Here’s a little piece of unsolicited advice from me to you:
When you’re living on the third floor of a 3-floor apartment building, and therefore keeping your heating turned off to save on money (the heat rising from below is sustainable), shut the bathroom door while you’re taking a shower. It will save you from a very, very painful exit from the shower when you’re finished. I learned this about 10 minutes ago. Yowza.
At any rate, allow me the privilege of inviting everyone to a vicarious journey (a la, me) into the life of a computational biology graduate student, late February of his first year.
I’m living in Shadyside, roughly 30 minutes’ walk door to door (25 if I hurry). There are numerous coffee shops within a 10-minute radius that frequently see my business, as the apartment itself isn’t overly conducive to productivity. Something about the “bed” and “couch” and “overpowered desktop computer” that are insanely distracting…
[ASIDE: On the note of the desktop, you may recall from a previous post that I was having processor issues. Intel was extremely helpful in this regard, prepping a replacement under the 3-year warranty, which will ship once they receive my defective unit. This process will take at least a few weeks, so for the time being, Ronon (my desktop’s designation) is out of commission, and V2 (my laptop’s designation) is taking the helm.]
I am enrolled in three classes this semester: 03-710 Computational Biology (henceforth referred to here as “710”), 03-711 Computational Molecular Biology and Genomics (aka “711”), and 10-601 Machine Learning (duly dubbed “601”). Each course has two 90-minute lectures each week, and two of the three courses have one 50-minute recitation each week. Doing the math, that’s 11 hours of class each week. I squeeze daily workouts into the mornings, leaving the afternoons free to conduct my studies.
Clear as mud? Excellent!
Excluding this week only, I spent the previous three weeks coding Java and C++ programs for 601 and 710, respectively. The former was for our third homework assignment, a detailed exploration of Logistic Regression for classifying variables with continuous-valued properties (as opposed to discrete properties that take on one of a few known values), as compared to the Naive Bayes approach (LR is more powerful but more complex; NB is less complex, but may sacrifice accuracy). The latter was an incredibly intricate implementation of generating hidden markov models from known protein or DNA sequences in order to classify unknown sequences.
Both these assignments were completed and submitted late last week, and other than the weekly Wednesday quiz in 711, this week invovled no deadlines. As any graduate student should be well aware, a lack of deadlines now can only equate to a deluge of deadlines later.
Which, of course, I discovered not-so-gently on Monday. The exchange went something like this:
[in 711 lecture]
Professor: “When does everyone want to have the midterm next week?”
Professor: “Monday or Wednesday?”
Me: “Um, Wednesday please.” (I want more study time)
Fellow Student: “Dude, we have a Machine Learning midterm on Wednesday.”
Me: “…WUT. Ok, I change my vote to Monday.”
Me: “Crap, we have two midterms next week?”
Fellow Student: “And a 710 midterm on Thursday.”
That brings me to the present hour. Let me go over in detail what exactly my to-do list entails for the next week.
In 601: Midterm next Wednesday, but not mentioned thus far is the fourth homework assignment that is due on Monday. This assignment covers Bayesian Networks – inference, D-separation, as well as structured learning and parameter estimation. Once that is turned in, we’ll have a review session covering everything we’ve discussed so far, from mutual information and decision trees to naive bayes and logistic regression to maximum likelihood estimation and bayesian networks. The midterm is open book and open notes, but given the exam is 90 minutes and covers a vast repertoire of information, it would academic suicide to believe one could flip through notes looking for answers during that time.
In 710: This is what I would consider my most solid position (for the moment, anyway). I have done very well with the programming assignments thus far as well as the quizzes. Nevertheless, I am quite sure that the exam will take every opportunity to delve more deeply into the theory behind markov models and dynamic programming than any homework or quiz has thus far. Particularly on HMMs, we got off easy in terms of the probabilistic theories that power the concept. Hopefully, studying will mostly consist of brushing up on weak spots. No news yet as to whether this exam is open or closed book.
In 711: I don’t know what to expect here. It probably has somewhat to do with the fact that this is my professor’s first course since receiving her PhD from Stanford, and while obviously brilliant, she has tended to rely on us for feedback on the progression of the course; she’s learning as much as we are. We’ve had weekly quizzes which, I feel, have also been less intense on the theory than they could have been. Thankfully, there has been a ton of overlap with this course and my other two, so similar concepts like Bayesian Networks, Hidden Markov Models, Expectation-Maximization, and Dynamic Programming will be reinforced, allowing me to focus more exclusively on sequencing technologies, Hardy-Weinburg equilibrium theories, linkage disequilibrium, and haplotype inference.
What to do?
Study, of course! Such a silly question 😛
I am off to go complete a write-up for 711 notes I took last week (it is a course requirement that each student types up two lectures’ worth of notes to be posted on the website) before my 710 lecture begins for today. After that, I have my weekly 601 recitation, which will be followed by a lovely dinner of homemade chili (made it last night from scratch, and it was DELICIOUS) while foraying into 601. Then tomorrow begins anew with 710 recitation in tandem with lots and lots of additional studying.
What’s that? It’s going to be Friday tomorrow, you say? What about going out, you say? First let me remind you, in case you haven’t figured it out by now (or are new to my blog) – I’m not exactly what you’d call a “stereotypical student”. Hardcore partying never has been – and I seriously doubt ever will be – part of who I am. But that little detail aside, I’m a graduate student, and Fridays are just another day in the 8-days-a-week graduate student work week.
And in the meantime, I invite you all to watch this 40-second video of a cat getting pwned by a rabbit. It reminds me a lot of my own cat (a black and multi-digited bundle of love named Salem) who likes to take up residence in the crate of one of my dogs (a charming and adorable nuisance of a chocolate lab named Darby). Darby, unlike the rabbit in this video, will never force Salem out, and will in fact scootch to the farthest corner of the crate so Salem has room to stretch out. Though the pictures I have of the two of them sleeping in the crate are truly priceless. 🙂