Justin Li


Updates

2010-04-28

I clearly haven't written here for a while. I have a post halfway written about an adventure on Craigslist (no, it's not from the m2m personals...), which I will finish eventually. For now, here's part of an email I sent to my family.

I just got done with the semester over the weekend. If you'll remember, I'm taking two classes:
- Advanced AI. Like I said last time, this course is oriented towards getting students to read current research and to think critically about it. I think the course did a good job of that - we had to read a large number of papers, all over the map in terms of area. It's interesting too to see how different students react to ideas differently, as well as what areas of AI they are interested in. The final project for the course is to reimplement a study, and see whether their claims hold up. In retrospect, I should have chosen a different paper. Mine was on a relatively simple method of understanding how the world around you changes due to your actions. Although it was an interesting paper, the algorithms and methods were not very well specified, which makes it hard to test. It was also short with very few references, so there's no extra material I could refer to either. I finished it, but I would have enjoyed it more if the paper was more substantial. - Programming Languages turned out to be more fun than I thought. The basic premise of the course is how to prevent programmers from doing stupid things, and how the computer can automatically detect these errors and notify the coder. We also read a lot of papers here, although a lot of them were older papers. The areas were also more biased towards people without much background, so although we looked at four different methods, we only studied one or two in depth. The final project here was up to us, and I decided to (again) reimplement a paper, this time checking for errors in their presentation. I discovered a few of them, as well as a general problem with the approach it was taking. Because this work is actually someone's thesis back in 2006, everything was documented properly, and I have very little trouble following their work. All this means that I should make my code public when I publish a paper.

It was because both of these projects were due last week that I didn't have time to write this email. I was awake both Monday night and Thursday night writing reports, and so had very little sleep. It was fun though; I felt invigorated by how much work I was doing. There's a quote from Ayn Rand's The Fountainhead (Dominique, for those curious), when a character says, "This is how I want to keep going - if there's a reason for it." That was how I felt on morning after staying up - that having a reason to drive myself that hard is the best feeling.

I'm also done with my teaching duties for now. The final exam for that class was on Friday, so we graded everything by Saturday evening. My discussion sections and the class in general is smaller this semester than the last. I got to know the students a little better, but not in pro[portion to the decrease in size. I did realize something though - that I should spend the first few session actively encouraging the students to talk to each other. It's really annoying when you ask a question and no one wants to answer it. So the discussions were not as lively, there were less jokes, and overall it was a little less enjoyable.

Oh, I got an award for being a good TA last semester, which is nice. It also came with an extra check, which is also nice.

On the research front... I'm really digging in now. For a short while back in February I was really uncertain how everything would turn out. That period of doubt clarified my feelings a little - I'm sure now that teaching would have a part in my future, more so than research. On the other hand, very soon after that, I found a topic that I'm pretty interested in, so much that I spent all of this morning reading. We'll see how it turns out. I got put on the spot today to give a 10 minute presentation at a workshop in three weeks time, which means I really have to get my thoughts together and present a coherent picture of what's going on in the field, and what I want to do.

comments powered by Disqus