edX - Berkeley CS188.1x: Artificial Intelligence Review

CS188.1x: Artificial Intelligence is an introductory AI course offered by UC Berkeley through the edX MOOC platform. CS188.1x covers roughly the first half of the material in the full on-campus AI course in the span of 12 weeks. Major course topics include search algorithms and heuristics, constraint satisfaction problems, Markov decision processes and reinforcement learning. The course assumes you have taken a first course in algorithms, are familiar with basic data structures, have basic python programming skills and are comfortable with mathematical notation. There isn't any particularly hairy math, but there are a lot of variables and symbols flying around at times. Grading is based on weekly homework assignments that allow unlimited attempts, 3 programming projects and a final exam that allows 1 or 2 attempts per question.


CS188.1x is a direct adaptation of the on-campus AI course. The lecture videos are edited versions of lectures delivered on-campus but instead of seeing the professor, we mostly see the presentation slides themselves with a voice-over from the professor. Direct adaptations of on-campus courses don't always work so well with MOOCs, but this course pulls it off perfectly. The professor speaks clearly and explains topics well. The lecture slides are extremely well-made, with clean text and even a bunch of cute robot and pacman art to go along with the content. The videos are cut down into digestible 5 to 15 minute segments and there are practice comprehension questions following most of the videos that allow you to take a second to reflect and digest the content.


Many courses that have great presentation fall flat when it comes to assignments. This is not one of those courses. The three pacman-themed programming projects are among the best programming assignments I've encountered in any online course. Each project consists of several parts that involve implementing AI algorithms you study in class in the context of a pacman game. The course provides you with all the code you need to run the game, a variety of convenience functions and skeleton code that you have to fill in with algorithms that accomplish the prescribed tasks. The assignments can be frustrating at times, but seeing your code in action with a little pacman racing around gobbling food pellets and ghosts is surprisingly gratifying. It also helps you gain a better understanding of how the algorithms work.


Berkeley CS188.1x: Artificial Intelligence is one of the best MOOCs on the web. It is so good that many students on the forums were eager to take part 2. Unfortunately the professors haven't gotten around to adapting the second half of the full AI course into a MOOC (they did express the desire to do so in the future) but they will give you access to an archived version of the full course upon request.


I give Berkeley CS188.1x 5 out of 5 stars: Excellent.