Friday, November 06, 2009


Good news, lads! Good news! The left hand side of the equation is our a posteriori estimate given the observations.

A frequent knock on scientists is that many of them can't write. Not even a little. In their wrestling matches with the English language, the latter often wins through a combination of astonishingly well-executed pretzel submission holds.

(This by no means includes those few scientists who can write clearly. If you're reading this and you think the above means you, of course it doesn't. You're mistaken, your prose always elegant and illuminating.)

On the other hand, one has the latest from the Jason group defense advisory panel.

The Jasons are famous.

"The elite JASON defense science advisory panel dismissed claims that high frequency gravitational waves (HFGW) could pose any kind of national security threat," wrote Steven Aftergood at the Federation of American Scientists not too long ago.

"The Pentagon should 'monitor enemy activities in sleep research' [said] a newly disclosed report (pdf) from the elite defense science advisory panel known as JASON," he informed even earlier in a post entitled, "Jasons warn of sleeping enemies."

"The JASONs were investigating the potential for U.S. adversaries 'to exploit advances in Human Performance Modification, and thus create a threat to national security.'"

Certainly, these are good things to know. And there can be no doubt that only Jason scientists could have delivered such information. That high frequency gravity waves pose no national security threat.

But in a recent paper, the Jasons turn their scalpel sharp skills to the problem of predicting things like catastrophic terror attacks.

"Imagine we have a discrete random variable with K possible values; as already mentioned one example might be the probability that a biological terror incident will originate from a group centered in country k, where k has a value between 1 and K," write the Jasons in a technical appendix on page 95 of "Rare Events."

"In particular there may be many bins that have n(subscript k) = 0, but it might be foolish to conclude that all these probabilities are actually zero if the number of data points is small," the Jasons add a little further down.

Yes, foolish.

"Attempts to predict the occurrence or the likelihood of extreme acts of terrorist violence on the scale of 9/11 should be discouraged because the available data are too sparse to permit the reliable modeling of such 'rare events,'" wrote Aftergood on Wednesday, in describing the alleged thrust of the new Jasons report. That summary can be found here in "Jason Cautions on Predicting Terrorist Events."

Over the years Aftergood has stoically provided a valuable service (among many) without complaint -- the reading and condensation of Jason reports into nuggets of easily comprehended wisdom. For this alone, he deserves a shout out.

"[The Jasons] cautioned that the complexity of the problem and the presumed urgency of the threat have 'led some to advocate the suspension of normal standards of scientific hypothesis testing, in order to press [predictive] models quickly into operational service,'" summed up Aftergood.

"Even in the original case of actual black swans, anyone sensible would have hesitated to extrapolate an observation of N white European swans to a prediction of zero black western Australian swans," write the Jasons on page 88 of "Rare Events," the analysis posted at FAS.

And in Africa no one goes into the jungle between the hours of 7 and 9 in the morning because that is when African elephants start jumping out of trees.

So do you know why there are pygmies?

There are pygmies because they went into the jungle between the hours of 7 and 9 in the morning.

Taxpayer-funded incomprehensibility contract carried out and implemented by the Mitre Corporation of McLean, Virginia.


Post a Comment

<< Home