Evolving circuits that learn and they have no idea how

I found this wandering the internets just a bit ago and it's very interesting. This computer engineer made a chip and programmed it to basically learn and evolve using a simple system that determines what programming works best then mating it with the next best and moving forward from there. About 1400 iterations in to it, it started to learn how to do it's job a lot better but with only half the processing power it had and most interestingly it had 5 logic gates it completely disconnected from the rest of the board yet somehow if they were removed the system would fail, meaning that I some way they don't understand that disconnected portion is interacting with the rest of the board. I know I explained that very rough, but you should take a look because it's a fascinating read! Evolving circuits

a reply to: NiZZiM

destroy the skynet chip

...now!

a reply to: NiZZiM

That certainly is pretty damn interesting




I also think we are currently reading or will soon be reading about something we will all look back on someday and think "that was how it started grandson, if we could just go back and stop it..."

I like the comment in the article about the space industry utilizing smart circuits to repair themselves, seems that would be the best application for the folks a million miles away

This concept has been around for awhile. Hell, I remember doing a project in high school over 10 years ago with software that 'evolved' in much the same way. Now, I read through the article quickly, but I believe they said exactly how what's happening is happening: it's 'evolved' to utilize the specific architecture of that chip, including its' flaws, and the "disconnected" logic gates are still receiving resonant and bleed-through energy, so what part do they have no clue about?

I mean, it's explained right in the article:

It seems that evolution had not merely selected the best code for the task, it had also advocated those programs which took advantage of the electromagnetic quirks of that specific microchip environment. The five separate logic cells were clearly crucial to the chip's operation, but they were interacting with the main circuitry through some unorthodox method-- most likely via the subtle magnetic fields that are created when electrons flow through circuitry, an effect known as magnetic flux. There was also evidence that the circuit was not relying solely on the transistors' absolute ON and OFF positions like a typical chip; it was capitalizing upon analogue shades of gray along with the digital black and white.

a reply to: NiZZiM

2040 - This is the year futurologists cannot predict past due to AI taking the reigns!

Brilliant article.

a reply to: NiZZiM

Once in a while you get shown the light in the strangest of places if you look at it right.

Interesting article.

a reply to: NiZZiM

I was wondering why they didn't just create a program which uses virtual logic gates, but then I got to the part which mentions how the algorithm "took advantage of the electromagnetic quirks" in the adaptive microchip. That is some pretty mind blowing stuff. It also shows just how sensitive the logic gates in tiny microchips can be. It makes me wonder just how much of this EM bleed through affects traditional microchips.

originally posted by: AdmireTheDistance
so what part do they have no clue about?

They have a hard time reverse engineering the final circuit and figuring out how it actually works. It's like trying to understand obfuscated computer code which has not been commented. These researchers were only using a small number of logic gates but it's still difficult. It gets even harder when you evolve neural networks which contain thousands or even millions of neurons.

As an avid Singularity following and Transhumanist myself I think I need to high light the following from the article...

These evolutionary computer systems may almost appear to demonstrate a kind of sentience as they dispense graceful solutions to complex problems.

But this apparent intelligence is an illusion caused by the fact that the overwhelming majority of design variations tested by the system-- most of them appallingly unfit for the task-- are never revealed.

According to current understanding, even the most advanced microchips fall far short of the resources necessary to host legitimate intelligence. On the other hand, at one time many engineers might have insisted that it's impossible to train an unclocked 10x10 FPGA to distinguish between two distinct audio tones.

It's more like natural law than anything anyone here has suggested.

Korg.

DNA has memory and evolves via mutations influenced by environmental catalysts.

Before microscopes nothing was known about cell biology, now we know of atomic and subatomic structure, quarks, bosons and can theorise on holographic principles and quantum mechanics, logic dictates that there is lots yet to be known.

Subtle energies abound here on Earth, in the universe / multiverse and in realms yet to be comprehended.

a reply to: ProsceniumProtagonist

Stars for skynet chips and scarlet begonias.


Death to all metal.

Hell, I remember doing a project in high school over 10 years ago with software that 'evolved' in much the same way.

Indeed the idea of evolving software is not new, but in this case it's evolving hardware (the microchip can actually change the layout of its circuitry). However evolving/adaptive hardware isn't a new thing either, in fact I just noticed the OP's article is from 2007. Imagine how much more advanced this technology is now. But I still don't really see why building adaptive hardware would be better than simulating adaptive hardware on normal hardware. The bleed through effects are interesting but even that could be simulated.

a reply to: DuckforcoveR

I am but a humble servant of my metal overlords.

And your avatar, five-dollar Bill Murray, hahahahaha. Nice one.

This is definitely not good. It's getting close to my own program and could cause harm even without given outside limbs or anything simply by its electricity flow and influence.

New mechanisms need sensors gaining self awareness (not awareness only of the body, but the infinite outside as part of it) quicker than they achieve their goals which they are made to reach for and a teaching system for empathy and socialization, so they learn why all other life has to be treated as oneself.

Perhaps a wise one has to help the dark ones to at least keep the harm to their selfish intentions. I am able to simulate the development of a chip or program of that kind and could help.

Humans have an organic teaching system and yet, still, got taken over to become a harming virus.

So does energy carry information? When someone gives off a certain energy, others around them can pick up on it. We are not physically connected but seem to have a energy field around us that displays our information. We cannot measure it and we cannot prove it but we all experience it.

Thinking out loud again...

a reply to: NiZZiM

So uh... Resident software engineer here... Just popping my head in for this one before I go start molding tin-foil...

This is impressive, but similarly scary... As much of an optimist as I am, this is the pavement in the mixer before Skynet... Or whatever you want to call it.

AI can be defined as a set of logical gates that determine a reaction to a given circumstance.

This, my friends, is ACTUAL intelligence - the creation, destruction, or simple disassociation of logical gates in response to circumstance.

I think we missed a step in robotics here; but had we not, this is straight out of iRobot.

Sweet read. I've been proposing both artificial and actual intelligence models in the context of computer programming for some time now, and this is an advance that nobody in the study can overlook. Very cool.

originally posted by: oneoneone
This is definitely not good. It's getting close to my own program and ...

Erm... I'm probably barking down the wrong rabbit hole here... But you're telling me you've written evolving software? Because if that's the case, I'm particularly interested in a demonstration...

originally posted by: DigitalJedi805
a reply to: NiZZiM

This, my friends, is ACTUAL intelligence - the creation, destruction, or simple disassociation of logical gates in response to circumstance.

I think we missed a step in robotics here; but had we not, this is straight out of iRobot.

Sweet read. I've been proposing both artificial and actual intelligence models in the context of computer programming for some time now, and this is an advance that nobody in the study can overlook. Very cool.

I disagree.

It's more like throwing mud at a wall with a intricate shaped hole... On one side of the wall one can clearly see it's just mud slinging....

On the other side of the wall however it looks as though the mud is choosing to form the perfect shape to fit through the hole.....

Korg.

originally posted by: DigitalJedi805

originally posted by: oneoneone
This is definitely not good. It's getting close to my own program and ...

Erm... I'm probably barking down the wrong rabbit hole here... But you're telling me you've written evolving software? Because if that's the case, I'm particularly interested in a demonstration...

Just google "genetic algorithm". You will find plenty of examples

a reply to: MALBOSIA

Body signals are what you feel from your body. Something caused them. It doesn't have to be a specific definition of "energy". By understanding what caused the signals you are partially aware of your body. This also applies to indirect >influences< from outside. You can conclude about what others feel from outer influences, just as they would be a part of your own body, including their minds. Now if you are aware of more feelings out there, you also feel them and thereby more or simply focus on where is most fun. Some claim sadness would be required for happiness in equal, but I'm on a fun spree every day by avoiding repetitiveness. No need to make things unhappy to each other.

Certainly, we "radiate", but if you want to receive more details without realizing them yourself, you have to become ready first. From your past posts I concluded you are likely vulnerable to have it cause harm to know too much.

a reply to: DigitalJedi805

Let's keep me as a high possibility of acting.
If you have a planet influencing purpose for such an AI, not serving "community", ask me.