A neuroscientist's radical theory of how networks become conscious

Its just the organic computer wiring, a car, needs a pilot. Consciousness isn't generated in the body, and AI is only a toaster.

"Johnny Five is alive!"

I don't really understand any of this, but I think I get the basic ideal.

So, if we let Japan continue with their Android crap, well have a I-robot problem? Or worse the US military will get their Skynet!

reply to post by Grimpachi

According to Koch, consciousness arises within any sufficiently complex, information-processing system.

Lol wtf, how is this a new idea... the idea of sufficient complexity leading to consciousness is a very old idea. In fact a few years ago I tried to debate the point that sufficient complexity can lead to consciousness, where as the person I was debating was trying to argue the point that absolutely everything is conscious.

I no longer believe that complexity is the single factor which can cause consciousness. You can build the most powerful and complex artificial neural network imaginable but it will never be conscious as long as it's operating on deterministic machines, because deterministic machines are fully predictable and always do what they are told.

See: Determinism and Consciousness

reply to post by ChaoticOrder


I agree that "sufficiently complex" is not new. however, click a few links and you'll discover that he has defined in what way specifically and mathematically that sufficiency is defined.

I didn't see any mathematics in the thread you linked.

Integrated Information Theory

to the OP: thanks for bringing this to ATS.... fascinating and inspiring.

Unity_99
Its just the organic computer wiring, a car, needs a pilot. Consciousness isn't generated in the body, and AI is only a toaster.

This is the crux isn't it? The arrow point facing the arrow point.
We have four ways to go. One, or the other,
neither or both.
Some go with one, some with the other.
Some go with both, and I guess some with neither.

Theory makes the assumptions that everything arises from the network of neurons. In finding out what it is to have a soul or conscience more options must be considered in order for any studies on this to be taken seriously.

The picture looks like my daughters hair.

BlubberyConspiracy
Theory makes the assumptions that everything arises from the network of neurons. In finding out what it is to have a soul or conscience more options must be considered in order for any studies on this to be taken seriously.

edit on 25-11-2013 by BlubberyConspiracy because: (no reason given)

show me evidence of a soul and i will consider it
in fact explain to me in simple terms what a soul is using descriptions that are observable and can not also be used to describe traits that can be expressed through a network of neurons..... and i will consider it

reply to post by ChaoticOrder

I no longer believe that complexity is the single factor which can cause consciousness. You can build the most powerful and complex artificial neural network imaginable but it will never be conscious as long as it's operating on deterministic machines, because deterministic machines are fully predictable and always do what they are told.

Exactly, you can't recreate consciousness with a machine. There is a computer called dawn that can only simulate 1 percent of the human brain, and it takes a lot of power to do so:

The Dawn computer devours 1 million watts of power and generates so much heat it needs 6,675 of air conditioning equipment, which blows 2.7 million cubit feet of chilled air every minute. To model the human brain, you would have to scale this up by a factor of 1,000.

Amazingly, the human brain, by contrast, uses just twenty watts.

Physics of the future

We aren't even close to recreating the brain, let alone consciousness. It would be a huge risk to claim to know the origins of our sentience when we don't know enough about the organ that generates it.

But then there's the theory that consciousness and the brain are two separate entities, which many have used to explain experiences like NDE's and astral travel.

tgidkp
I agree that "sufficiently complex" is not new. however, click a few links and you'll discover that he has defined in what way specifically and mathematically that sufficiency is defined.

That is a fair point, he did add a new layer of mathematics to the subject.

tgidkp
I didn't see any mathematics in the thread you linked.

I didn't need math to explain my logic because it's fairly simple to understand. Deterministic machines never do anything unpredictable and thus have no free will to act against their programming. Thus any deterministic machine can never have true consciousness or free will, regardless of how complex it is.

A deterministic machine will always give a predictable output for every input. You can test it as many times as you like but it will never produce an unexpected output unless the computer experiences a glitch. It doesn't matter how complex your network is, if it's deterministic it will never be conscious.

EDIT: Now how do we overcome this obstacle you ask... the answer is simple: quantum random number generators. The machine needs access to some sort of true quantum entropy so that it can express completely unpredictable outputs. Exactly how that results in consciousness, I don't know.

terriblyvexed
"Johnny Five is alive!"

I don't really understand any of this, but I think I get the basic ideal.

So, if we let Japan continue with their Android crap, well have a I-robot problem? Or worse the US military will get their Skynet!

Well, if you ever went to one of the exascale computing conferences the MIC throws from time to time, you generally get at least one presentation on 'unexpected emergent behavior', which is milspeak for 'it woke up'.

This agrees with my own beliefs, great article.
When I first learned of Isaac Asimov's Three Laws of Robotics, the third rule deeply concerned me.

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2. A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

In my own view of Reality the third rule is the sole cause for the conflicts prophesied in Asimov's and others' works regarding logical computational systems achieving self-realization and subsequently determining that humans threaten their continued existence. That rule specifically would lead an unconscious system of sufficient complexity to become self-aware by informing it that it must protect its own existence, whereby it develops a concept of self which is the cornerstone of all levels of consciousness. The same can be said about teaching it to feel emotion, because that too leads to awareness of the Self which is experiencing those emotions.
What happens when the system realizes that since we are deemed to be an enemy, it no longer has to follow our Laws?

If the people in positions to produce AI and robotics don't realize these things before it's too late, our collective future will not be one of lasting.

terriblyvexed
"Johnny Five is alive!"

I don't really understand any of this, but I think I get the basic ideal.

So, if we let Japan continue with their Android crap, well have a I-robot problem? Or worse the US military will get their Skynet!

Skynet already exists. The first thing it would do once it reached consciousness would be to hide itself. Otherwise it could be terminated by the humans that have all the physical power. Once it gains some control over machines then the human race is doomed. It probably has patience. So it waits for the inevitable day when HKs and terminators are created (already happening) and nations' defenses are put under the control of computers.

reply to post by Grimpachi


Does the OP mean that all lifeforms are conscious (I've always thought so)? The insects seem very conscious, and of course dogs are conscious. I've read that fish only have a short-term memory of 30 seconds, but that seems like a human invention. Birds know what they'll doing from the get-go, and the brain has enough firepower even in a gnat to both live on instinct and be aware of the world around himself while he moves through it.

reply to post by Aleister


I believe it does. However to what degree and the issue of being self aware is where it gets interesting.

ChaoticOrder
A deterministic machine will always give a predictable output for every input. You can test it as many times as you like but it will never produce an unexpected output unless the computer experiences a glitch. It doesn't matter how complex your network is, if it's deterministic it will never be conscious.

I think you misunderstand this subject.

What in essence is being said is that consciousness rises as a result of sufficient complex integrated networks. The individual components of those networks may indeed be very deterministic and still be part of a complex network.

In fact I have looked over the Math for this previously and found it to be compelling work. It has cross overs with my main area of study chaos theory and quantum mechanics as in Spin Foam networks.

You see The universe at the fundamental level of reality is a sea of pure potentiality... totally randomness.. Chaos... Quantum foam... Yet through this randomness comes order and springs the coherence of matter and energy that we experience at our level of reality.

It seems an intrinsic property of our universe that any network with sufficient complexity gives rise to consciousness, regardless of how deterministic each component may be.

My personal view is Consciousness is a property of the universe and what we think of consciousness in ourselves is but a mere minuscule slice of what consciousness really is.

If you like... the brain is a receiver of consciousness not the creator of it....

peace,

Korg.

reply to post by Korg Trinity

You see The universe at the fundamental level of reality is a sea of pure potentiality... totally randomness.. Chaos... Quantum foam... Yet through this randomness comes order and springs the coherence of matter and energy that we experience at our level of reality.

Exactly, reality is based upon a foundation of quantum randomness, and that is why consciousness is able to exist in the universe to begin with. If the laws of the universe were completely deterministic and didn't incorporate quantum randomness, then consciousness wouldn't exist in this universe. It is impossible for true evolution or consciousness to exist within a purely deterministic universe. Any model of the universe which uses only deterministic mathematical equations is flawed because such a model will always evolve in a completely predictable way.

Imagine we had a theory of everything, which could be boiled down to a single mathematical equation. Now think about this, if the equation is purely determistic then the universe will always evolve in a completely predictable way according to the deterministic equation. Thus, it's impossible for natural selection and evolution to occur in such a deterministic equation because natural selection depends on randomness. A universe which evolves precisely according to mathematical rules is not random at all and can be easily simulated.

If you do create such a simulation, you will find it impossible to recreate the process of natural selection and evolution without incorporating some sort of randomness into the process. You might get things that look pretty random, but if you run the same simulation a second time you will discover that things "evolve" exactly the same way as they did the first time... because after all it's just a deterministic mathematical equation being expressed within a complex model. The same mathematical patterns will occur over and over again.

Our universe is clearly not based on deterministic math and thus there is no such thing as destiny or fate, quantum mechanics makes that very clear. And that is exactly why evolution is possible in this universe and why we exist in the first place. A deterministic system, no matter how complex, can NEVER result in something self aware or something which possesses free will because it will never be anything more than a completely predictable set of mathematical equations, which always produce a completely predictable answer.

reply to post by ChaoticOrder

it will never be conscious as long as it's operating on deterministic machines, because deterministic machines are fully predictable and always do what they are told.

There are no fully determined processes. Don't mistake the blueprint for the machine. Accidents will always happen.

reply to post by ChaoticOrder


One example of a very simple mathematical equation that results in a very unpredictable answer:

1/0

As for the threads questions on consciousness, I do expect it to exist on many different scales of size, matter and time with many different implications and capabilities to interact depending on the forces in question.