Join 3,521 readers in helping fund MetaFilter (Hide)


Clever /monkey/human/robot/alien
January 28, 2011 12:45 PM   Subscribe

'On the Hunt for Universal Intelligence' 'How do you use a scientific method to measure the intelligence of a human being, an animal, a machine or an extra-terrestrial? So far this has not been possible, but a team of Spanish and Australian researchers have taken a first step towards this by presenting the foundations to be used as a basis for this method in the journal Artificial Intelligence, and have also put forward a new intelligence test.'

'The implications of a universal intelligence test also impact on many other disciplines. This could have a significant impact on most cognitive sciences, since any discipline depends largely on the specific techniques and systems used in it and the mathematical basis that underpins it.
"The universal and unified evaluation of intelligence, be it human, non-human animal, artificial or extraterrestrial, has not been approached from a scientific viewpoint before, and this is a first step," the researcher concludes.'

And what about our wetware? Neuroscientists are on the hunt.
posted by VikingSword (11 comments total) 8 users marked this as a favorite

 
Intelligence does not necessarily mean sentience. Obligatory link to Blindsight.
posted by infinitewindow at 12:51 PM on January 28, 2011 [1 favorite]


Is there a website or something where I could take this new universal intelligence test?

Have I already failed the universal intelligence test by not having noticed a link to it?
posted by Flunkie at 1:06 PM on January 28, 2011


Or at least a place where it's described in more detail than merely "uses the so-called 'Kolmogorov complexity"?
posted by Flunkie at 1:12 PM on January 28, 2011


I was skeptical that they'd be able to describe a scientific method for determining what the concept of intelligence should consist in. But then I clicked through, and it sounds like they've just assumed some definition of intelligence and what they're looking for is some technical method of testing it in babies, robots, and aliens.

Also I simply don't understand this bit:

The implications of a universal intelligence test also impact on many other disciplines. This could have a significant impact on most cognitive sciences, since any discipline depends largely on the specific techniques and systems used in it and the mathematical basis that underpins it.

Is this a non-sequiter? Or am I missing something? Sure, cognitive science depends largely on the techniques it uses ... but does it really depend on a technique for measuring intelligence in babies, robots and aliens?
posted by creasy boy at 1:13 PM on January 28, 2011


Flunkie, you can always contact the researchers. Link to the abstract.
posted by VikingSword at 1:14 PM on January 28, 2011


Somebody's gonna trot it out, might as well be me: g, a statistical myth.
posted by nzero at 1:16 PM on January 28, 2011 [1 favorite]


Neat... At a high level it seems like their work is about applying some of the fundamental principles from information theory and complexity theory to enable comparison between two computers. My basic/curious questions with this approach would be 1) that intelligence is itself vaguely defined; how would you resolve this in a formal definition? and, 2) K() is uncomputable, so what does it even mean to talk about the Kolmogorov complexity of Michelle Kwan?
posted by polymodus at 2:16 PM on January 28, 2011


They need to test out their intelligence test on cats.

Scientists used to think cats were colorblind and a lot less intelligent than they now realize, because cats refused to go along with their standard animal tests.

"I don't care if you have put my favorite snack behind the green door. I know I'm being played, and I refuse to be played."
posted by eye of newt at 8:49 PM on January 28, 2011


eye of newt "I don't care if you have put my favorite snack behind the green door. I know I'm being played, and I refuse to be played."

Equally applicable to humans, dogs, pretty much everyone maybe except AIs. Intelligence testing is wildly skewed by social expectations and context, and social benefits and costs perceived by subject and tester.
posted by aeschenkarnos at 2:57 AM on January 29, 2011


Website with preprint

http://users.dsic.upv.es/proy/anynt/
posted by bleary at 4:08 PM on January 30, 2011


The website I posted above links to a page on c-tests with online tests:

Computational Measurement of Intelligence
posted by bleary at 4:15 PM on January 30, 2011


« Older Mike Doyle is making a series of intricate decayed...  |  Merit badges for nerds, mother... Newer »


This thread has been archived and is closed to new comments