semester endgame
May. 5th, 2012 12:52 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
I meant to finish my phd qualifiers this semester. How this works here is: you pick a committee of three professors, and you and the committee pick three problems. Then you do the problems, write a paper for each one of them, and present it all. It's "supposed" to be done over a summer, and one where you're not doing anything else. But I keep having these great internships and getting married to
lindseykuper so I haven't had a free summer yet...
Due to a combination of perfectionism-ing my first problem, trying to do other stuff this semester, and plain time-mismanagement that was my fault, I didn't finish all three problems. I had explained my situation to my committee, though, and they were OK with me presenting just the first two, during finals week.
Then I proceed to freak out about a particular aspect of the second project, and I got way too fixated on it. I was trying to make a particular machine learning library work that just wasn't going to work (too inefficient to train in a reasonable time), and I spent like a week and a half on absurd performance hacks, at the cost of accuracy. And then I ended up missing our agreed-upon deadline for the second paper. I got an email from Mike about how it's too late to put in second paper, and maybe I should just present the first project; he didn't sound too mad, but I felt totally defeated, depressed, stupid, awful and worthless.
This was on Sunday (29 April) at about noon. Presentation is on Monday, of course. And I hadn't slept much the previous two days, of course. I said to myself that I was going to have enough to *present* about the second project, even if I wasn't able to give the second paper to my committee. And I found the parts of the project that I had already done, and they were OK. Then I found that there was another approach to do basically the same thing that I had been trying to do, except that it works. (specifically: LibLinear, which ships with MaltParser, is beautiful and fast and rad at classification, whereas Weka's Logistic classifier is extremely slow and sucky)
So of course I was up all night running experiments and making slides. But then on Monday morning, Lindsey brought me to the bagel shop to pick up Presentation Bagels, then to campus, and I gave my presentation. It was pretty OK! (slides here, if you're interested!) My committee seemed fairly impressed, gave me some pointers for how best to finish up the second project, and gave me an "over the summer" timeline for the third one. Notionally the third one should not be so hard, and anyway, we don't have to plan a wedding.
When I woke up on Tuesday, I remembered that I had told Lindsey I'd get the car fixed, since it's been making a weird scraping noise and she was supposed to go to Chicago to get her passport fixed on Wednesday. Oh snap! So I called around and found a car repair place that could look at it that afternoon. I also remembered that I had a meeting with Fil to do a conference call with another research group at 2:30.
So I'm at the car repair place, and they think they have a handle on what's wrong with the car, and it's like 2pm. But car repair places can drive you to where you need to go, right? Except their driver is off somewhere else. And then it's 2:08, and I'm running from the car repair place back to the house, in my clothes, with a backpack. I get to the house at about 2:16, and then I'm out the door by 2:17, on my bicycle. 2:26pm, I am in the office, and soaked in sweat. Conference call success; Fil didn't ask about it.
One of these years maybe we/I'll learn to not do things at the last possible goddamn instant.
(epilogue: the car is fixed, Lindsey got her passport squared away, and we're about to head to Toronto to race!!)
![[livejournal.com profile]](https://www.dreamwidth.org/img/external/lj-userinfo.gif)
Due to a combination of perfectionism-ing my first problem, trying to do other stuff this semester, and plain time-mismanagement that was my fault, I didn't finish all three problems. I had explained my situation to my committee, though, and they were OK with me presenting just the first two, during finals week.
Then I proceed to freak out about a particular aspect of the second project, and I got way too fixated on it. I was trying to make a particular machine learning library work that just wasn't going to work (too inefficient to train in a reasonable time), and I spent like a week and a half on absurd performance hacks, at the cost of accuracy. And then I ended up missing our agreed-upon deadline for the second paper. I got an email from Mike about how it's too late to put in second paper, and maybe I should just present the first project; he didn't sound too mad, but I felt totally defeated, depressed, stupid, awful and worthless.
This was on Sunday (29 April) at about noon. Presentation is on Monday, of course. And I hadn't slept much the previous two days, of course. I said to myself that I was going to have enough to *present* about the second project, even if I wasn't able to give the second paper to my committee. And I found the parts of the project that I had already done, and they were OK. Then I found that there was another approach to do basically the same thing that I had been trying to do, except that it works. (specifically: LibLinear, which ships with MaltParser, is beautiful and fast and rad at classification, whereas Weka's Logistic classifier is extremely slow and sucky)
So of course I was up all night running experiments and making slides. But then on Monday morning, Lindsey brought me to the bagel shop to pick up Presentation Bagels, then to campus, and I gave my presentation. It was pretty OK! (slides here, if you're interested!) My committee seemed fairly impressed, gave me some pointers for how best to finish up the second project, and gave me an "over the summer" timeline for the third one. Notionally the third one should not be so hard, and anyway, we don't have to plan a wedding.
When I woke up on Tuesday, I remembered that I had told Lindsey I'd get the car fixed, since it's been making a weird scraping noise and she was supposed to go to Chicago to get her passport fixed on Wednesday. Oh snap! So I called around and found a car repair place that could look at it that afternoon. I also remembered that I had a meeting with Fil to do a conference call with another research group at 2:30.
So I'm at the car repair place, and they think they have a handle on what's wrong with the car, and it's like 2pm. But car repair places can drive you to where you need to go, right? Except their driver is off somewhere else. And then it's 2:08, and I'm running from the car repair place back to the house, in my clothes, with a backpack. I get to the house at about 2:16, and then I'm out the door by 2:17, on my bicycle. 2:26pm, I am in the office, and soaked in sweat. Conference call success; Fil didn't ask about it.
One of these years maybe we/I'll learn to not do things at the last possible goddamn instant.
(epilogue: the car is fixed, Lindsey got her passport squared away, and we're about to head to Toronto to race!!)
no subject
Date: 2012-05-05 05:05 am (UTC)Uh, maybe we should go to bed or whatever?
no subject
Date: 2012-05-05 02:24 pm (UTC)no subject
Date: 2012-05-06 10:36 am (UTC)