Using Spammers To Advance Artificial Intelligence

CAPTCHAs are a ubiquitous feature of the internet. Many websites use then to distinguish between real human beings and software attempting to add content to the site. Usually the real human beings are the ones you want to permit, while the software is attempting to add spam to the site, so it's a useful problem to solve. Essentially, a CAPTCHA is a mini Turing Test to try to distinguish between humans and machines.

I came across an article pointing out that we could harness spammers to help solve artificial intelligence problems to advance the field of research. Essentially, spammers have a lot of financial incentive to solve the problems. When they solve these problems, we simply present them with new unsolved problems.

This situation is hilariously clever. It's also the natural progression of the CAPTCHA arms race. So, the next time you get annoyed by spammers on websites that are protected by CAPTCHAs, think about the spammer's contribution to computer science!

The Sokal Affair

Merry Christmas! The holidays aren't a big deal to me, but time off from work is always nice. The holidays make me miss my sister, though.

But depressing thoughts aside, I stumbled across this interesting incident where a physicist got a joke article published as an experiment to see if the journal would do it. It makes me feel a bit better than those pretentious, convoluted articles really are what they seem to be sometimes.

On top of that, recently, some folks used a computer science paper generator called SCIgen to get a paper into the IEEE database.

Edsger Dijkstra

I recently stumbled across this archive of Edsger Dijkstra's work. This guy is a legend* of computer science. He was so adamant that computer science should be taught as something independent of actual computers that he himself never owned a computer (the site contains a number of pdf manuscripts... literally hand written). This is one of his works in which he discusses why the common approach to teaching computer science is bad. It might not be an appealing read for non computer scientists, but as someone who is interested in teaching (possibly computer science), I found it to be very interesting.

*Dijkstra's Algorithm (a core approach to finding the shortest route between two points on a graph), and a letter condemning GOTO statements are known to the vast majority of current computer scientists.