Thursday, April 26, 2007

Modeling Mess-opotamia

There's a story in the new issue of Science called "Pentagon Asks Academics for Help in Understanding Its Enemies" by Yudhijit Bhattacharjee about how "a new program at the U.S. Department of Defense would support research on how local populations behave in a war zone"

I like the reaction of one critic:

The plan has drawn mixed reactions from defense experts. "They are smoking something they shouldn't be," says Paul Van Riper, a retired lieutenant general who served as director of intelligence for the U.S. Army in the mid-1990s. Human systems are far too complex to be modeled, he says: "Only those who don't know how the real world works will be suckers for this stuff."

I tend to agree with this general... and his not so subtle stereotyping of 'academic types' is funny... but then again any (non violent) thing that will help the situation even a little bit should be considered. I just share the general's skepticism of modeling.

at the end of the story:

Accomplishing those goals is a tall order, Page admits. "Despite tons and tons of data from U.S. elections," he says, "we are still not very good at predicting how people will vote."

Building comprehensive and realistic models of societies is a challenge that will require enormous amounts of empirical data, says GMU's Levis, a former chief scientist of the U.S. Air Force. But it is doable, he says, adding that the field will benefit greatly from linking social science researchers and computer scientists. "The goal here is to win popular support in the conflict zone," he says.

No comments:

Locations of visitors to this page