Over the past year or so I’ve participated in a number of massive open online classes, or MOOCs as they’re coming to be known. The experiences have varied, but overall have been very positive and quite educational.
Introduction to Artificial Intelligence (pre-Udacity)
First I took the initial Artificial Intelligence class from Norvig and Thrun, which was the start of the current stage of online learning. This was my first class of any sort in many years, so it took a bit of getting used to as far as homework and .quizzes. I have a degree in Computer Science and casually studied AI back in the 80s and 90s, but I’ve never done anything significant with it beyond that. My college senior project had been to get radio controlled cars to cooperate to push a ball towards a goal. My project partner worked on the vision system and I worked on planning. So I had some practical experience in getting an AI system to actually work. But that was the last time I’d done anything AI-ish. Because of that, I went into the class thinking of it as a fresh start. As it turned out, it mostly was. The subject, at least as taught by Novig and Thrun, has moved to a much more statistical foundation over the years, which was different from the more logic-based approach I’d learned way back when. The class turned out to be a good overview of the basics of the current state of the field from two of the leading practitioners of the art.. The presentation was rough around the edges, the videos often crude and rudimentary. Sometimes it would be a closeup of a hand scribbling on paper. The quizzes, homeworks, and tests had minor issues here and there (primarily not considering all possible lines of thought), but the class itself did a good job of communicating the material. It was a pilot course and it showed it, but the quality of the material made up for the presentation.
Circuits And Electronics (MITx)
Next I enrolled for MIT’s 6.002x class, an introductory circuit analysis course. This course had much better presentation (They had the benefit of having seen the AI course, which had to have helped.) and was considerably more demanding than the AI course. The course was very heavy on mathematics, with a lot of emphasis on deriving proper equations for circuits and circuit elements. Very challenging approach to the material, but it paid off with a stronger understanding in the end. Working things out for yourself is a much more effective method of learning than just memorizing equations and approaches. I’ve struggled with understanding electromagnetism and electronics in the past, I’ve just had some sort of mental block on the material. It was quite satisfying to have to work hard at this class and achieve understanding as a result. Easily the hardest I’ve worked at something that was initially so frustrating in many years. From that viewpoint, it’s not at all an exaggeration to say this was a life-changing course. The instructor and the class brought out the desire to excel at difficult tasks that’s been missing in my life. So worth every hour of frustration along the path. Extremely happy with this course and the results.
I also took Coursera’s Introduction To Crypto course. Again, a subject where I had some self-taught familiarity but no formal instruction. Ultimately not too difficult as programmer comfortable with manipulating bits and bytes. The lectures were quite dense and fast-paced, to the point that several times I would have to rewind because my attention had wandered for 30 seconds and I know longer knew what he was talking about. Personally I see that as a plus- the course covered a lot of material in its short timespan. Not material I’ll have much cause to use in the end, as the course says repeatedly, you probably shouldn’t be implementing crypto systems yourself. The risk is high that you’ll make an error that won’t be apparent to you but provides an attack vector for those well-versed in the field. But the programming homework was good as a push to learn some basic python. And it certainly doesn’t hurt to have a medium-level understanding of crypto, how it works, and what some of the issues are.
Algorithms: Design And Analysis (Coursera)
Next was the Coursera Algorithms course, which I intended as a refresher. It worked well at that level, reminding me of material I’d seen in the past, along with some new concepts I only had vague ideas about. The instructor was excellent, with a real passion for the subject and an ability to explain the material succinctly and in a way that made ready sense. The course would be good for anyone with a CS background looking for a refresher. The programming assignments were relatively straight forward for basic solutions, but they also prompted me to consider how to make them more efficient in time and space. Worth the effort expended.
Listening To World Music (Coursera)
Next I started Coursera’s World Music class. I approached this class with some enthusiasm since world music has been an interest of mine since my days in record stores and college radio back in the 80s. Also, I was curious to see how they would manage the essay portion of the class. Every week required a few paragraphs on one of the topics that have been covered that week. The essays were judged by a peer-grading system, with each student expected to provide scoring on at least 3 submissions from other students. The essays were to be judged on the ideas presented as well as the mechanics of the writing. I had been concerned about how this would work in practice, but I felt my scores were fair for the most part, and reflected where my essays were flawed or were strong. I didn’t really enjoy grading other people’s essays, however. I didn’t mind providing feedback on the ideas presented, but judging the mechanics was often difficult, particularly when the author was clearly not native to English. In the end, I felt this aspect of the course required more energy than it paid back.
I’m not sure how that problem could be solved. I see the need for that sort of feedback, I just didn’t enjoy it personally and did not find it rewarding. The course itself seemed quite disorganized and in the end I lost interest and drifted away from the course. This was the first course I started but did not finish, which was an unpleasant feeling. I’m not completely sure why my interested faded as much as it did, nor if it was me or the course.
I have just started the Coursera Introduction To Financial Engineering class. This is a topic I’ve long been curious about, but I have next to no knowledge beyond general economics. The class has pre-reqs of calculus and statistics, as well as some basic programming knowledge, which is more than most online classes have required thus far. If nothing else, I should gain some experience using R. This could be a challenging class, we’ll see.
I also started the Udacity Differential Equations course as a refresher. It’s based on using simple Python code to solve and understand differential equations using numerical methods. The first 2 units are based on a simple version of calculating a rescue plan for an Apollo 13 like incident, so it’s pretty fun so far. The videos are *very* slick and the course is by far the most polished of any I’ve seen so far. The videos are a whole different level from anything I’ve seen in a MOOC. The class is also completely self-paced, which I think I will also enjoy.
Signed up for some future courses. Modern American Poetry at Coursera. That one might have the same essay issues as the World Music course, but the intro video from the instructor made me think he was quite comfortable with the format. I know little about the subject, so it should be interesting. Also signed up for Solid State Chemistry at EDx. This one looked like the most promising of the fall offerings, though none of them were the math or physics courses I really wanted. I’m not quite sure I can handle the course, but I’ll give it a try. Also added number of Astronomy courses at Coursera starting in the winter.
Astronomy/physics is my main area of interest interest for MOOCs. A lot of the other classes are marking time until more of those become available. I’ll take all the classes in those subjects that I can manage. EDx, at this moment, is my preferred home. If they have courses I want, all else being equal I’ll go with them. I very much like their philosophy of keeping the classes at a challenge level similar to the top universities. There’s a place for the Udacity philosophy of accessibility, and I can see making use of their classes for quick and easy coverage. And there’s a place for the somewhat simplified Coursera courses. But what I really want is the genuine challenge that the 6.002x pilot provided, and I hope EDx can live up to that with future classes.