On Consciousness

[steenkh]

I think you are dodging the fact that modeling is a mathematical discipline.

Not all models are mathematical models. When you build a model of the interior of a star, you are using maths all the way up. Financial models are statistics all the way, which is also math, but a model of a brain could rely only in parts on math, as I said above. The point is that so far, we only have biology to base it on. I am not aware that anybody have made a mathematical model of the brain. Are you?

So these neural networks just appeared out of nowhere then? No mathematical modeling required?

Actually, they are called neural networks because they are based on a crude biological model. They are not capable of doing anything on their own, but have to learn by reward and punishment. When they can do what they were intended to do, nobody knows how they do it.

Neural networks have also failed completely in achieving anything like consciousness, and modern attempts at achieving artificial consciousness are taking the path of creating an emulation that uses much more detailed simulation of the elements of a brain than before.

 

[Belz...]

Machines don't build cars by themselves.
Try again.

Since I assume you fully understood the point of my analogy, I shall henceforth conclude that you are deliberately avoiding my point.

 

[Belz...]

1 - Machines will never be conscious.
2 - We don't need conscious machines because we have conscious people.
3 - Those who disagree with (1) or (2) are egotistical.

If I didn't know better, I'd think Kaggen has an ideological aversion to the idea that machines can devellop consciousness. Gee, I wonder what the source of that could be.

 

[dlorde]

... If you have been following this thread, you will know that we already have crude systems called neural networks that are able to perform tasks without anybody, and certainly not any programmer, knowing how it is performed. This is not something you can put in a formula, at least not until afterwards.

We also have neural networks based on digital computing that reproduce the same functional behaviour as biological neural networks in the brain, suggesting that at least partial emulation is possible.

 

[dlorde]

Neural networks have also failed completely in achieving anything like consciousness...

Has anyone seriously attempted anything like consciousness with neural networks? It would seem somewhat optimistic given current state of the art...

 

[trebor]

Has anyone seriously attempted anything like consciousness with neural networks? It would seem somewhat optimistic given current state of the art...

I would guess that consciousness is only something that would be recognised in a neural network after the fact...

This might be a project to watch though:
http://www.humanbrainproject.eu/

 

[Prometheus]

Fascinating.

“...you actually can point to a physical substrate to memory,” what the researchers call an engram....the recent research is the closest they have gotten to pointing to a spot in the brain and saying, “That is the memory.”

In the research reported Thursday, Dr. Tonegawa’s team first put mice in one environment and let them get used to it and remember it. They identified and chemically labeled the cells in the animals’ brains where that memory was being formed. The mice were not shocked in that environment.

A day later, in a completely different environment, the researchers delivered an electric shock to the mice at the same time that they stimulated the previously identified brain cells to trigger the earlier memory.

On the third day, the mice were reintroduced to the first environment. They froze in fear, a typical and well studied mouse behavior, indicating they remembered being shocked in the first environment, something that never happened. The researchers ran numerous variations of the experiment to confirm that they were in fact seeing the mice acting on a false memory.

The tools of optogenetics, which are transforming neuroscience, were used to locate and chemically label neurons, as well as make them susceptible to activation by blue light transmitted by a fiber optic cable. With these techniques the researchers were able to identify and label which neurons were involved in forming the initial memory of the first environment, and to reactivate the labeled cells a day later with light.

 

[steenkh]

Fascinating.

Indeed.

I would like to know if the engrams are located in the same spot in different mice. It would be surprising if they are.

 

[Mr. Scott]

Neural networks have also failed completely in achieving anything like consciousness

Linky?

 

[steenkh]

Linky?

No linky. My informations are outdated and from a time when I still had contacts in the field.

Anyway, if consciousness had been achieved then or later, I doubt we would have the debates in this forum that we do.


Sent from my iPhone using Tapatalk 2

 

[deadendmind2013]

From what? And how?

Well, "supposedly" from AI.
Consciousness is a product of evolution. There is gene expression and neocortex and CNS. Those things are not really easy to simulate.

As I said to Pulvinar before, I dont think that it really helps to juxtapose or replace "real consciousness" with "real numbers" (for the sake of a thought experiment) and it is confusing to think about it. (I think that a number is far more mysterious than consciousness). But maybe it is just confusing for me, I really have a problem when it comes to similes and metaphors and thought experiments.

 

[Belz...]

Well, "supposedly" from AI.
Consciousness is a product of evolution. There is gene expression and neocortex and CNS. Those things are not really easy to simulate.

Why do you say that ?

Running is also gene expression, and it's easy to imagine reproducing that. I wonder why consciousness is afforded special status. I wonder, I wonder...

 

[Arcade22]

Consciousness is a product of evolution. There is gene expression and neocortex and CNS. Those things are not really easy to simulate.

And neither is "simple" things like human-like balance, movement, coordination and object-manipulation. Why would "consciousness" be special or different?

 

[tsig]

And neither is "simple" things like human-like balance, movement, coordination and object-manipulation. Why would "consciousness" be special or different?

Because people want to be special, they were told they were their mothers' special darling and they want to be gods' special snowflake. Otherwise life is grey and meaningless, there's no art or inspiration or joy just lumps of matter.

 

[Mr. Scott]

Because people want to be special, they were told they were their mothers' special darling and they want to be gods' special snowflake. Otherwise life is grey and meaningless, there's no art or inspiration or joy just lumps of matter.

Yes, the idea that we are wet robots some people find highly offensive, but a fact is a fact regardless of how it makes you feel.

 

[deadendmind2013]

Why do you say that ?

Running is also gene expression, and it's easy to imagine reproducing that. I wonder why consciousness is afforded special status. I wonder, I wonder...

To be honest, it's because I had the Blue Brain Project in mind, where gene expression is somewhat and to some extend simulated? I think I read somewhere in wikipedia about this, I dont really know of any link though about any scientific literature publication. I am just not confident about some predictions from some scientists that say that we ll create an AI within 10 years from now. I guess I read too much science fiction. Maybe I am wrong, who knows?

Why consciousness? Well, I ve been trying to read this thread from page 1 and I find the arguments here fascinating. I am just here to learn. Also, it combines Physics, Psychology, Biology, Physiology... Also, I think this thread here is the best one can find on the web about the concept of consciousness.
Anyhow, I get the feeling that we ve started on the wrong foot here. I do not mean to offend anyone by saying that I am not really confident about the Blue Brain Project.

 

[Belz...]

Re: the 10 years from now thing, I think they've been saying we'll have fusion in 50 years for about 50 years, now.

 

[Fudbucker]

I am not claiming any single element to be the cause of consciousness. Currently, we do not know what causes consciousness, but we know where we can find consciousness in nature. Accordingly, if we emulate that in sufficient detail, we know we will achieve consciousness. There is very little math involved.

If you have been following this thread, you will know that we already have crude systems called neural networks that are able to perform tasks without anybody, and certainly not any programmer, knowing how it is performed. This is not something you can put in a formula, at least not until afterwards.


No. As far as I know, nobody has presented a mathematical model that approximates biological processes, although at the lowest level you may use maths to decide when an emulated neuron should "fire". Putting a lot of these neurons together is not the same as having a mathematical model.


First of all, it is a straw man, and secondly it is nonsensical. Why on Earth would you want to hear that something being simulated is not the same as the original? It is so blindingly obvious that one wonders why you are asking the question in the first place. Perhaps you have been reading too much philosophy?


Obviously. And your point is?



Even when I was young I noticed that French philosophers had a way with words that sounded very fancy but was really without content.

Mathematical and theoretical biology is an interdisciplinary scientific research field with a range of applications in biology, biotechnology, and medicine.[1] The field may be referred to as mathematical biology or biomathematics to stress the mathematical side, or as theoretical biology to stress the biological side.[2] It includes at least four major subfields: biological mathematical modeling, relational biology/complex systems biology (CSB), bioinformatics and computational biomodeling/biocomputing. Mathematical biology aims at the mathematical representation, treatment and modeling of biological processes, using a variety of applied mathematical techniques and tools.
http://en.wikipedia.org/wiki/Mathematical_biology

 

[leonAzul]

Just a lurker here, but I am amazed that the false dichotomy presented in the OP has escaped notice among "skeptics" who have continued this thread for nearly 5,000 posts.
(Those who identified this early on know who you are.)

There are radical differences between a fabricated computing machine and a living brain.

Using machines to model consciousness is a valid method, yet it still relies on careful observation of evidence of consciousness as it is, which to date has always involved a living brain in a living body, if it intends to be anything but a gambit for job security.

 

[deadendmind2013]

Yes, the idea that we are wet robots some people find highly offensive, but a fact is a fact regardless of how it makes you feel.

I don't think that's the case. I think pretty much everyone knows that a human being is a form of a sophisticated robot whether they like to admit it or not. I mean come on, this is not the 80's. I also don't think that even the people who claim that we are all "spiritual beings" believe what they 're talking about. They just want to feel like that and for whatever psychological reason they choose to believe it. Plus it pays better. If you're a talented writer, say like Deepak Chopra, it sure does.

But, if I may ask, what is the difference? Regardless of how we call ourselves, our categorical identity does not change significantly.
Even if we were called robots instead of humans, you think that there would not be any religion and wars? There would not be any brands and advanced models and mythical robots, even some messiah robot? There would not be a robot made in Japan, another made in China, another in Canada (so to speak)?

But can robots create robots by reverse engineering?
Yes.
My point is that when that happens, then we will have total equality between AI consciousness and human consciousness. Before that, there can be no clear definition. There is simply no measure. Therefore, consciousness is a human condition until now. Whatever progress is going to happen within 10 years, I believe will be in the sphere of AI awareness and not consciousness per se. That is why I said that human consciousness is fundamentally different. Though, ok, I take back the adv. Human consciousness is different.