Centerfield

But computers have no emotions, simulated expressions maybe but not true emotion.

Wow, you want to take this one up? Well, I think Searles makes a strong case against strong AI. The Turing test shows only that machines can simulate intelligence. It certainly cannot prove conscious awareness, or that a machine “understands” the language it is speaking. It has been awhile since I read debates on the topic and Penrose was a tough last read.

Briefly, I would argue that Penrose is essentially right, though I don't know if quantum effects in organic microtubules are the vital kick of human awareness. It might explain the speed of our awareness not easily explained by estimates of neural transmission. What Penrose is saying however, is that WHAT the machine is made of and how that material is interconnected to Programs, adds a vital quantum effect to cognition that software(formal systems limited by Gödel’s proof) and mechanical hard drives can only SIMULATE. HAL is a long way off, and we should be very careful to distinguish advanced simulations of personality from conscious being. If you argue that monotheism is the ultimate projection of consciousness (that we do every day when we say hello), we should make the same mistake with machines?

Put another way, each human cell has awareness and an active relationship with the environment. Each cell "handles" quite literally the environment following natural rules we don't even fully understand. Each cell is programmed to respond to each other and the world interacting with the subtle physical inputs of the environment. The mechanisms are at the molecular level and possibly the sub atomic level with communication between cells, extremely complex. Explain the handedness of organic molecules.

AI tries to discount the hardware Jon. Hardware is just a means to manifest the program. Just like software alone overlooks the complex hardware of CDMA chips. We are not just a program, but also the interaction of programming with complex matter. Penrose is simply saying that life is a fusion between the miraculous characteristics of matter and complex programs encoded in DNA, which is itself a material undergoing constant transformation, and obeying complex laws We have fewer genes than we thought and this emphasizes the point. To achieve such complexity with fewer genes means that matter, matters. The neural environment, its spatial complexities, the resonances between components and quantum effects all contribute in a way that simple binary circuits and software programs can at best simulate.

Can binary intelligence ever be conscious? Maybe. I once tinkered with a Sci Fi where clouds of floating silicon and elements formed seas of electronic junctions. Our human junctions are far more complex and may well involve “tunneling”. Over time these chip communities evolved into binary consciousness with programs written by natural selection. Yes, a machine can become conscious, but not because it runs a program, but because its hardware interacts in a way with a program that leads to consciousness and emotions. The Turing Test can only show simulation of self awareness, which I think Searles shows. In some small sense, our program in reality is alive. In your view, I could be a machine just simulating Maxtrue and the Turing Test will never know. Combining computers with organic structure like DATA in Star Trek is another matter and in his case his positronic brain was as important as his program.

Perhaps, religion is the belief in consciousness in the absence of formal proof.

Computer says to scientist: “I find your need for “consciousness” quite unnecessary. I think, therefore I am”.

Anyway, interesting topic.

At this point, the difference between a computer and a human brain is really not speed -- for any defined computation the computer is substantially faster already. The difference is in the complexity of the internal interconnections -- the network, if you will. And even an advanced computer network pales in comparison to a human brain. Or even a relatively advanced animal brain.

A good case can be made that a lot of the differences between individuals stem from the differences in the way the nerve cells in the brain hook up as they grow. That's why even identical twins, although genetically the same and raised in the same environment, can develop rather different personalities and different interests. And individuals change as they grow because new connections keep being grown.

Computers, on the other hand, are hard-wired from the outset. (Programming just doesn't seem to achieve the same functionality.) No doubt you could make one which had the ability to add circuits, but to my knowledge nobody has done so. Or even worked out the algorithms for such additions.

So I expect it will be quite a while before a computer which can grown and learn like a human being comes along, assuming it can be done at all. A Turing machine, which can appear to be a human being over the course of an interaction? Sure. But a computer which can grow and learn over the course of years? Not soon.

Computers are not smarter, they are more powerful and capable. Being smart infers that it can think for itself and make judgments on its own. There is no such thing now in a computer nor is it likely to happen.

Computers can only deal in a tangible world. They can only make judgments based on algorithms. Chess is, at its based, an inherently logical and mathematical game. This adapts well to a computer. Bobby Fischer would have trouble playing a computer since he also loved to play mind games.

Computers will never be able to be good judges or politicians. There is a lot of intangibles that do not code well into mathematical functions. One can try; but it is not easy to assign numerical values to something that is not a constant range from one person to the next. You can program base levels and assumptions; but they will never be as dynamic as the human mind because there are areas that do not use logic as a basis. This is where the human brain differentiates from the computer. The computer is a slave to logic. The human brain is not. A computer can try to simulate the human brain and may come pretty close. Possibly close enough that it will be hard to tell the difference. However, it can only surpass the human brain in tangible areas. It can fake it in the intangible areas; but there will likely always be something to flag that it is not human.

The computer is a slave to logic.

Just to be picky, the computer is slave to its programming and construction parameters. You can certainly program a computer to be illogical.

There are computers that can "learn" and computers that can change their programming. Yep, computers with a certain degree of self-programming. DOD would love them on missiles. There are computers which simulate evolution and the results are quite amazing. They create their "neural networks" through a natural selection we impose on them as the "evolve". Do these computers “think”? No, but they point to open architecture and an evolution of computational “freedom”. There are computers that can create music. They might not “appreciate” what they do, but we are talking baby steps here. The point I was making concerned self-awareness and consciousness. That was what Searles was discussing. It is imaginable that several super computers could simulate personality to the point a blind folded individual might not be able to tell a human responder from a computer. Add nanotechnology and holographic memory plus super conducting circuits and there is no reason computers might not one day simulate many things we take for granted as human.

Morality it seems, however, issues from our self-awareness and the value it gives all we think about. Moral decisions and ethical behavior imply the existence of consciousness.

There maybe numerous material systems that might mimic the natural evolution of our organic minds. Why not? Do we honestly think all conscious and intelligent life in the universe MUST be modeled after our likeness? That was the point Penrose included in his search for our brain's uniqueness. He feels today’s computers cannot duplicate the complexity of mind because circuits and programs alone, cannot duplicate physical reality, which he argues, contribute to our consciousness. Silicon based life, solid-state life of greater complexity and other possibilities must exist at least theoretically in the universe. It is quite anthrocentric to think otherwise. It is even possible to imagine that one day; super computers can create other machines that can build flexible circuitry and open system programming -a machine that can refashion itself. Something tells me complexity and "fuzziness" can be captured in non-human systems and it is not impossible to say these new forms CANNOT BE CONSCIOUS. Unfortunately, the Turing Test will not tell us..and I do not know what test will. How do you "know" another is really conscious?

There are already the beginnings of quantum computers and DNA-based computers. We should not be so closed-minded that we think our brains are the only living and conscious machines life (and through our living activity) can produce. The advent of such sentient forms may be premature, but there is nothing in science that says such wonders will never exist, or CANNOT ever exist.

Read Doug Hofstadter's I Am a Strange Loop or Dan Dennett or Marvin Minsky. Basically rather than being something extra special above and beyond animal or computer representational systems, human consciousness is a workable fiction. Humans consistently act like we're conscious so it's a viable working assumption -- but there's really nothing "there" other than a deep, fast representational system. As Hofstadter puts it, the self is a hallucination that hallucinates itself.

When computers get strong enough at representation, they too will seem to be conscious. But you can't just add speed, it's a software problem.

Is that an improvement on "Gödel, Escher and Bach?"

Now I’m just being light but;

What do you mean, "rather than being something extra special above and beyond animal or computer representational systems, human consciousness is a workable fiction."?

First, who said this difference is a radical departure from animals, or that animals aren’t conscious too? Computers yes, but there are some pretty aware animals. They lack the genetic products that give rise to representational systems (language of “words” being one of them ) required for human thinking. That’s a hardware/software problem. Second, the interaction of genetic products and their expression involve more than following software instructions. Try thinking without dopamine or acetylcholine. Perhaps their "definition" of software is a bit exclusive.

And that nasty software problem? Maybe the Gödel Part mentioned in the title of the book above? Representing representation without forms of representation is an interesting trick for a software writer. There are many things that are true that cannot be proved.

Hofstadter might have called "I am a Strange Loop", “i am a Hallucination", or, “i'm more me When i Sleep", but I’ll check it out. I followed the AI debate a while back and I notice the expectations have fallen very short. By now, I would have expected a computer to spot terrorists at an airport, let alone converse with me about the abortion debate.

Its nice to know there is essentially no great difference between me and my toaster oven, save some strings of numbers. I have this suspicion that the physics of cognition are a bit more complicated than is assumed when comparing the brain to software.

Gene expression and the “self” is more that software running on a computer. It is easy to explain human consciousness by denying it, but then Skinner did that with love..

But computers have no emotions, simulated expressions maybe but not true emotion.

What's the difference?

I don't know if quantum effects in organic microtubules are the vital kick of human awareness.

Most brain researchers don't see evidence of quantum effects. Penrose is a little lonely on this one. I think it's unlikely.

As Max says, there is a healthy field of research on simulated neural nets. I wrote a quickie simulator as an assignment for a graduate class, and had a college roommate who was designing neural net hardware. It's fascinating stuff.

A neural net simulator starts with layers of cells, randomized as to connection strength between each pair of connected cells. To learn to solve a problem, we give it an input, and at at how far off the output is from a correct answer. The net looks at that criticism and uses it to improve.

These neural nets have many similar properties to our brains. They learn. They get smarter or stupider as you add or subtract cells. They can recover from cell addition or subtraction, with limited effect on operation. They work with variable-strength connections between cells.

Neural net algoriths are still a research subject. As we understand brains better and better, the models get updated t0 reflect that understanding. The state of the art is much farther along than when I took the class in '90.

My neural net homework solution was pretty small (100 cells?), and not too smart. But when you scale it up far enough, well, then you have your judges and politicians. Note that computers are excellent at several games with strategy implications.

As the book Gödel Escher Bach touched on emotions are one of the ways we solve problems, frustration over not getting a solution tells us we might need to change tactics or even try an entirely new path. How does a computer program duplicate this in problem solving? As far as problem solving goes another question I have is the problem to be solved, in other words we pick the problems we want solutions for, and emotions often guide us. Will AI have a similar capability to look at “its” environment ,make a judgment, and take actions to improve or modify that environment? If an AI system designed a building for itself, would the structure be completely functional, with no design elements that would be superfluous or decorative?

Yes, from what I glean as an outsider, neuronal computing and even DNA-based computing are fascinating.

I wouldn't brush aside Penrose's and other's thrusts however. 1. Ionic communication junctions (ICJ) provide multicelluar organisms A COMPLETELY DIFFERENT COMMUNICATION SYSTEM from neurons. And as one link below suggests that quantum effects ought to focus on intracellular activities and the ICJ nodes of communication, rather than long range communication. 1. There are other communication systems at work in cognition and consciousness, which are a bit invisible to present neural and computer simulations. (see link below on Glia cells) 2.. The "processing" organization of the brain as opposed to neural nets (they are different in the sense of feedback loops, programs of genetic expression, chemistry and hard wiring) are perhaps even more important than the branches on a tree and are not really incorporated in present modeling (so I read). 3.

Even if Penrose is wrong about quantum effects, there are certainly unexplored fine detail of brain function and chemistry that must contribute to awareness and thought. One such aspect would be molecular resonances that seem to coordinate wave patterns of neural firing. Even if we could formalize neural activity to numbers, the events involves materials with properties, signals with gene expression not wires, software and chip lasers. Despite these things, I am not suggesting that reductionism might lead to sufficient modeling to produce invention that is self-aware and conscious. I’m not sure I would trust moral decision to such a form, though their are many humans I wouldn’t trust either.

Anyway, an interesting thread and perhaps you could update it with the cutting edge news involving AI and neuronal nets. I leave the following threads for anyone interested.

Explaining consciousness -an interesting overview of articles

The problem Penrose and others approached in the early nineties

Barking up the wrong tree? Look at the trunk

A response to criticism

the complex variables presently unknown about actual components of biological neural nets

some on line reading concerning consciousness if anyone is still awake after reading all this

And if Rachel Ward was just a machine, would it matter if she dreamed? Not to me....LOL

Jon,
I wonder if it would be too much to add something here on your thread. You started me thinking and although the following might not hold interest, forgive the length. I think what follows has a lot to do about centrism and the political extremism we see now as the NYT reports this morning that antiwar is ratcheting up pressure on the Democrats to reverse administration foreign policy. I am going to show a connection between materialist AI advocates and political conflict.

Towards a New Understanding of Spiritual

Back in High School physics class, I wrote a short essay called “Supranature, the End of Modern Dualism”. Little did I realize I was a budding Liberal Naturalist . I had read a book called Supernature and it started me thinking. In my paper, I advocated that modern scientific thinking (and public opinion) was divided into two camps; the materialist (atheists) and the dualists (religionists who also accepted the truths of science). Of course there were religious fundamentalists, but naively, I thought their days were numbered. The fact that three Republican candidates now profess a belief in creationism, shows me how wrong Orwell and I were. I do have a problem with a C and C accepting the literal interpretation of the Bible, whether Jewish or Christian. I did not think America or the world would witness a rise in both extreme secularism and religious fundamentalism. Back in the seventies, I believed attempts to explain human reality be reductionism (Skinner and other scientific thinkers) or Dualism (scientists who believed in religion AND science) would eventually be transcended by a new physics. The Tao of Physics and other best sellers advanced “new thinking” about the laws of nature. Prigogine and others pointed towards an intrinsic characteristic of materiality for self-organization and self-referentialiy. Many scientists began to explore chaos and declared there is a fundamental statistical nature in phenomenon itself . Although my understanding of formulas and equations were limited, I sensed what they were suggesting.

Today, the same lines still divide us, but let me try to sell you a used car. Material secularism and Dualism still are pitted against each other. Some scientists believe that the laws of nature as understood now can explain cognition, human behavior and perhaps even awareness. Other scientists believe spirit, consciousness, limited volition are fingerprints of God and will never be explained because they constitute something “beyond” materiality. This is what I call modern Dualism. Some Liberal Naturalists offer a middle path, which like centrism seeks to unite valid elements of contrasting extremes. Let me suggest that one can be spiritual but still believe that spirit, consciousness and will are an emergent characteristic of complex systems pushed far from equilibrium. These qualities reflect the inherent properties of matter and constitute more than a “workable fiction”. I understand a certain Gurdjeffian cynicism, but not even he precluded the ability to become truly conscious and willful in his attempt to enlighten the world with his teachings. Mysticism can still be mystical if it issues from the power of yet undiscovered Natural Laws. Remember, Newton was above all, an Alchemist. Gravity, to him was one of the alchemical laws driving the majestic universal machine. If science doesn’t recognize this new physics, it erodes an essential usage of words like, free will, consciousness, love, morality and ethics. I helps those on the Right to claim secularism and materialist make a mockery of the use of those words. What is the case for morality when essentially their is no spirituality, love, consciousness, or that these things don't really mean what the dictionary says they do? Is love just a workable fiction?

An atheist can be moral, just, loving and conscious. These words are not the exclusive province of the religious. There are in fact, more incarcerated people who claim they are religious than those who say they are atheists. An atheist can be spiritual and believe in certain mysticisms. Many atheists, who feel this way, call themselves Agnostics with the caveat that God can never be a force that would negate the beauty and reason of existence. One can be an agnostic or atheist and believe all is material under the notion of Supranature. We don’t even know what all that Dark matter is out there. I suggest that many in the AI debate are hardcore materialists who would rather marginalize consciousness and love in there reductive attempts to explain everything. I see the same sentiments echoed on the political Left. Perhaps Liberal Naturalists offer some hope to the conflict between the Left and a Right that still has faith woman came from Adam’s rib. The three greatest thinkers in the West’s recent history, Newton, Darwin and Einstein, all believed in a “qualified” God. Not one was reduced to reductive materialism nor prone to wild fundamentalist theology. Perhaps they too were centrists and walked a middle path relative to the culture of their day.

Damn, footnotes don't post. Anyway

1. see first link in my last thread for "Liberal Naturalists".

2. see works by Progogine

From earlier posts;

"Something tells me complexity and "fuzziness" can be captured in non-human systems and it is not impossible to say these new forms CANNOT BE CONSCIOUS." should read;

"Something tells me complexity and "fuzziness" can be captured in non-human systems and it is not RATIONALLY POSSIBLE to say these new forms CANNOT EVER BECOME CONSCIOUS."

"Despite these things, I am not suggesting that reductionism might lead to sufficient modeling to produce invention that is self-aware and conscious" should read;

"Despite these things, I am not suggesting that reductionism WON'T lead to sufficient modeling to produce invention that is self-aware and conscious"

Given my less than computer-like proof reading, viwers beware of my propensity for typos which often produce opposite meanings....LOL

A computer wouldn't make these mistakes, but to err is human and self-correction divine.

Interesting questions, Max.

Yes, I'd agree with your rough classification. One interesting question is whether the recent fractal theories, improved vision of the cell as DNA-driven machine, Hawking's theories, and widened consciousness of power of automation have changed the numbers of atheists and Blind Watchmakers lately. Or is about the same as ever?

I was watching a Nova last week that suggested that Newton was a literalist. They interviewed somebody who claimed he spend lots of time on his own version of a Biblically short history.

Oh, and a good simulation of you WOULD make the same mistakes. Neural net simulations definitely include mistakes as an important feature.

The PBS on Newton was very interesting. He was searching for that Philosopher’s Stone. He had to keep his views about the Trinity secret. The reductionism /vitalism debate is very old. I think "new physics" offers a way out by defusing most dualism and allows atheists the same clout as religionists with words like moral, love etc. It is interesting that researchers are trying to find ways to incorporate elements of physical reality into software and hardware such as fractals, genetic language, quantum effects etc. You can see the complexities of the biological neural net in the rebuttal response by Penrose's partner. The body also produces crystalline substances that might provide some workable environment for quantum phenomenon. It is also amazing the recursive and disparate levels of communication in the body that contribute to cognition.

As I said, atheists can be spiritual and mystical. Some atheists believe there is a conscious field in the universe that perhaps our energy flows into when we die. Anything is possible provided the Divine is material and doesn’t violate the Laws of the Universe. Others see alternate universes; higher dimensions etc. as presenting a wonderful and unknown field where just about ANYTHING is possible. What atheists object to is the description and directives of our human notions of God and the existence of the immaterial. Just because language allows us to formulate sentences, does not mean such sentences have meaning. Our faith others are conscious does not mean we can project this concept into a God that actually exists outside our language game. Once atheists and the religious can agree on a ground of shared morality, half the battle is over. Instead, we see extremism pushing ideology into opposing camps with Salvationist and Liberation philosophies opposing each other.

I don't doubt that humans will one day build something that is actually conscious or self-aware. If Prigogine is right, there is an uncertainty in everything that is the cause of our limited free will. Even some existentialists require this uncertainty to explain how out the future remains indeterminate. Of course, human stupidity centers on our denial of how much like machines we are. Most of us spend most of the time living in "workable fictions" of self. We declare negotiations and intentions can change behavior. Let’s wish and it will be so. Such is the delusions of human machines.

I am interested in reports of natural selection simulators creating "computer species". That is right out of a Sci Fi I can't recall clearly. A scientist goes to an Island and "evolves" his little creations into a world altering result. The more we can get software to reflect elements like chaos, parallel processing, open architecture, self-modification and adaptation etc., the more the results will shock us. We are the organized reflection of complex materiality in search of more powerful mirrors.

Is it Monday already? LOL