Hulver's site || Because we're afraid to quit our jobs.

Diaries
Everything
Who's Online?
FAQ / Help

bleh

By gzt (Fri Apr 18, 2014 at 02:51:07 PM EST) gzt, work, done, near future, singing, choir, church, sing faster (all tags)

"sing faster!"

CHURCH: So I still have a bit of a cold. My throat isn't sore and I can breathe through my nose, but I'm a little phlegmy and my breathing is a little shallow, so while I can still relax and sing, sometimes the notes don't go where I want them to (at the low end, sometimes anything between Bb and D was just the same note by the end of the evening, unfortunately) and I have to breathe more frequently. by the end last night, i was a little rough. I'm the only bass, too.

Ugh, much handwaving when it comes to this problem set I'm supposed to finish. The CLT can be tricky.

I need to get a bit of work done in the near future.

< Always use utf-8 | Not that bad. >

Login

Related Links

Hit Counter

48 Husi users have viewed this.

Comment Controls

bleh | 3 comments (3 topical, 0 hidden)

yeah, so some statistics graduate student came up by the mariner (4.00 / 1) #1 Fri Apr 18, 2014 at 05:13:06 PM EST

to me the other day and started asking me some question about minimax-ish problem involving matrices and shit. and she was calling the image of the linear map defined by a matrix its "column space" and the kernel its "null space" and i was all like, omg, how can you get anything done thinking about things this way?

then i was like, dude, your problem reduces to optimization on a simple, compact subset of RR^n. the solution is guaranteed to exist and you can find it by standard methods of multivariable calculus.

then i was like, i can't believe i actually knew the answer to a question about something from statistics.

argle bargle by gzt (2.00 / 0) #2 Fri Apr 18, 2014 at 06:21:21 PM EST

it really is annoying that, as much as they make statisticians do stuff with "linear models" and optimization problems, they don't make people take courses in numerical analysis or numeric linear algebra (or, often, any real linear algebra).

you just described, like, a good percentage of problems statisticians have to solve. Sometimes they're not that "nice", but often they are.

[ Parent ]

: yeah. by the mariner (2.00 / 0) #3 Fri Apr 18, 2014 at 08:21:44 PM EST

her question wasn't numerical though, she was really asking an existence question. she managed to confuse herself into thinking that the problem wouldn't have a solution unless it was in the intersection of the "column spaces" of two matrices (because that was the situation in which she thought she could solve it by fiddling with matrices).
anyway, it made me feel that all the time i've spent learning/doing real math wasn't a total waste, since i'm still better at doing "real life" problems than a lot of people working in more practical fields and the real math actually helps in thinking about "real life," in those rare situations that i bother to think about it.

[ Parent ]

bleh | 3 comments (3 topical, 0 hidden)