“Got kitesurfing on the mind, mixed with some search & classification tech, and a dab of political ranting”

Oh-oh NSA Data Miners in Demand

Posted by direwolff on January 25, 2007

Just caught this link courtesy of Bruce Schneier’s blog post titled “NSA Hiring Data Miners“.  I’ve railed about this before, and I’ll rail about it again, predictive or user analysis systems are nice for commerical uses where they can afford to get things wrong without adversely affecting the rest of any person’s life, but in government systems this is pure lunacy.  If a contextual or behavioral or collaborative filtering system recommends the wrong thing or groups me inappropriately I may have to deal with some mistargeted ads or other slight inconveniences online.  But if a government system marks me as a terrorist or a danger to society as a result of being off by one action I may have taken at some point in my life totally unrelated to who I am or ever was, then that’s a HUGE problem that they’ve never been prepared to fix.

Companies like Autonomy use in part Bayesian statistical models to determine the categorization of various documents.  Basically, they get the users of the system to feed it anywhere from 10 to 20 samples of documents that should belong to a particular category to train the system, and from there *magic*, the system is able to identify new documents that should be categorized accordingly.  Nice in principle, but in practice such systems struggle reaching a 70% accuracy rate.  The problem gets worse because no human can see how or why a new document failed to get categorized appropriately.  The only way to deal with it is to submit the new document to train the system that it should recognize documents like this new one as being part of that category.  I’m only using this as an example of some of the statistical methods being employed for various tasks including some of the ones that some of our government intelligence departments are using for some of their citizen/terrorist data mining activities.

While I’m not saying that there aren’t some very good advancements in these techniques,  the NSA wiretapping fiasco, the recently departed “No-Fly List”, and a host of other invasive programs from the NSA, the TSA, the CIA, and the Department of Defense (on its own citizens), puts us in the awkward position of not being able to trust our own government to deploy something here with the best interest of its citizens in mind.

One Response to “Oh-oh NSA Data Miners in Demand”

  1. No way……. I’m all for it it…. I can’t wait to sue …. I want my $10M ;)

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: