Polyworld: Using Evolution to Design Artificial Intelligence

Added: Nov 13, 2007

From: googletechtalks

Duration: 66:38

Google Tech Talks November, 8 2007 ABSTRACT This presentation is about a potential shortcut to artificial intelligence by trading mind-design for world-design using artificial evolution. Evolutionary algorithms are a pump for turning CPU cycles into brain designs. With exponentially increasing CPU cycles while our understanding of intelligence is almost a flat-line, the evolutionary route to AI is a centerpiece of most Kurzweilian singularity scenarios. This talk introduces the Polyworld artificial life simulator as well as results from our ongoing attempt to evolve artificial intelligence and further the Singularity. Polyworld is the brain child of Apple Computer Distinguished Scientist Larry Yaeger, who remains the primary developer of Polyworld: http://www.beanblossom.in.us/larryy/Polyworld.html Speaker: Virgil Griffith Virgil Griffith is a first year graduate student in Computation and Neural Systems at the California Institute of Technology. On weekdays he studies evolution, computational neuroscience, and artificial life. He did computer security work until his first year of university when his work got him sued for sedition and espionage. He then decided that security was probably not safest field to be in and he turned his life to science.

Channel: People

Tags: education  engedu  google  googletechtalks  talk  talks  techtalk  techtalks 

Rating: 4.77 (195 ratings)    Views: 39225' favoriteCount='36    Comments: 25

qarnos Says:

Sep 14, 2008 - I challenge the claim that Polyworld is pure natural selection with no fitness function. Virgil mentions that there's a cost function applied to the networks. Obviously, if there is no cost function, then the networks will be filled with a lot of neurons that do nothing, as long as they are being "carried" by a useful neuron. The problem is the cost function is arbitrary. Unless the cost function is subject to natural selection, this is not purely natural. It is essentially a fitness function.

DbzmasterdbzNecross Says:

Sep 16, 2008 - inthefade- Err...there are already AI that should be able to do that. In certain games they can temporarily remember moves you do and use them, I'm pretty sure, and if not, it's entirely possible and just hasn't been done yet.

inthefade Says:

Sep 16, 2008 - You're talking about something different. I'm not talking about AI *learning* what you do, and then trying to counter it, I'm talking about AI that uses a randomly mutating genetic algorithm for behavior and then a fitness function to naturally select more successful behavior. This is a significantly different model for AI, and you are right, it is entirely possible, and maybe someone has tried it already.

EmotionIsIgnorance Says:

Sep 23, 2008 - in spore they dont evolve they made it look like you do but in fact this is far more closer to evolution.

Colombian459 Says:

Sep 28, 2008 - What? lol then we will never achieve human level intelligence!

rmessenger Says:

Sep 29, 2008 - "Natural selection" is used as an analogue. It's not really natural, but it is selection. You can't have any kind of selection (or evolution) without a cost or fitness function. In nature there are sets of probabilities that act as the "fitness function" although not explicitly stated as a "function"

Luftwaffe101 Says:

Oct 2, 2008 - sp sp sp sp sp spit it out!

batymun Says:

Oct 5, 2008 - dicks and cocks

cyborgtroy Says:

Oct 6, 2008 - Yeah Spore is rly annoying bc it's a game, not anything about evolution. If Spore were evolution you wouldn't even be playing :P

amayami Says:

Oct 8, 2008 - Wouldn't kill him to rehearse his debate before showing up...too many uh's and uhm's.

Olucatei Says:

Oct 9, 2008 - Maybe if they added an output neuron that gave signal to an additional input neuron, the organisms could have a short pseudo-memory. I know nothing of these systems, just a thought.

TrannyGirl15 Says:

Oct 15, 2008 - The cost function is there to mimic real physical constraints. Animals, in the real world, need to eat something in order to survive...

deepdivestudios Says:

Oct 18, 2008 - Can anyone (including intelligent designers) come up with an example of ANYTHING that is NOT a cause-effect product? And isn't that what evolution really is? From that perspective, don't we use evolution in design, no matter what approach we take? I'm being a smart-ass... sorry. Interesting vid.

Maskuerade Says:

Oct 29, 2008 - very very interesting . this brought a new viewing lense (perspective) on how i have always viewed artificial intelligence. This is REALLY really, alluring. I should look into studying it.

lancstony Says:

Oct 31, 2008 - Subtitles would be helpful.

danielgc101 Says:

Nov 3, 2008 - Agreed.

Doomfrost Says:

Nov 9, 2008 - Very nice video, though yeah he seemed nervous, he was perfectly understandable. I would like to see a day where game AI could have "human like" thinking.

TheArchitectOfDreams Says:

Nov 13, 2008 - Smart guy, needs to work on his speeches. Slow down... dude. ^_^ Other then that, understandable

vvvvtwo Says:

Nov 18, 2008 - Its like he fears for his life...does Google make people do speeches at gunpoint?

mikeyo1234 Says:

Nov 19, 2008 - Agreed... his delivery was very jittery and not flowing... which leads to his speech requiring excessive concentration.

icedthai Says:

Nov 20, 2008 - Nuclear decay does not have to be cause-effect induced. A particular atom may decay now or wait 14 billion years. Arguably perhaps we don't know how to define the atom well enough to prove rigorously that it is also cause-effect grounded. There are other examples along these lines that involve "quantum decisions".

EmmyKokoro Says:

Nov 21, 2008 - Something like... "uhm... so, yeah! uhm... now, uhm... something like that." :P

SoulPhantomX Says:

Nov 21, 2008 - Cause and effect? so if someone attacks you(cause) you have the option to run or fight for survival (effect?)

SoulPhantomX Says:

Nov 21, 2008 - Does that mean that AI can't form relationships or base their decisions on incoming situations that may effect others around them, the sole purpose is self preservation, but sometimes in life we have to sacrifice for the greater good, this ability defies his rule of energy conservation.

SoulPhantomX Says:

Nov 22, 2008 - I suppose more complex interactions with not only the environment, but with users, and other agents(AI forms?) can be incorporated later on. As he said mid presentation, this is the raw of the raw. Thanks for the presentation, this was informative and thought provoking, I have a lot of questions, and theories on the subject.