New Comment
18 comments, sorted by Click to highlight new comments since:

Civil Engineering: Making Green Buildings

Electrical Engineering: Information Theory (looks awesome enough that I'll try to take it even though I probably will not have time)

[-][anonymous]10

Edit: Classes added after the original post.

New classes: (start in Jan/Feb 2012)

Cryptography

Design and Analysis of Algorithms I

Information Theory

Making Green Theory

Anatomy

Gah, too many interesting classes all at once. As much as I do have an interest in crypto, I think I'll skip that one for now because it looks like they're not going to be having much in the way of cryptanalysis.

I think I'll just be taking ML and possibly PGM. But... too much interesting stuff.

LEARN ALL THE THINGS!

So wait, what exactly is this? I'm guessing some form of online education, presumably a level above Khan Academy's videos (not implying anything negative about Khan Academy)?

Is there personal interaction with the professor? Are there assignments and or grades? How much time would an average class on XFrequentist's list take compared to an in-person class at a relatively good American university?

And finally, the most important question, at least for me: Do they accept anyone who signs up?

[-][anonymous]90

So wait, what exactly is this?

A series of free online lectures, combined with a schedule, quizzes, assignments, official forums/community (with which the professors interact), graded assignments (including programming assignments where appropriate) and sometimes exams.

I'm guessing some form of online education, presumably a level above Khan Academy's videos (not implying anything negative about Khan Academy)?

Yes. The original AI classe was inspired by Khan Academy, and that in turn caught the attention of the professors doing the DB and ML class.

Is there personal interaction with the professor?

The professors answer upvoted questions in videos or written responses.

Are there assignments and or grades?

Yes there are assignments. There are homework's that are graded and some classes have midterm and final exams. The latter two are measured to see how well you do in the class, when (if?) you receive your certificate of completion (which is not by default accredited, it is just a written statement from the professors not Stanford) it states your percentile.

How much time would an average class on XFrequentist's list take compared to an in-person class at a relatively good American university?

The original AI class was aiming for the same difficulty as the Stanford class. Due to various logistical reason they couldn't really pull it off. The classes are in my opinion less intensive but still at least freshman uni class level.

And finally, the most important question, at least for me: Do they accept anyone who signs up?

Yes, it is open to everyone. That's why you get sign up numbers in the 10,000s or even 100,000s.

Thank you for an extremely informative response!

I'll be taking all of those, and also human-computer interaction - only because the teacher seems high status in his field (of course, the great thing about these 0-cost courses is that if they turn out to be a disappointment, it's nothing to drop them).

In addition, if you haven't already taken it, Machine Learning will repeat.

Taking ML, it's great.

[-][anonymous]20

ML is wonderful.

Do you use any strategies in addition to Luke's efficient scholarship tips in order to handle four challenging classes at once (and presumably a job and a life)? AI was all I could handle; I had to drop ML.

My strategy for taking AI and ML was knowing some of the materiel before hand :).

Is it possible to access the material afterwards, especially from the AI class? During semester I have no time since I am currently an university student myself.

AI class videos are all on YouTube. Since their quizzes are also videos (don't ask) that's pretty much all of it. The AIMA book is much deeper than the course, but the quizzes get you to think a little extra.

How useful are PGMs? (I'm taking the ML class right now and I only want to spend a whole semester on PGMs if they're worth that much attention.)

Very, depending on what you're doing. They allow rich probabilistic models useful in scene recognition (CRFs), HMMs are hugely useful in state detection, Bayes Nets are used (amongst a ton of other things) Decision Analysis. Bishop's ML book is full of them.