Practical AI with Python and Reinforcement Learning (Udemy.com)

Good course to learn Reinforcement Learning. It's getting outdated, though. It uses gym, which is deprecated and is now superseded by gymnasium. I could adapt some lectures to gymnasium, but in the lectures using keras-rl2 I couldn't, because the library was written for gym and is now archived since May 2021, so it wasn't updated to support gymnasium. I had to make a virtual environment with the legacy gym to complete the last lectures. Otherwise, solid course.

I loved this course except for one thing. The packages used in the course are several years old by the time I took it so much of the code failed. I had to spend a significant amount of time upgrading to newer packages and that changed a lot of code. While this was actually useful for learning, it greatly expanded the time needed and some things never worked. These courses should specify what versions of all packages are needed right at the beginning. Better yet, they should be upgraded when significant package changes happen.

For me personally, I was given a task at work to apply an academic paper to real world problems. It was using DQN and I only knew about it from theory, not application. My PhD dissertation was in the application of Deep Learning to radar frequency machine learning (RFML), so I am familiar with ML concepts but not specifically RL. This course has helped me tremendously. It is not just the application but the theory and intuition behind it. And it was a great experience. There were some errors in the code but nothing that I could not figure out with research which is what I do for a living (R&D). Would easily recommend this course and instructor to someone else.

This course has been stopped from updating. All the Reinforcement Learning section is outdated and no one has answered questions related to RL, not jut for me, but for any one else as well. I see Jose has answered questions up until last week. But no questions have been answered related to Reinforcement learning and GYM library, project execution issues. Its really so unexpected of Jose. I have taken so many classes from him. But this one forced me into a different experience altogether. Disappointed,

The most complete, step-by-step introduction to Reinforcement Learning in OpenAI. Much better professor than pretty much all others in the same subject here in Udemy. Congratulations. Only hoping now for a follow up with real life, non-Atari implementations of RL that could be applied using these techniques.

Same method of delivery as you can expect from the instructor. The course content is fun and engaging. However, despite instructor encouraging students to ask questions, I have noticed the questions were left unanswered by instructor and TA. This has become more frequent in many courses from this instructor, which I find it very disappointing. Answering questions would help students and many others would might have the same questions and develop understanding from the course.

Very excellent course. I appreciate that this course was probably more difficult to design than others. I very especially appreciated the demonstration of the manual implementation of the replay buffer, as well as the demonstration of conversion of the hand-implemented Snake game to the OpenAI Gym environment. The selection of Keras-RL2 was appropriate for its level of abstraction -- not too low, not too high.

Exceptionally simplified yet indepth lessons on Reinforcement Learning - particularly the math involved. I enjoyed the lessons really a lot, and using this as the basis to learn further in reading books and understanding higher concepts in Reinforcement Learning.

This course was an excellent introduction of reinforcement learning and OpenAI's gym environment. The explanations were clear and easy to follow, the references to the original papers and Richard Sutton's book were also useful in getting more in-depth knowledge of the RL theory. The notebooks were very easy to follow and made the exercises straightforward. I also appreciated the background topics of CNNs, Tensorflow and Keras. While I didn't need them for the class as I have experience with these areas, I plan to go back and review them, as I find I always pick up things I didn't know in Jose's classes.

I've been trying to learn RL from watching David Silver's lectures on the DeepMind youtube, and the book. It's almost like they make it complicated on purpose. I appreciate you explaining the RL equations in plain english and break it down in code.