Feed on

gawsh, i’m starting to appreciate how difficult it is to develop algorithms. it seems that the function representing the novelty & utility of an algorithm vs. effort exerted looks like 1-exp(-x): the first couple of breakthroughs in a field come easily, but improving on them can be quite hard. in my case, i’ve been working a lot lately on how to go about inferring evolutionary events and phylogenies from sequence data. initial methods for phylogeny learning (like parsimony and distance matrices), at least in hindsight, appear relatively obvious but still completely changed the way people thought about evolution [this is the initial, sharply sloped part of the curve]. a little more work (maximum likelihood and bayesian approaches) yielded nearly as large gains [the function's slope is starting to decrease]. i feel like i’m in the terminal end of the curve, where a huge amount of effort may yield only marginal benefits.

that is, if i make any gains at all. this is what i think makes grad school much harder than med school: going into grad school, you have no idea how in the world you’re going to get yourself out. at least in med school, you know that if you can sit through a couple more exams, you get to add to your parchment collection. grad school possesses all of these wonderful uncertainties, like: what happens if your experiments don’t work? or your ideas are all wrong? or as i’m beginning to fear right now, you simply lack the mental chops to solve the problems you’ve bitten off? no lambskin for you.

bah, less than two months into my doctoral work and i’m already stressing myself out :)

i’ve decided to post some photos i took over the past weekend, to take my mind off of the sub-problem that’s triggered all this angst. (in case you’re curious, the problem is: given a phylogenetic history of a gene – the gene tree – and the phylogenetic history of several species – the species tree – how do you efficiently enumerate the minimum number of gene duplications and losses in the species tree that produces the given gene tree. the problem’s actually been solved before … no one explains it well, however, and i feel forced to reinvent the wheel).

in any case, christina and i met up with my good friend andrew, his siblings, and another on sunday out on the far eastern tip of cape cod. we took a bike trip from a town named brewster to a tidy little place called nauset beach.

throughout the day, we did new englandy things like see a lighthouse:

play on the beach:

swim in 50 degree (fahrenheit!) water [photo credit: andrew and his un-watersealed and anti-anti-fogging camera enclosure]:

and frolic in provincetown, which undoubtedly boasts one of the world’s highest drag queen-per-capita ratios:

and, of course, here are the obligatory artsy photos i took:

oh, who am i kidding. i try and make all of the photos i take artsy. some just suck less than others.

the rest of the cape cod photos cower here.

in any case, the beautiful weather on the cape provided conclusive proof that research & productivity are inversely proportional to how sunny it is outside. christina and i couldn’t get enough of being outside that day.

ok, missioned accomplished – a bit less stressed and ready for bed!

Bookmark and Share

2 Responses to “algorithms, phds, and cape cod”

  1. on 01 Jun 2006 at 11:24 am Anonymous

    hey nice photos! I didn’t know you would swim your nice digital SLR out to the ocean to snap that great photo of you and Andrew… *cough* plagerizer *cough* :) :)

  2. on 01 Jun 2006 at 11:31 am Lawrence David

    haha, photo credit duly inserted!

Trackback URI | Comments RSS

Leave a Reply