Log in

Previous Entry | Next Entry

Here goes! Whee!

Zokutou word meterZokutou word meterZokutou word meter
741 / 50,000
(50,200 +) (1.5%)

741 words on Book 2. None too impressive so far, is it?

Today we had a network outage for the entire work day, from 08:00 or so to after 17:00. Apparently an air conditioning unit failed, leading to some overheating in a machine room that tripped a breaker in the basement, which in turn took down some switches and a firewall server. The end result was that we could connect to machines outside the department until someone from central Computing and Network Services fixed it. Details are to follow when I find out more.

tmehlinger and thekuffs seem to be going full steamish on massforge. Late this evening I got mail from dextertech saying he posted some new stuff, and indeed, it looks interesting.

Sanjoy Das and I had a talk about massforge and about some dynamic programming applications to inference (particularly value and policy iteration solutions to decision problems) and learning (particularly MDPs), graphical models, dynamic real-time search, simultaneous localization and mapping (SLAM).

The penultimate grad council meeting of the fall went swimmingly. masaga and pvntejawi: if you're reading this, the 2006 Grad Research Forum and Topeka Capital Research Forum will be announced online next week. There is a nominal amount of prize book money, and students will have the chance to go and present oral talks at a CS/Math/Physics forum session or a poster about their research. Please spread the word.

Today I had a biology faculty member ask for a bioinformatics algorithms course that did not require a first programming course. Now, I could understand this if it literally meant first programming course as opposed to first CS course, but no, the faculty member meant CS. As you might expect, this did not wash with the department head, and as I told my colleague, it just doesn't work that way.

That raises an interesting question for all you computational biology types, though: how much programming do you think someone has to have to understand protein folding codes, OpenRasMol, BLAST, etc.? Suppose for the sake of argument that the idea is to know "just" how to run such software and interpret the results. I have my opinions, but I'd like to hear yours first.



( 2 comments — Leave a comment )
Nov. 2nd, 2005 12:54 pm (UTC)
"That raises an interesting question for all you computational biology types, though: how much programming do you think someone has to have to understand protein folding codes, OpenRasMol, BLAST, etc.?... "

Well, to run all these things - not a heck of a lot. RasMol and Blast you just run. Protein folding is more complicated, but I suppose if said biologist understands the thermodynamics of the algorithm, actual understanding of the algorithm becomes less important.

But that never suited my taste. I always wanted to jump in there and learn what it was doing. I mean, like ClustalW and such, it helps to understand the interplay so you can adjust the gap penalties, and you can make sure that it's codon-aware.

At Case there are two into bioinformatics class. Gene 508, and EECS 458. 508 = intro to unix and some neat tools, EECS 458 is algorithms of bioinformatics.

That's how we partition that stuff here at CWRU.

Nov. 2nd, 2005 08:44 pm (UTC)
Legitimate grouse
I think, the Bio Prof's request to have a algorithmic less intensive course is a reasonable request. We need to encourage stronger participation from non-CS people. Having more folks with a biology background enhances the classroom experience significantly as compared to one dominated by CS types.

At SUNYSB, they pull out all stops to encourage Biology students to enroll for the grad Computational Biology course, many of whom do not possess any background in CS.

They shifted from a more algorithms intensive textbook like Gusfield to Jones and Pevzner, which provides a more accessible treatment without shying away from topics like HMMs, Clustering algorithms etc.,

They also have differential marking in exams for Biologists and Computer Science students. Also, much of the course credit is carried by the term project, where you're given a choice to do either a more Biology oriented project or a Computer Science intensive one or both. Assignments are formulated in a way that encourages both groups to collaborate.

I think a little "dumbing down" may not be so bad, considering the benefits of having a more heterogeneous student population. Ideas flow freely and you get to learn a lot more biology that way.
( 2 comments — Leave a comment )

Latest Month

December 2008

KSU Genetic and Evolutionary Computation (GEC) Lab



Science, Technology, Engineering, Math (STEM) Communities

Fresh Pages


Powered by LiveJournal.com
Designed by Naoto Kishi