COMS W4701

Four projects, a midterm, and a final. I wish we got more guidance on what would be on the tests, and the project descriptions are vague which means you'll probably lose points unless you triple-check your program with TAs and the professor.

Easy midterm and final (Although Voris will be an ass during them so yay?) Assignments were unnecessary tedious, but the AI parts were easy

4 projects - 20% each 1 midterm - 15% 1 final - 20% class participation - 5% (totally arbitrary)

1 midterm, 1 final (both reasonable), and a number of programming assignments that you can spend as much time on as you want to.

Depends on how much you want to learn. Not much work at all - easy to get a B, besides for project. Even easy to get A.

I have never been more let down by a Columbia class....ever. Artificial Intelligence is supposed to be the most interesting field in CS. Programming robot intelligence? Learning about machine learning? Discovering how your video games are so smart? Nope, but we will spend half of the semester on depth-first search, breadth-first search and other search algorithms. And then Stolfo will skim past everything you actually wanted to learn in AI. That being said, Stolfo is the sassiest professor I've ever seen and a blast to watch on x2 speed (online lectures!!). Laughs will be chuckled, but by the end of the lecture you'll have a faint idea of what he was talking about.

Dear might possibly be the worst professor I've had at Columbia so far. He made easy concepts extremely difficult to understand and did not explain concepts well at all, would highly recommend people to not take this class with Dear. Exams were extremely difficult and expected students to be able to completely code out homework assignments without providing knowledge on how.

The rest of the reviews are pretty accurate about this class but I figured I should add in a bit to just explain Prof. Voris. Let me start by saying Prof. Voris is a terrible human being and might be one of the worst people I have ever met. He would often belittle students for asking questions related to class material or about class policies. His assignments were based off what we learned in class with the majority of the difficulty coming from the insane amount of fluff put on top of them. Like he gave us games to encode a solver for and I must have spent 1% of the time on the solving algorithm (the important part of the assignment relevant to the class) and 99% of the time dealing with the absurdly detailed rules of the game that had no reason to exist beside to waste our times. Now back to the real problem Voris. Every time a student asked him for something, like if he could post the blank midterm online after we got it back so we can study it for the final, he would answer in the most demeaning way ever. Every question he felt was a personal attack on himself. Now I am by no means a psych major, but I would say he had a huge fear of being seen as inferior. He would oftentimes go on rants about how in his college, which is some local school that most people have never heard of, he had it so much harder and how Columbia students were entitled and a bunch of cheaters. This r…

Radev's a jovial guy and a decent teacher. However, I didn't learn much in this class since Radev focused a lot on useless topics. We literally spent 60% of the class studying logic of some sort. I would rather have learned more about search algorithms and machine learning. Radev coached the International Linguistics Olympiad team so he's a big fan of linguistics and it shows in his teaching. He spent a lot of time talking about gramamars and outlining NLP at a highly impractical level which I really hated. I think you can learn more about AI by taking a MOOC online from Coursera or Udacity. There are 4 programming projects which are all very time consuming. He gives you 2 weeks for each and they take at least 6-8 hrs. There's no upper limit since you can improve the assignments as much as you want in an effort to get the very generous extra credit (up to 20 points on a 100 point assignment that counts for 20% of your grade). Only 2 of the projects had anything to do with AI. The first one was to be coded in LISP which has the dumbest syntax ever - it's like a poor man's Haskell with a lot of parenthesis. All in all, can't recommend taking this class since it doesn't really cover much. Radev's a nice guy and probably the best teacher for AI you're going to find but the class itself is pointless. Tests are ok. If you read all the slides you should be fine. If you understand all…

A very good class and professor. Taking Artificial Intelligence with Professor Mckeown convinced me to pursue AI and computer science further. Very highly recommended.