Saturday, March 28, 2009

Debriefing on the NSF 'Future of Network Science' workshop

I got back late last night from a military-run, NSF-sponsored workshop on the future of network science. The goal of the workshop was to brainstorm and try to develop a roadmap for developing 'network science' as a discipline in its own right.

Sadly, the workshop failed. ("I don't think your plan was very well thought-out... Master.") The organizers' method of doing this was ill-conceived, because they seem to think that this can be done in a physics framework and were hoping that there might be some sort of 'periodic table' of networks. I think that's too narrow-minded and tries to put a square peg in a round hole. Also, almost none of the leaders in the fields (from any of the various groups going on---in physics, CS, social science, engineering, etc.) were there. Hell, originally the workshop was going to conflict either with Sunbelt (so that none of the social scientists would be able to go) or the APS March Meeting (so that none of the physicists would be able to go). As it turns out, almost none went anyway. The organizers should have gotten around 10-15 leaders from these various fields, made sure they could come to the workshop, and then organized it around their schedule. (The 10-15 might have come from a list of, say, 30 stalwarts.) Then whoever among the rank and file (i.e., people like me) could make it then would come up to the numbers they wanted. (The size of the workshop was about right.) Also, based on the introductory comments and buzzwords, the organizers' knowledge seemed to be circa 2003.

We had lots of arguments about what a network is, and some people annoyingly insisted that it must have some sort of "flow", which would have the unfortunate side effect that things traditionally studied by bio networks people and many of the situations in the social sciences that are rather hard (that people still struggle with, as most of the research in practice deals with gross simplifications of these things) for which many of the important concepts were developed to get a handle on would a priori not be considered networks. So is the stuff for which a lot of the stuff was traditionally invented supposed to be outside of the scope of what we want to develop? That didn't seem to bother them, but I call bullshit on that one! This wasn't adopted, but some people kept bringing this up over and over again and really impeded progress. Personally, I think the suggestion of following the the Supreme Court's definition of defining pornography would have been a good way to go. :) Hence, one of my take-home messages of the workshop: "Network science is like pornography. We'll know it when we see it."

There were a few "awesome" comments during the workshop (both about the science and about other things):

"Monies will get exchanged between your government organization and our government organization."

"I can say some things that will make it sound like a network."

"If you're going to dress me up with Barabasi-type stuff, I'm going to walk out." (This was stated by a very annoying person. I was extremely tempted to pretend to believe that stuff just to see if he'd leave.)

"I know there are neurons. I know there's a lot of them." (says the medical scientist sarcastically)

"Some parts of some things have some properties of networks [I think "some of the time" then followed, though I forgot to type that here originally]." (This was particularly well-phrased, and this is pretty much exactly how I'd characterize things.)

Larry Leibovich really kicked butt as one of the voices of reason at the workshop.

"This led to an interesting and long discussion, which did not resolve anything." (This could be used to describe most of the discussions at the conference, though only some of the discussions were actually interesting.)

The random well-dressed guy was here (and actually I expected him to be here after seeing him show up as a military person in the 'How Kevin Bacon Cured Cancer' video).

No comments: