academic cultures

[gusl]

I'm soon going to start my PhD in Statistics, but I usually still think of myself as a machine learning person (or as a computational statistician). My department (Columbia) is a modern and open-minded one, and has hired faculty whose PhDs are not in Statistics.

But it seems like, at some point, I will have to make a decision about what research community I want to belong to. Perhaps most importantly, whether to publish in conferences (the CS model), or only in journals (the Stats model). What should I be taking into account when making this decision? Should I simply follow the faculty who I most want to work with?

Comments

[altamira16.livejournal.com]

In EE, submitting 4 page papers to conferences was a little step in submitting longer papers to journals.

[chrisamaphone]

i find it odd to think about these as decisions you have to make. it seems like these should emerge from the kind of research you do, and that in turn will emerge from who you pick as an advisor.

maybe i'm not getting something here. it sounds like you're expecting a lot more autonomy than most grad students i know have, though.

[gustavolacerda.livejournal.com]

At Columbia Stats (and perhaps most Math & Stats program), PhD students seem to be freer than in CS programs. This may be a consequence of the funding model: our funding is guaranteed for 5 years, and is not tied to research grants.

[gustavolacerda.livejournal.com]

This question was put in my mind by a faculty member there.

[chrisamaphone]

in cmu cs, our funding is also "guaranteed" (in the sense that they don't involve us in the process of securing it or tell us where it comes from unless we ask) for as long as we make sufficient progress. i don't see how making you worry about the publication model makes you more "free"; that's something i'm happy to let my advisor decide.

[gustavolacerda.livejournal.com]

Regardless, I think you would find it more difficult to switch supervisors to a faculty member who had no grant money.

Most students in Stats don't think about the publication model, because they are not considering faculty jobs in machine learning and assume that they will be in Stats forever. (I do not know where I will want to be)

awareness is the first step

[fancybred.livejournal.com]

But it seems like, at some point, I will have to make a decision about what research community I want to belong to.

I think that it's good that you're aware of this. But I don't know if there's anything more concrete you should be doing at this point in time, except to continue to stay informed about the workings and research directions of both communities (which you can do even if your advisor does not).

[wjl.livejournal.com]

Regardless, I think you would find it more difficult to switch supervisors to a faculty member who had no grant money.

This is actually one of the things CMU CS prides itself on -- you work with whomever you work with best, essentially, and it shouldn't be tied to funding.

[gustavolacerda.livejournal.com]

Sure, perhaps CMU is among the most egalitarian among CS departments, (i.e. faculty with little grant money don't find themselves unable to get students; faculty with lots of grant money pay a big cut).

But there is a constraint here, which you might be forgetting: grant money must be spent, and (in many cases) deliverables must be delivered.

(As a result of this), departments have to make sure that there is a correlation among faculty members between number of students (including postdocs and other staff) and grant money.

Such a correlation can be explained (or implemented) by:
* free students having similar taste to NSF folks
* decision of whom to admit: better-funded subfields admit more students
* pressure to work for grants that need to be filled, after admissions

I imagine all 3 apply in all CS departments in the USA.

I do not have evidence to back up any of this. I'm only speaking from what I think is common sense.

Re: awareness is the first step

[gustavolacerda.livejournal.com]

yeah.

[afanesvoltaje.livejournal.com]

Watching my husband apply for jobs while trying to straddle two fields, I think you're very right to be thinking about this now.

Perhaps it would be worth talking with people familiar with hiring trends in each area and asking about how publications via the other model are seen?