 mai 2020

Excellent course

par RUI D

2 août 2021

Nice Course!

par Yanlin L

19 avr. 2020

GOOGD GOOD

par Chang, W C

8 nov. 2019

Enjoyable.

par 남상혁

18 janv. 2021

Very good

par Tran M D

22 mai 2020

Excellent

par A4

1 janv. 2020

awesome~

par ARTEM B

1 mars 2021

Great!

par Justin O

22 mai 2021

Great

par Adrian Y X

4 avr. 2020

I will write a longer review for the entire Specialization later, but this course does well to sum up all of the other progress you've had made thus far on the Specialization. However, you'll find that from Course 2 onwards (and this one especially), very little hand holding is given for the programming assignments. Command of numpy and python at good level are expected. Personally, having worked with OpenAI gyms before starting this specialization helped me immensely. As the instructors state, this course lays the foundation for future studies. The field of RL is simply so complex that even foundational work is challenging. Overall, a great course.

par Steven W

11 mai 2021

They mostly discuss the importance of real world experience and hyperparameter tuning in this class. The content it did have was solid and the instructors were great. The "capstone" was creating an agent to solve the Moon Lander problem, and much of the code was already written.

I would have really preferred getting experience with a real RL framework like RLLib or acme, rather than the toy libraries used by the book. It would have also been really nice to have a little more freedom and challenge, such as making us actually create an agent to solve an MDP of our own choosing and definition.

par Henry C

16 oct. 2021

A decent course to wrap up the RL specialization, with a "project" that demonstrates a "real-world" application of RL.

The word "project" is in quotes because it is structured as a (short) series of fairly short assignments with very heavy hand-holding, so very similar to previous courses.

My only complaints with this course are that the project is a bit too hand-holdy and that the course overall is quite short and thin in content. I would estimate that this course is around 1/3 the length of the previous courses in this series.

par Jing Z

2 juin 2020

The project is a decent example to go through in order to review what we learned from previous courses. However there are few key things supposed to be addressed as well: 1) What exactly the reward function is in the final project (C4M1 practice is badly designed); 2) How can we build an environment on our own; 3) Apart from Mean Squared Value Error to be minimized, what are other loss functions to choose from and what's the consideration behind.

par Dmitry S

10 janv. 2020

Good course. Summarises and puts everything in context. But would benefit from having larger programming assignments (which would make it more challenging as well) when less things are provided out of the box, and from a bit more extended and systematic overview and walk-through of the material.

par Ahmed S S A

5 mars 2020

Great course, thanks a lot really. But I do hope if we did visualize the environment to see how my agent behaves and then saves the RL agent to use it offline after being trained. Really thank you so much for making RL clear to me and interesting too :) <3

par Alaaeldin Z

24 mai 2021

I liked the project. I hoped it would be harder and enable the students to design the whole agent and environment code and be evaluated with a human grader. But overall, I was able to practice the concepts I have learnt throughout the specialization.

par Surya K

3 mai 2020

A cherry on top of the cake. This course helped me understand how to think about a novel problem and formulate and build an RL system from scratch. I thank Course Instructors, University of Alberta, and Coursera for this beautiful specialization.

par Lik M C

23 janv. 2020

The project is interesting. But the implementation left as assignments is too simple. There are too many guidance running in assignments. If more flexibility is allowed in implementing the project, it should be even more interesting.

par Mateusz K

16 nov. 2019

In my opinion, the capstone should've included more development and or programming. I liked having to develop NN action-value function approximator, but the parameter study was a bit too simple (should've had more code content).

par Narendra G

24 juil. 2020

The capstone project was great, it helped gain insights for developing a full RL agent. The RL problem though was a simple one, a more complex problem real-world problem implementation would have made this course perfect.

par Farhad A

18 juin 2020

This course provides an excellent start. It could have been a little better, though by incorporating some more deep neural nets probably and touching on some of the state-of-the-art. Anyhow, I'm glad that I enrolled.

par Wahyu G

4 avr. 2020

Not as complex as previous courses in the specialization but gives a nice refresher and lets us see the bigger picture of how the algorithms learned in the previous courses fit and differ. Amazing course!