Completely independent subproblems are delicious.

Oct 14, 2007 02:03

First off, apologies for yet another post. Studying and locking oneself away from human interaction apparently invites venting.  Second, apologies to people I used to talk to on AIM, for my complete silence lately.  I should be back to my usual level of internet communication with AIM and Skype and LJ in about a week.  I’ve mostly been offline because I’ve been busy, which was caused because I let myself get way behind in all of my classes, which led to the nice weekend-o’-cramming mentioned later in this post. But thanks to a paper due Friday and a midterm Tuesday, I will be caught up on just about everything by the end of the week.  I’m not sure how long this glorious state will last after that, but hopefully I can maintain it.

So, cramming.  I’ve never actually crammed before.  Generally my study strategy consists of ignoring the test until a couple days beforehand and then looking over what we’re supposed to know simply to remind myself of things and make sure everything is fresh in my mind.  In this case, however, I hadn’t read any of the textbook (normal) and I’d missed enough of the classes that even the notes I had didn’t make much sense (not normal).  So I devoted today to teaching myself the fundamentals of artificial intelligence.  Or rather, those fundamentals covered in chapters 3 through 5 of my textbook.  7 through 9 I’ve left for tomorrow, and Monday will be mostly spent going over the specific kinds of problems that should show up on the midterm.  Now, this seems like a task doomed to failure, but it’s actually been going fairly well.  It’s even been fun, in a mildly masochistic way.  This textbook is easily one of the most useful and easy-to-read textbooks I’ve ever had, hence the quote for this post, and the topic has always been one that’s interested me.  It was more than a little disconcerting to realize, halfway through learning about heuristic search strategies, that I could almost see myself being happy working with people and researching this kind of thing.  Still, I pledged my loyalty to English long before Computer Science, so it’s unlikely I’ll consider a non-literature-related field any time in the near future.

Another amusingly distressing effect of this cram session, (aside from causing a large handful of my tea bags and the entirety of memoria_teneo’s Dove chocolate to disappear), was the complete devotion of all but a few of my thought processes to ideas related to problem solving algorithms.  This, in turn, caused me to try to think of different ways to represent the relation between season and weather, (“Don’t Fear the Reaper” came up on my playlist and it mentioned seasons. This is my only explanation.), and later to consider an algorithm that would sort through appropriate responses to cuteness, taking into account situation, people involved, and various social and personal constraints, (I can blame zapf for that one.)

If it were even possible for me to write algorithms that produced appropriate reactions in social situations, though, (and I am fully aware that it would be well beyond me), I’d mostly use them to help me figure out how to deal with people.  I think I’d start with not talking to myself in public. And perhaps not standing in Usdan, holding an AI textbook, listening to Once More With Feeling, and talking about fighting beholders in DnD. *cough* Not that I did that today, of course.  At the same time that I was being beyond geeky, however, I realized that my mildly awkward feelings toward a person I’ve been interacting with recently were a direct result of a Rocky Horror cast party last year.  I had to devote some of my attention to non-computer science tasks while studying, lest I go insane, but it seems far too ironic that my thoughts should polarize on socially awkward geekery and attractive, mostly naked people.  Ah well.

computer science, cramming, ai, strange thought patterns, studying, social awkwardness, geekery, rocky horror, stress

Previous post Next post
Up