we're going to do it with computers

[alexr_rwx]

Talking with Mike today, I brought up some of my frustrations with the CS program here, wondering whether I should switch to Cognitive Science or something else.

In a lot of ways, cogsci sounds kind of like what I want to do, and as I understand it, the program here is quite good. Discussing with Mike today, though, we uncovered two objections to my switching over: (a) they only have so much funding, and probably won't admit many students in the near future, and (b) to be in the cogsci program, it's pretty important that you care about cognition.

He said, and this made me feel really good: he likes that my goal is building software, because that's what he's out to do now too. Questions of psychological plausibility are only so interesting. (and they are kind of interesting.) But we're not going to do machine translation with brains, we're going to do it with computers, and hopefully it'll be useful to people.

(the downside, of course: I may have to take some dumb CS distribution requirement courses, which I feel like I shouldn't have to take because, y'know, I've already got an MS in this, and it's senselessly difficult to get my classes from gatech to transfer here.)

Comments

[lyceum-arabica.livejournal.com]

on the transferring of classes... this is a pretty small-town department, there might be more than one way to do something. if you're trying to transfer a class, you could talk to the profs that teach the IU version, maybe even ask to take last semester's final or something, and demonstrate that you not only took a course at GA with the relevant title, you also know everything from the IU course (which for a number of cases can be accomplished with a weekend of studying).

[oniugnip.livejournal.com]

That's a really good idea. I'd really rather not take theory again, but I bet I could do alright the 501 exam, with a little bit of brushing up...

(although I've heard that Leivant uses some pretty unusual notation -- t/f?)

[lyceum-arabica.livejournal.com]

true... but he also uses it on the written theory quals, which are all posted online, and in his class notes (which lindsey or someone else may have... or even the website, I don't think that's protected) Shouldn't be too big a deal to review.

[sstrickl.livejournal.com]

I sympathize. I had similar problems with transferring any credit from GT to apply to the lower levels of the PhD program at NU. Once I realized, at best, I'd be able to skip just one class, and that would be either Theory or Algorithms—classes I probably could have done with a refresher in—I decided to just do it all over again.

[oniugnip.livejournal.com]

Yeesh! Do you have to take classes anymore? (or do you, take them, if there's something interesting?)

I've actually got two classes transferred so far! Thad's class and Ron Ferguson's inference engine class... we'll see what else I can wrangle.

[sstrickl.livejournal.com]

No, I finished my class requirements in spring 08, so I don't need to take classes now. However, I am taking the History of Programming Languages class Matthias is teaching this semester, as the last time he taught it was spring 2004, and thus I missed it by that much :)

[lindseykuper]

Schools like IU (and maybe NU too?) don't seem particularly well equipped to deal with people who come in with an MS already. I'm also wondering if it's a particular kind of person who tends to switch schools after the MS, and a particular kind of school who attracts that kind of person, or if it's usually just individual circumstance. For me, I think switching partway through a Ph.D. would be a really bad idea. It would destroy my momentum.

[sstrickl.livejournal.com]

In my case, I was in the MS program at GT, not a PhD program, so... it was basically my way of taking a couple of years to decide whether research was really what I wanted to do, and really, it didn't help as much on that score as my later industry stretch. It was really just a way of putting off the decision about industry or PhD in a fit of procrastination :)

[winterkoninkje.livejournal.com]

If you're thinking of jumping ship, it seems like Informatics or (possibly) Complex Systems might be a better fit for you than Cog Sci... FWIW.

[winterkoninkje.livejournal.com]

Also, like I was saying to Lindsey, it may be easier to try to get the courses waived rather than trying to get credit transfered. Many departments are finicky about credits, even when they're relaxed about whether you actually take the specific "required" courses. I'm sure there are some interesting systems, theory, etc courses which would teach you something not covered by the MS; try to talk them into letting you take those in lieu of the reqs.

[oniugnip.livejournal.com]

We discussed that too, actually! I think Informatics sounds a lot like me, but they're not doing much NLP over there.

I think Fil's work is really interesting, though.