Physicist Proposes New Way To Think About Intelligence
-Chris Gorski, Inside Science News Service
Chris Gorski is an editor for Inside Science News Service.
Zach Cox said...
The general idea that this relies on is quite sound. The underlying idea is the definition of consciousness.
Consciousness for any motile organism is the organism's ability to take its sensory inputs and use them to create a model of the future.
The only purpose of the model thus created is to allow the organism to successfully negotiate its environment.
This success, in turn, is the definition of what it means to say natural selection will select in favor of the organism.
Like every property of living organisms such a thing seems to be on a continuum of sophistication, all the way from cockroaches to birds, and mammals that are what we humans think of as conscious.
Sunday, June 2, 2013 at 9:23 PM
'The researchers suggest that intelligent behavior stems from the impulse to seize control of future events in the environment. This is the exact opposite of the classic science-fiction scenario in which computers or robots become intelligent, then set their sights on taking over the world.'
Does this make sense to anyone? I realize the paper has nothing to do with robots and this is just mind candy, but it seems like if we're going to bring robots taking over the world into the conversation a paper arguing that 'the impulse to seize control over future events in the environment' is the hallmark of intelligence should be granted as one more reason to lay the welcoming mat out for our robot overlords. Or is the author saying that robots can't really be called intelligent *until* they've taken over the world? That seems like a high bar to me, one which surely anybody believing humanity to have achieved is going to face an uphill battle when confronted with our seemingly totally rudderless teleology.
Wednesday, May 1, 2013 at 11:57 PM
Is it correct to tie intelligence to conciousness? Many things are concious without being intelligent. We have created machines that appear to be intelligent without necessarily being concious. Many things are alive without being either of the preceding two. Creating equations that produce behaviour that appears to mimic intelligent learning is possibly all that happened to us a few (million) years ago. After all, if it looks like a duck etc...
Monday, April 29, 2013 at 12:03 PM
Mark Waser said...
So . . . . what is the equation? Is it any different from the second law of thermodynamics as ramified through the universe? Why is the article all advertising and no "meat"?
Monday, April 29, 2013 at 11:19 AM
If this ran in a big and fast enough computer and was connected to the internet, it could cause the "singularity"...
Monday, April 29, 2013 at 2:34 AM
Actually, I think it sounds quite interesting. If it works, my guess is that it will probably turn out to be a variant of what is sometimes called temporal difference learning. This is a family of generalised learning algorithms which have been under development for some time in AI and computational neuroscience.
Sunday, April 28, 2013 at 8:13 PM
Didier Greusard said...
Very interesting! Gives me a lot of food for thoughts. And here again I can see how the deeply grounded western materialist view of the world leads to gross mistakes. This research is good step toward the right direction and will be full of opportunities, but this team should not forget that intelligence emerge primarily from consciousness. Every living beings share this common property to directly perceive, know, or in one word be conscious of its surrounding environment. Before computing information you have to acquire the basic information. Direct perception, awarness of the environment are the most basic information required for living intelligence to take place.
Saturday, April 27, 2013 at 5:09 AM
It does seem a little hard to believe.
Friday, April 26, 2013 at 3:09 PM
I'm a physicist myself and I love and respect mathematics... but I don't believe what Wissner-Gross and Freer state.
It seems to me just pseudo-science.
But I might be wrong (even if I don't believe so).
Friday, April 26, 2013 at 3:06 PM