Altruistic robots produced through evolution

[Mesozoic Mister Nigel]

I keep thinking the title says "Autistic robots produced through evolution".

Anyway, to address the subject of Darwinian selection, which I suspect they used to differentiate it from "natural" selection, (and could have refined further to call "Darwinian-modeled selection) in a sense, it IS the robot's natural environment, as both robots and environment were created by humans.

And either way, it's really damn neat that the robots evolved in an unexpected manner.

[Triple Zero]

evolutionary computing is full of awesome surprises like this.

[Rococo Modem Basilisk]

evolutionary computing is full of awesome surprises like this.

Like that time that the two FPGAs learned to use each other's circuits through electromagnetic resonance

[Requia ☣]

And, astonishingly, the robots were even able to evolve altruistic behaviour, in a task that involved pushing tokens around. Some could be pushed single-handed, earning the robot one 'fitness point'; others required two robots, gaining the whole group one point.

I get confused where the line between mutually beneficial and altruistic is drawn at the species level.  At a genetic level, isn't it all selfish?

For example, I think a better test for altruism would be if helping another bot (for simplicity, of the same species - not a widely differing genome) push their token didn't necessarily grant a direct benefit to the individual who was helping.

example is not altruism.
example is selfish.
robot that tries to push 2robot tokens gets more points than if the same robot didn't try to push 2robot token.
if robot benefits, robot not altruistic.
indirect benefits count.
it would be altruism if the robot did not benefit, but other robots did.
evolutionary pressure can't create altruism. because pressure only exists if there is benefit.

for the rest:
AWESOME. BALLS.



PS i know people will be difficult about this 'altruism can't evolve' point so let me quote a respected expert:

indirect benefits count.

The robot does not get more points if it takes the 'altruistic' path.  It gets the same number of points, but it benefits other robots, not just the two involved.  This *actively harms* the robots taking the 'altruistic' option, since it hurts their reproduction chances as individuals, but increases it for the group.

[Captain Utopia]

The artificial life simulation Tierra does have something more akin to natural selection. In that simulation machine code programs have to evolve their reproduction routines themselves as well*, so the ones that do this most efficient and fast reproduce more. That would not be artificial selection, right? In this simulation, only the mutation is artificial, a tiny percentage of random bitflips in computer memory.

Tierra is interesting because it evolved parasites, which would consume resources from the hosts, which then evolved to almost defeat the parasites, but then the parasites evolved yet again, to defend themselves, etc.

At that point they realised that they had forgotten to implement the mutation routine -- all this sophistication was achieved through nothing more than splicing genes.  They were surprised as their assumption was that mutation was required for any interesting behaviour to occur.

[Jasper]

I would like to see these people work with the people who made this:

http://www.technologyreview.com/computing/22339/?a=f

In fact, the current prototype can operate about 100,000 times faster than a real human brain. "We can simulate a day in a second," says Karlheinz.

[Requia ☣]

They already are using one of those.  Just not in hardware form.  Simulated neural nets are nothing new.

Also, I really wish people would stop comparing simulated neural nets to brains.  It's not the same thing.

[Bu☆ns]

Despite efforts to make the chips as biologically plausible as possible, Markram admits they are still crude compared to what can be achieved in simulation. "It's not a brain. It's a more of a computer processor that has some of the accelerated parallel computing that the brain has," he says.

[Reginald Ret]

And, astonishingly, the robots were even able to evolve altruistic behaviour, in a task that involved pushing tokens around. Some could be pushed single-handed, earning the robot one 'fitness point'; others required two robots, gaining the whole group one point.

I get confused where the line between mutually beneficial and altruistic is drawn at the species level.  At a genetic level, isn't it all selfish?

For example, I think a better test for altruism would be if helping another bot (for simplicity, of the same species - not a widely differing genome) push their token didn't necessarily grant a direct benefit to the individual who was helping.

example is not altruism.
example is selfish.
robot that tries to push 2robot tokens gets more points than if the same robot didn't try to push 2robot token.
if robot benefits, robot not altruistic.
indirect benefits count.
it would be altruism if the robot did not benefit, but other robots did.
evolutionary pressure can't create altruism. because pressure only exists if there is benefit.

for the rest:
AWESOME. BALLS.



PS i know people will be difficult about this 'altruism can't evolve' point so let me quote a respected expert:

indirect benefits count.

The robot does not get more points if it takes the 'altruistic' path.  It gets the same number of points, but it benefits other robots, not just the two involved.  This *actively harms* the robots taking the 'altruistic' option, since it hurts their reproduction chances as individuals, but increases it for the group.

it accepts a new source of points, is this not an increase in potential points?
2 foodsources are better than 1.
in a medium with glucose and fructose bacteria that can use both do better than those that use only glucose.
i cant make it any clearer.

[Requia ☣]

A) It's not food.

B) It hurts the robot to go down this path rather than be selfish.

[Reginald Ret]

A) It's not food.

B) It hurts the robot to go down this path rather than be selfish.

hurt?
how?

because if that is the case you are completely right ofcourse.

[Requia ☣]

according to the article, giving points to the other robots increases their odds of reproduction, which in turn hurts the cooperative bots since only so many get to reproduce each round, only the bots that are selfish have a high chance to reproduce.

[Jasper]

They already are using one of those.  Just not in hardware form.  Simulated neural nets are nothing new.

Also, I really wish people would stop comparing simulated neural nets to brains.  It's not the same thing.

They are similar, if simpler.  Anyway the advantage with those chips I linked is that they're physical neural nets.  Simulations require a great deal of computation to run on normal procs, this doesn't.   The article said the  robots being tested were using only a few dozen "neurons", while these chips employ hundreds.  And they're scalable.  So, no, they weren't using one of these.   Stop thinking about "neural nets aren't brains" and "really real altruism for realness" and start to think about the meaning of a robot that had a physical neural net with orders of magnitude more sophistication.

[Jasper]

Also, altruism isn't a matter of game theory, which is how these robots are operating.  Altruism comes from intentionality (and sentimentality), which these robots tend to lack.

[Rococo Modem Basilisk]

In the context of game theory, altruism tends to be modeled as a form of naiive and culturally-induced enlightened self interest -- since the superorganism (the community as a whole) benefits from particular types of altruism, they are encouraged regardless of the effects on the individual since if most of the individuals in the community engage in the altruistic behavior most of them likewise get the benefits of the altruism of the others.