| Author
|
mind to computer -- is it ever possible?
|
zooter
IsraTrance Junior Member
Started Topics :
44
Posts :
771
Posted : Jul 1, 2004 16:05
|
ok, this time one in favor of AI  (im trying to be neutral here)...
long time back i read an article wherein robots were pitted against each other in a very similar "evolution" condition where only the fittest "genes" would contribute to the next pool of robots....(here, i don't know what genes mean, but i assume that's the neural net part)....
now the interesting thing is one of the robots actually ran away!!!!!!!!!!!!!!!!
here is a small article...i couldn't find the link to the entire story
http://www.theage.com.au/articles/2002/06/20/1023864460978.html
So, then again....are we back to square 1???
psyk: that's what every1 thinks (IF only i had this, IF only i had that)...just kidding...even i think the same way & i'm sure most ppl in this forum think the same way 
|
|
|
ZilDoggo
Started Topics :
4
Posts :
663
Posted : Jul 1, 2004 16:54
|
zooter,
hehehe.
if the scientists can't make a decent cage then how do you think they will build a brain??
(actually, this is copied from someone at slashdot in the discussion on this story.., i just thought that it's a nice argument,. )
greets,
aka., |
|
|
|
Amygdala
Amygdala
Started Topics :
12
Posts :
175
Posted : Jul 1, 2004 18:36
|
Hehe, cute story
Actually the idea of evolving agents is quite common - even for solving simple mathematical problems. One encode an algorithm (being robot-control, number-chrushing or whatever) as a set of artificial genes, that can be easily manipulatied (mutated) or paired up (reproduction by mating) to form new algorithms... The new algorithms are evaluated for performance, and the best ones are given higher chances of either mutating or mating. After a number of generations, one should end up with very good algorithms for solving the problem.
This is usually called evolutionary or genetic algorithms, and are the topic of much research. One major downside, is the amount of time it takes to breed a good solution, and the fact that it takes a pretty darn good solution to the initial problem, to actually evaluate the genetic algorithms. But it's fun
- A |
|
|
|
zooter
IsraTrance Junior Member
Started Topics :
44
Posts :
771
Posted : Jul 2, 2004 07:00
|
Quote:
|
On 2004-07-01 16:54, ZilDoggo wrote:
zooter,
hehehe.
if the scientists can't make a decent cage then how do you think they will build a brain??
|
|
offtopic:
SPOILSPORT  hehehe...that was a good one
 Interviewer: "So Frank, you have long hair. Does that make you a woman?"
Frank Zappa: "You have a wooden leg. Does that make you a table?" |
|
|
|
Top-down
Inactive User
Started Topics :
7
Posts :
119
Posted : Jul 13, 2004 03:52
|
|
zooter
IsraTrance Junior Member
Started Topics :
44
Posts :
771
Posted : Jul 13, 2004 07:18
|
Top-down, nice links
Amygdala was saying about exactly these 2 experiments, but "lost" the link 
Interviewer: "So Frank, you have long hair. Does that make you a woman?"
Frank Zappa: "You have a wooden leg. Does that make you a table?" |
|
|
|
