Lin Yang is by far the worst lecturer I've ever had.

The lectures don't make much sense but the textbook is ok, and the class is project based and you learn a lot from the projects.

Do not recommend it at all!
Awful explanation, no slides, no recording! I did not understand what he is trying to tell. Not only me, other also failed to get his point.
Exams are extraordinary hard, the average was 65/100.

Since this is a 'graduate' and 'special topics' class, I recommend that people take it only if they are interested and willing to put enough effort into learning the material. I loved that we focused on the theory part, if you're planning to work on this field or in a related area, it makes it easier to understand the proofs and ideas used in recent papers. HW assignments were based on the proofs we covered in the class and often required you to do a literature search and understand some of the fundamental papers in the field (this may be a downside for some people but imo it is how graduate classes should be). There are no midterm or final exams, instead, you do a group project where you present a paper and write a report, which was fun and helpful. It's one of the classes where I learned a lot, unlike the others where you either memorize or repeat the same things all over again.

The instructor is extremely smart and knowledgeable in the subject, but it could be better if lecture notes were a bit more organized (since it is a proof-heavy class, it is sometimes hard to focus or follow the proofs if there is a typo or the text is not readable).

Lin Yang is a great Professor. I'd recommend this class if you are looking for a light workload with engaging numerical computing material. This class had 1 quiz every week which were very easy as the professor would help out and allow students to exchange ideas. The midterm was also very fair. The group project is fun because you apply what you learned in class. The homework was also very easy as it would take about 1-2 hours maximum for each homework.

This class is an introduction to reinforcement learning. It starts with the basics of RL such as Markov Decision Processes and ends at... well I wouldn't know because most students (including myself) stopped watching Prof Yang's lectures. They were that bad. In general, this class is really disorganised, does not cover the material well and has no interest in teaching anything.

Regarding the organisation, these are some of the events that took place:
-Mid-Week 1: there wasn't any Bruinlearn page for the class, so students emailed Prof Yang until he made one.
-Week 4: the Prof finally hires a TA.
-Week 9: final project proposal is marked 5 days before the entire project is due.
-Week 10: the "midterm" exam, scheduled to happen on Week 5, takes place.
-3 Weeks after end of term: a grade is published to MyUCLA, with nothing else said.

More or less everything happened either late, never and with no feedback from the instructors.

The lectures are not much better. Most of the material is borrowed from a public UCL lecture series from 2015 and is replicated using inconsistent slides full of errors. Sometimes these errors are fixed, but then the fixed versions are not actually uploaded for students to use (even though we asked for this repeatedly).

Looking back, I should have seen the warning signs in the first week and dropped this class immediately. To get an A in this class, I pretty much studied every university's RL class save for the one I was actually paying for. This class is a disgrace.

This is not an easy class.
8hw 20% (some from textbooks with solutions, others from additional book without solutions)
1 mid 30%
1 final 50%
Luckily we did everything online. And more lucky, final was one-day limit but unable to do them all.
One thing annoyed me was that i got 0 point in one of problems in the midterm no matter what I solved up to 2 pages. Still zero points without a partial point.

Seems nice but probably due to being new is terrible at teaching. He will explain a theorem or algorithm by just putting the math on the board without first explaining what its used for nor what the english-explanation of the method is; he just dives straight into the math with no context repeatedly. That being said his grading was very easy, the textbook is pretty good, and he can easily clarify any issues you have from reading.

Exams and homework are very hard but the curve is very nice.