Shadows of the Mind
Every year, I try reading at least one "epic" work that challenges my thinking processes. Last year, it was "On Intelligence" by Jeff Hawkins. That was a great work emphasizing a new approach to understanding and perhaps simulating intelligence: rather than being the result of computational procedures, intelligence is actually based upon the retrieval of stored memories that are utilized to make forward-looking predictions.
This year, I've been working through "Shadows of the Mind" by Roger Penrose. Like Hawkins, Penrose has little faith in computation (either human or computer) in order to understand such attributes as human awareness, consciousness, and intelligence. The difference is that while Hawkins lends some hope for a computer simulation of intelligence, Penrose explicitly rules it out. In fact, he mathematically proves that there are cases where an algorithm can't be proven, but the human mind can comprehend its inner workings and end results (see pp 74-75). He demonstrates this with a procedure that points to an algorithm that allegedly never halts, and thus continues forever. What he shows is that there are cases where the procedure simultaneously indicates the algorithm will halt and will not halt, hence inducing a contradiction that can only be resolved by admitting it continues forever.
Penrose discusses the famed Gödel Incompleteness Theorem and it's implications for formal systems established for proving mathematical propositions. He states that Gödel showed that the consistency of the formal system can't be proven by the rules or axioms of the formal system itself; that in effect, the proof of this consistency is external to the formal system. This is different than other interpretations I've read about his Theorem, namely that there will be at least one general theorem, based upon those rules, that will be true but can't be proven. In any case, he discusses these implications for computer intelligence and robotics and concludes that they can only be programmed with human directives and can't "think for themselves".
Step by step, Penrose dismisses any linkage between algorithmic computation and human awareness. Regardless of whether the system is sound or unsound, or is known or unknown, he is steadfast in his conviction that it can't be simulated. I have to say that I have some reservations about his claims. I'm not convinced with the complete soundness of some of his arguments here (have I myself fallen prey to Gödel's Theorem?!) and feel that he might be missing some logical cases that would weaken those arguments; but I myself wouldn't be able to construct a counterargument to them. Nevertheless, I'm impressed with his general line of reasoning -- he's a brilliant thinker and theoretical mathematician.
Penrose classifies 4 types of views of human awareness, consciousness, and intelligence:
1) These attributes are the results of computational processes. This is the strong view of Artificial Intelligence (AI).
2) These attributes can be simulated by computational processes. This is the weak view of AI.
3) Science can describe these attributes, but they are neither the result of nor can be simulated by computational processes. This is Penrose's position.
4) Science can't even describe these attributes. This is the "mystical" position, and Penrose claims that Gödel himself falls in this camp. He argued that Gödel's Theorem might point to the existence of a "mind" distinct from the brain, outside of the formal system!
The second part of the book, which I haven't read yet, starts with the assumption that consciousness can't be described in classical terms. Rather, it's based upon the indeterminism of quantum mechanics. Consciousness is the result of the collapse of quantum wave functions and the superposition of these events! He argues that these events occur in cellular structures called microtubules.
Well, I'll have more to say when I'm finished with the book. But for now, I recommend you read this book. It's more difficult reading than "On Intelligence" because of the formal arguments and mathematical reasoning presented here. Feel free to comment on the subject matter.
posted by David Epstein @ 7:46 PM
12 comments
12 Comments:
I have a lot of respect for Roger Penrose, and highly value his thought on this and other topics. I do, however, have a lot of respect for Jeff too, and do believe that HTM technologies are a remarkable pattern recognition engine, and are capable of so much more than people give them credit for. With the release of the software package for Windows, I think we'll see some interesting solutions come out of this in the next year.
It seems as if Penrose is saying that the reason computers can't think like humans is because thinking depends on quantum effects. Computers are made of matter, as are humans. It seems to me that quantum effects are possible for both.
Joe, I think what Penrose is contending is that human consciousness and thinking are based upon biological-quantum processes, not merely quantum ones. These quantum effects (i.e. quantum wave collapse, superposition) are occurring within and across cellular structures called cytoskeletons, specifically inside the microtubule components of these structures.
If his theory holds up, then there is the question of whether these effects could be replicated in bio-computers. The broader implications of this would revolve around the emergence of awareness, consciousness, thought, and intelligence from such artificially created quantum processes. Of course, these quantum processes are indeterministic by nature, so whether they can be "replicated" is the fundamental question!
I haven't read the 2nd part of the book which discusses the quantum theory of consciousness, but I would imagine that based upon his assertion that human awareness is not based upon any algorithmic computations, nor can be modeled by an algorithm, he would rule out the algorithmic creation of the indeterministic quantum processes within a bio-computer. I'm thinking he would contend that consciousness is a unique biological attribute that isn't deterministically created, but naturally emerges out of biological existence.
I just had a thought. There are 2 issues addressed in my last post:
1) Regarding the possibility of artificially creating an indeterminate process, is there an algorithm (deterministic by nature) that can simulate an indeterminate process? Conventional wisdom would say "No".
2) Human awareness or thought can't be modeled by an algorithm, nor is the result of any algorithmic computations. This is Penrose's contention and he proves at least one case where this is true.
My thought is that #1 and #2 might be related: an indeterminate process can't be deterministically created; hence, an indeterminate stream of processes like conscious awareness can't be deterministically created or simulated.
But perhaps there are other types of algorithms (i.e. neural networks, genetic algorithms, combinations, or undiscovered ones) that can learn, adapt, or evolve to such a point that they can trigger their "own" process that spawns the indeterminate processes that are desired (the quantum "effects" previously mentioned). If so, then we would find a negation of #1. This negation would somehow have to lead to the negation of #2 (i.e. consciousness can be created), or the negation of #2 would have to emerge from the negation of #1.
Let me just address one issue here: The idea that atoms in organic organizisms have different quantum properties than atoms in inorganic organizisms sounds like vitalism. I rather think that life and intelligence are emergent properties which arise out of the complex configurations of elementary particles which have the same properties whether they are part of a living organism or not; or for that matter even if the atom goes from being part of an inorganic body to being assimilated by an organic body. I think it is an unnecessary complication to suppose that there are two irreducible types of quantum particles, one type for living and another for non-living things.
Regarding in/determinism I would like to think about it a bit more; but I know there are pseudo-random number generators which appear to the user who doesn't know the internals of the program to be random; but the ones that I have read about are all cyclic: the output is a repeating very long list of numbers. Chaotic systems are deterministic, but the result of the next iteration of a point in an almost period part of the map will not repeat as the pseudo-random number generators will. There are also problems in these systems arising from our inability to measure exactly and from the way a chaotic system is defined (SDIC, etc.)
The computational model criticism is more interesting. In "The Fabric of Reality" David Deutsch uses an adapted version of Cantor's diagnolization argument to prove there are virtual realities that no universal Turing machine could compute. I'm not sure I'm ready to talk about this.
Another thing I am not ready to talk about, but find interesting is something Ray Kurzweil says in "The Singularity is Near". He claims that in order to find the organizing principle behind evolution we have to see beyond complexity to the idea of order. The order he is talking about seems to be the order found in thermodynamics, or better yet, information theory. He distinguishes between randomness and unpredictability. It's over my head for the time being, but interesting as can be.
I don't think Penrose would argue that there are two different sets of quantum particles, one for organic and the other for inorganic organisms. Rather, he would argue that the quantum coherence underlying consciousness arises as a result of an organic process rather than an inorganic one.
He also addresses several of the issues we're raising: randomness, pseudo-random generators, chaos, neural nets. He rules out all of these (and other) phenomena as a possible "creator" of computer awareness or intelligence.
Regarding Cantor's diagonalization, he also uses Cantor's techniques (the diagonal slash) to prove that there is at least one case where an algorithm is understood to be true, but can't be proven to be true.
The principle of order in evolution is not understood too well, but I would agree it's one that should be explored. I see more hope in theories like Stephen Jay Gould's Punctuated Equilibrium leading to an eventual order rather than some gradualist or implicit order existing in nature.
By the way, I'm using the ideas of "consciousness", "awareness", "understanding", and "intelligence" interchangeably in my argues. Penrose himself, however, gives careful treatment in defining and utilizing these properties for his own arguments.
I'm having a problem with information overload right now (or maybe it's laziness, not sure about that). I am reading "Grammatical Man", "Godel, Escher and Bach, The Eternal Golden Braid", "The Singularity is Near", "Visions: How Science Will Revolutionize the Twentyfirst Century" and, when I have time "Harry Potter and the Goblet of Fire" all at the same time. Let me make an exception to that, for the last couple of days I have been looking up articles in Wikipedia like "thermodynamics", "entropy", "free energy", "entropy and life", etc. Man, that will soak up some time. I'm intrigued by this idea of measuring order as opposed to measuring complexity, the goal being to be able to compare and contrast the fitness or adapability of two different organizations (or organizims or structures or formations) in terms of problem solving ability or fitness, or perhaps as a measure of which of the two systems is more evolved. I'm getting responses that I think are obfuscations, like "be careful not to reify something that is not really there" and "don't talk about that, that's social darwinism". If there is something real there, (and I think there is. Imagine if there was no such thing as order) then it seems to me it can be either beneficial or harmful, depending on how it is used. Also, to neglect exploring a potentially powerful idea (not that I saying that this is necessarily a powerful idea, but rather that censorship ought to be avoided as a bad practise) gives up first rights to the idea to the one who exores it first, and I think it is negligent to not things about the world that might be of benefit to all (or at least to many).
Going back to the thread of my previous post, I think that if Penrose is saying that life, consciousness and intelligence are prperties that emerge from the organization of matter, then I think that is preferable to vitalism or somthing like vitalism. But.. how and where do the quantum effects come into this? One thing that puzzles me about this is the reductiionism/holism debate. In holism the interactions of many potentially simple agents of a system affect the whole system bringing into existence a new thing, and the properties of that new thing in turn feedback upon and alter the properties of the agents the system is made of. However, quantum particles are supposed to be, in some sense, reductionist building blocks that everything else is made of. Darn, class is starting, gotta go; but the above could be considered to be a reasonable request for more information, if not a well-formed question.
Included below is an email conversation between David and myself. I took this off-line because I wanted to test the waters and see if I had crossed the line with some of my comments, and thought it best to do it off-line. David suggested I post the email conversation on the blog. so here it is, unedited, spelling and grammar errors and all and not in chronological order:
If you'd like, you could post this to the blog. Either as a response to the current blog entry, or just create a new one.
Joe Cash wrote:
Good reply!
I understand about not having time, but I also know that intellectual discourse is an appetite that needs to be satisfied, too. I think you will be coming back for more sooner or later.
Having a conversation like this off the blog seems kind of like “cheating” somehow, but the previous email was an aside because I wanted to see if I had stepped over a boundary, and didn’t want to do it on the blog. Under the current circumstances responding to your email on the blog would not make sense to other readers because they don’t have the context of the email to explain what we’re talking about.
We’ve been talking about these things, so I have been cruising the web and finding things that are relevant. One of the things I found is a description of Social Darwinism that states something like the following: The strong survive, the weak should perish. I don’t like that belief much. Everybody is weak at some time or another. Doesn’t seem very human. On the other hand, it does seem to me that it might be useful to have a way to determine, given two individuals or two organizations which of the two contains more order. Note that the one that contains more order need not be more complex, at least according to Kurzweil. A simpler organism that solves a problem more efficiently than a more complicated organism would contain more “order”. Given two businesses with different organizational structures, is there a way to measure which is more ordered so as to predict which will be more efficient? How about for organisms? Entropy and information seem to be leading candidates, now I am trying to understand them better. When I brought the subject up in the computer lab, a teacher in the math dept. said is sounds like social Darwinism, as if to say it is topic that should be avoided. That might be my interpretation and he might not have intended that at all, but I do live in the Central Valley.
I do have a little difficulty with the idea that the Nazis would have been evil even if they did not do anything violent or oppressive. I think we all have bad thoughts, it is acting on them and/ or justifying bad action (i.e. believing they are right and would be right even in an ideal world where there are no competitive pressures forcing us to compete to survive) that I think makes a person bad. That being said, while I hate war, I think I would have considered World War II a just war worth fighting had I been alive then. The Nazis were bad.
I was not arguing that Nazis did some good things, I was trying to say that not all things Hitler did should be considered to be bad simply because they are associated with Hitler. Hitler breathed; that does not make breathing air bad. It was the violence that makes a person bad. Problem now is, coming up with a really good definition of the word violence.
Well, fortunately for you, I’m short on time, too. So, I’ll sign off for now and talk with you later.
Joe Cash
Lecturer, California State University, Stanislaus
Office: P284
Phone: (209) 667-3183
From: David Epstein
Sent: Sunday, September 23, 2007 2:22 PM
To: joe cash
Subject: RE: [Idea Symposium] New comment on Shadows of the Mind.
Hey Mr. Joe,
I was attending Yom Kippur services the last couple of days, so I didn't have a chance to reply earlier. The only thing I'm offended by is you didn't wish me Happy New Years (L'Shana Tova). ; )
I saw you're last blog entry but haven't had time to reply. It's getting harder to find time. It will be harder once the baby arrives in 1 month! Oh well, so enjoy the time now perhaps.
I see by your blog entry and this email that I better break this down into topics:
Topic 1: Order behind/besides Complexity
You mentioned in the blog you were interested in the subject of Order behind Complexity, or apart from it. That appears to be a worthy topic to pursue. My own thoughts are that Complexity is a pointer to any prospective Order. The intuitive relationship is that the more complex the system, the greater the difficulty in achieving or sustaining the order. But there are undoubtedly more subtle and hidden factors to consider. Like the speculative 5th forces of nature (e.g. supercharge), there might be some self-organizing factor that more readily promotes order in more complex systems.
I think this could be a New Topic on the blog.
Topic 2: Darwinism and Fitness:
Regarding "fitness", in a strict Darwinian respect, this has to do with the ability of the species to reproduce. It has come to represent other factors like "skill", "ability of accomplishment in a field", etc. Obviously, an individual of a species is more apt to reproduce if s(he) is able to survive and adapt to a changing or challenging environment. The great debatable claim is that a more fit being will be able to more readily adapt than a non-fit being; that is, the fitness, or propensity to adapt and perform, is already ingrained in the individual. I happen to believe it's a combo of nature and nurture, but this would be a long-winded discussion and there's already been a ton of reading material written about.
But in regards to non-survival fitness factors like "skill" and "ability", an individual can fail in one area and then succeed in another, or could even re-evaluate his approach in the area he failed and then succeed. Insofar that such "re-evaluations", more studying, reformulations, etc, are modes of adaptation, then it's absolutely imperative to consider these as "fitness" factors in an evolutionary system. We are more likely to call these "standards". Well, what's wrong with standards for determining achievement levels? The point is these standards are implicit in both survival areas (i.e. requiring food, water, shelter, etc) and non-survival ones (achievement in a field of study).
So fitness and adaptation are 2 of the primary factors of Darwinism. And the 3rd is ... variation. Variation of the species is evidence of an evolutionary movement, because mutations in the gene pool, reproduction, changes in the environment, etc, will result in such variation in the species.
Topic 3: Social Darwinism (and Morality):
I'm certainly no scholar on the topic, but my take is that Social Darwinism (SD) enters the picture when we consider Darwinism in the social realm. Darwin didn't deal with this topic, but other contemporaries like Spencer certainly did with his ideas of "progressivism" and "survival of the fittest". Malthus to some extent dealt with it. From what I was reading online, it wasn't until after WWII that the movement took off.
The point, however, is whether one theory of evolution (Darwin's), which primarily deals with the evolutionary *biological* movement of a species can be translated or transplanted to the *social* realm. Is it feasible? Is that transplanted operation a scientifically sound one? It's similar to trying to apply Quantum Mechanics concepts (sound ones at the micro level) to our macro everyday world.
The other issue is what is the interplay or separation of any SD forces and Morality. Are moral actions the product of evolutionary forces, and if so is the evolutionary directional from "bad" to "good"? Well, if that was the case, then we were born "bad" just like the advocates of Original Sin believe. But unlike the Fundamentalist remedy to become good by "accepting the literal word of God", the moral-evolutionists would probably argue that we've become hard-wired to aim for the good, and that those who fail that test of attempting to achieve the good, or who appear to naturally aim for the bad (i.e. evil), are bound to be unfit. According to this theory, the struggle between good and evil is itself subjected to the forces of evolution.
Insofar as societies, cultures, nations, civilizations, etc, are able to adapt to a changing, challenging, and sometimes hostile environment, they are more apt to survive. I accept this part of SD to be self-evident. A society that is under attack must learn to defend itself. Self-defense is a mode of adaptation, and also a method of adoption. Adoption is a tool to be utilized once the choice is made to survive.
What I don't accept with what I know about SD are any "apologies" or "excuses" about evil acts, that atrocities should be seen as the end results of deterministic socio-evolutionary forces. SD proponents (most anyway) would undoubtedly be against those acts, but would be simply explaining those acts in what they deem to be a scientific lexicon. But is SD a science? Is it a branch of socio-biology, political science, anthropology, a combination of these? Or is it itself a socio/political ideology? Perhaps it's nothing more that pseudo-science disguised as legitimate scientific inquiry.
Do moral principles exist outside of the boundaries of evolutionary forces? Are our yearnings, actions, and choices to do good in the world, and to fight against the bad and evil that lies therein, merely survivalist instincts and/or merely to adapt to a hostile environment? I think we would agree that there is something greater to life than that, regardless of whether we interject God into the picture!
This too could be a separate blog topic of course. But I've set up the blog as a forum to spur on new ideas, understanding their implications, technologies, ...
Topic 4: Social Darwinism, Naziism, Capitalism:
Yes, you're right that Naziism is often associated with SD. We could even view the victory of the Allies over Naziism and Imperialist Japan as the end result of SD evolutionary forces. But SD is associated with Capitalism as well (some people argue that Naziism was a corrupted implementation of Capitalism or Corporatism). The theory is that unfettered, free competition between individuals and societies will promote a better socio-economic order than any centrally planned society. Competition will weed out the misfits and the strongest will survive and we'll all be better for it. That's the reduced core and crux of SD as it applies to Capitalism (of course the proponents of SD will say we're over-simplifying it).
Topic 5: Can Evil people do Good things:
You're also right that Hitler (and the Nazis) was/were primarily evil because of his/their violent acts. However, they were also evil because they espoused a racist, anti-Semitic, anti-gay, anti-Gypsy, anti-freedom, pro-"Aryan" ideology, and were in power to implement this ideology. Even if they didn't commit acts of mass-murder, persecution, disenfranchisement, and violence, they were still evil.
Regarding whether or not they did any good things. They did good things, but for whom? Who benefited from them? The entire Nazi society was founded upon unconscionable evil and injustice. Can any good come out of that? Yes, that's a good philosophical issue to discuss, and yes, the alleged good products & services that arise out of that evil could be applied to a "good" society.
You mention vegetarianism as something that's not necessarily evil. Who would consider it to be evil, other than doctrinaire carnivores who won't eat any vegetables?! Even nutritionists who argue that a vegetarian diet doesn't provide enough proteins or essential vitamins wouldn't argue that vegetarianism is evil!! ; )
Anyway, just joshing you here. We would both agree that vegetarianism might very well be a good thing. But was Hitler himself truly a vegetarian? Google on the topic and you'll find evidence to support that he occasionally ate meat (ham, squab).
And the Volkswagen? We could debate how good of a car it actually is. Regardless of its alleged quality and affordability, it was built with a great amount of slave labor. Can something good be built upon an evil act? Goes right to the heart of that classic question: Do the ends justify the means.
Topic 6: Censorship:
Not sure how this topic arose. Who's censoring whom? Were you feeling censored by the person who said that your discussion of "fitness" is Social Darwinism, and you shouldn't talk about it? How is that censorship? Was he preventing you from discussing that position? Could you not have responded that "if we don't freely discuss this idea, then we're possibly inhibiting the promotion of ideas that could benefit individuals and societies", or "if SD is in fact an insidious ideology, shouldn't we openly discuss it so we can formulate a rational response to combat it?"
Sometimes, we just can't get this open discussion we seek. Other times, we don't have the time. We don't have the means or time to discuss these issues with everyone we meet. So we must carefully choose our discussion partners and what topics will bring us the greatest return on our investment of time & energy.
Your idea of censorship putting people with no conscious in power is well taken. I feel we're living in such a period right now. Can we truly debate the policies of the current administration? If we discuss the War in Iraq and question the wisdom of waging that war, the response we get is "This is a time of war and you shouldn't criticize the commander in chief in a time of war." Can we then respond that "You chose to fight this war, that it wasn't a war of self defense, and that your claims that it was out of defense (i.e. the weapons of mass destruction) were at best unfounded"? Then if we criticize the large amounts of casualties over there, their response is "They're fighting over their for your freedom, for you way of life". Can we respond that "No you're not. They're fighting for your misguided policies"?
Regarding some forms of censorship being good (i.e. laws prohibiting certain behavior), that too is well taken.
Topic 7: Potential Blog Topic (Justice Systems):
Sure, why not. Feel free to start a blog topic on it. Designing a new justice system is not necessarily admitting that the current system has failed en mass, but rather that it needs to be re-designed to serve the needs of our current society. Also, it could be a good theoretical exercise which could be useful at some later time. For some reason, I was thinking this is akin to building a new software platform for supporting existing applications.
Unfortunately, my time is limited so I don't know how much I can contribute; but I will always try responding to your postings or emails.
Topic 8: Dr. Wayne Dyer:
Is he for real? I'm listening to him on KQED as I'm typing this email. He used to be Mr. Pop-Psychology. Now he's actually talking about Taoism, coping with current events (i.e. the war), and a lot of other useful areas. Who would have thought?!
Take care,
David
Joe Cash wrote:
My friend David;
I have been thinking, and I hope I have not offended anyone (especially you) with my comments on the blog about social Darwinism. Please allow me to explain: I think social Darwinism is associated with Hitler (though Francis Galton and the founder of Planned Parenthood were advocates before Hitler took it up.) While Hitler and his ilk have done horrible things, not all the things that Hitler did are evil. Vegetarianism does not have to be evil, nor does making a more affordable car like the Volkswagen. Hopefully walking and breathing are not evil, yet I am sure Hitler did all these things. What Hitler did that was evil was resort to violent acts. That is what should be focused on when criticizing Hitler, in my opinion. Talking about or researching social Darwinism is not evil in and of itself, nor is it evil just because Hitler did it, or even because it was an essential part of Hitler’s plans. It is the violence that made Hitler evil (IMHO).
To say one should not think or talk or write about something is a bad thing. It’s censorship, it’s oppressive (it may even be a violent act?), and it is negligent in the sense that letting others be the first to acquire power because they were the ones who were free to discuss and explore it while others are prevented by censorship or approbation is a dangerous thing; it may put the power advantage in the hands of those who have no conscience.
To be frank, I don’t know much about social Darwinism. What I was talking about in the blog was the idea, which I had read about, that order (thermodynamic negentropy or information) could be used as a measure to determine which of two species is more evolved. A teacher in the math dept. here at CSU suggested that I was talking about social Darwinism; and I guess I reacted to that on the blog, even though the blog had nothing to do with this math teacher’s comment. My apologies.
Expanding on the idea that censorship is bad, I might make some exceptions to the idea that all censorship is bad. For instance, the law against yelling “fire” in crowded public building where there is no fire is probably a good law. I also think making serious verbal threats to hurt someone physically or conspiracy to do the same might well be discouraged, and it seem like breach of contract is violence to property rights. To me this line of thought begs the question: where does a person end and the rest of the world begin? It seems to me that violence is done to a person or an individual. Is property then a part of that individual? If not, then can we truly claim theft as being a violent act? Theft of, e.g. a garden hose laying on a front lawn does not hurt the body of the owner, so does the owner have the right to use force “defend herself” against this “violent” act, or should we consider theft to not be a violent act?
Let me take a step back here to talk about a potential blog thread. How about the idea of starting a discussion which analyzes current justice systems and attempts to start from scratch and discover or design a justice system based on basic principles? The reason I suggest this is because I think there are many injustices in existing justice systems and talking about this might help us in our political decisions. Do you have any interest in that? What do you think are the chances that something like that would draw any interest or replies?
Well, I’m rambling now. I must be getting tired. I’m going to go have a weekend. Hope yours is a good one.
Joe Cash
Lecturer, California State University, Stanislaus
Office: P284
Phone: (209) 667-3183
________________________________________
Tonight's top picks. What will you watch tonight? Preview the hottest shows on Yahoo! TV.
________________________________________
Moody friends. Drama queens. Your life? Nope! - their life, your story.
Play Sims Stories at Yahoo! Games.
Going back to the "Tue Sep 18, 02:18:00 PM PDT, mitigator (Joe Cash)", and replying to specific points:
* "But.. how and where do the quantum effects come into this?"
By Penrose and Hameroff's theories, I would say these effects are occurring within the cells themselves, specifically within the microtubules. I haven't read enough of the 2nd part of the book to say for sure, but it looks like superposition and even quantum entanglement are at play to bring these effects into existence.
* "One thing that puzzles me about this is the reductiionism/holism debate. In holism the interactions of many potentially simple agents of a system affect the whole system bringing into existence a new thing, and the properties of that new thing in turn feedback upon and alter the properties of the agents the system is made of. However, quantum particles are supposed to be, in some sense, reductionist building blocks that everything else is made of."
Yes, this is puzzling. How do the interections of particles bring about the holistic functionality of an organism? Where does the forces of emergence, if they can be described in these terms, enter the picture? Again, I would imagine that the answers lie with the phenomena of:
1) superposition (i.e. the "overlapping" of quantum events, or more technically the addition of amplitudes of wave formations from interfering events).
2) quantum entanglement (when 2 events are interrelated even though they're spatially separated).
I did not know about microtubules before I read your response, so I thank you for bringing them to my attention, but I still don't see how this justifies Penrose's assertions that artificial intelligence is not possible (I can't quote him, but I think that is what he means).
If a structure such as a microtubule in the cytoskeleton of a cell can exploit these quantum effects, then why can't people (or some other sufficiently equiped beings, for that matter) build artificial structures to imitate those natural structures which will also exploit the quantum effects in a way similar to the way the natureal structures exploit the the quantum effects? I think this question points to a principle that needs to be answered by anyone who supports a argument that AI cannot be achieved because of obstacles of a materialistic nature. That is to say, if nature can exploit these physical properties of matter, then why can't an engineer build something that also exploits these properties? Even if it is just a molecule-by-molecule imitation of the organic structure, it is still artificial. Once said imitation exists, it proves that it can be done. At that point, maybe engineers can improve upon it.
The argument from computibility seems to me to be more interesting. I don't understand well enough what is and what is not computable,but doing a little research I found that the set of all computable functions has a denumerable number of elements, while the set of finitary functions on the reals has an uncountable number of elements. I have read the finitary functions on the reals can be approximated to any degree of desired accuracy by a turning machine. The halting function is also not computable. So, functions that are not computable do exist. Are non-computable functions necessary for intelligence? I don't know. I don't think we can know until we have a definition of intelligence that a "quorum" (so to speak) of scientists can agree on.
But, that might be a moot point. If an artificial entity can continue to function independently long enough to reproduce several times with a proper amount of variation, then the entity will probably be able to evolve. Give it the ability to repair itself independently and it can probably improve itself during its operation (it's "lifetime"). Non-computable functions are probably not necessary to create such an entity. It would probably have some intelligence (however we define it) built into it in order to accomplish its survival, repair, self-improvment, and reproduction. If the offspring of this entity do evolve, then the I.Q. of some of the offspring will probably increase from generation to generation, or even within the operational "lifetime" of a single individual of this "species". Under such circumstances, being able to repair and improve itself, an individual of such a species may be able to operate indefinitely, allowing it to enhance its ability to improve itself over time. Some people have suggested we define intelligence in terms of ability to survive. If that works, then such a species may be considered, according to this measure, to be more intelligent than humans. They might become competitors for the human ecological niche. I suspect Penrose is (consciously or unconsciously) objecting to the idea of AI to avoid such a scenario. If he is wrong and we belive him, then we may be stumbling into said scenario with our eyes closed by self-delusion. I think we need a better solution.
JJC: "If a structure such as a microtubule in the cytoskeleton of a cell can exploit these quantum effects, then why can't people (or some other sufficiently equiped beings, for that matter) build artificial structures to imitate those natural structures which will also exploit the quantum effects in a way similar to the way the natureal structures exploit the the quantum effects?"
DE: There are 2 obstacles that need to be overcome here:
1) According to Penrose, human awareness (and intelligence by extension) can't be modeled at all. It doesn't matter if intelligence is the result of classical or quantum processes.
2) Even if awareness/intelligence could be modeled, a quantum-rooted process is indeterministic and thus there's no assurance that it could be artificially replicated.
Regarding #1, Penrose doesn't rule out scientific explanations of human awareness. He rules out deterministic algorithms that would either generate or model this awareness. Such an awareness arises out of conditions we don't understand, not yet anyway. So "exploiting the effects", either classical or quantum ones, presupposes that there can be a deliberate approach (algorithm) to deterministically steer the system to a desired end result (the emergence of intelligence). Penrose rules this out.
One reason he disregards robotic intelligence has to do with Godel's Incompleteness Theorem. This necessitates the existence of the theorem proving the consistency of a system lying outside the bounds of that system itself. Thus, the robot can be programmed with all types of rules and directives, but the consistency of this programmed system would exist outside of the realm of its "awareness" (humans, on the other hand, are not *aware* of this limitation!). The robot wouldn't know about reformulating or enhancing that system, other than with the rules it was given. So it would never conceive of anything.
Regarding #2, suppose it was possible to exploit effects to some desired end. This could be done classically, but not at the quantum level. The indeterminate nature of quantum processes prohibits exact determination of events. Look at Schrodinger's Wave Equation. We can only speak in terms of probability of the existence of a specific wave formation, motion, and location in space/time.
Yet, quantum events are steered all the time to deterministic *classical* ends. For example, photons and other particles are artificially excited to in lasers to form coherent energy patterns that can do all types of neat things like cut through matter. However, the energy levels and the timing of such emissions are uncertain due to Heisenberg's Uncertainty Principle.
Same thing with atomic and chemical reactions. These can be artificially induced. But here too there's the element of unpredictability. We can push an electron to a higher orbital state, and we can also cause it to bolt from the atom; but we can't also know it's energy level or momentum.
So the question is whether intelligence is more akin to the "coherence", or an end result of the "uncertainty" or "unpredictability" described above.
Regardless of whether it's quantum-rooted, it would need to bypass modeling process and would require some type of "spontaneous generation".
In order to artificially generate awareness/intelligence, we would need to discover a way to induce a series of indeterministic processes that would generate awareness, regardless of the modalities of that awareness (i.e. one way or another, it would be generated). Once that was achieved, then more advanced systems of awareness and intelligence could be induced through learning, pattern recognition, refinement, & error correction to harvest a more effective approach to steer the quantum probabilistic model to more favorable "indeterministic" behavior! In other words, we would become better quantum poker players.
JJC: "Are non-computable functions necessary for intelligence?"
DE: Penrose says they are.
JJC: "I don't think we can know until we have a definition of intelligence that a "quorum" (so to speak) of scientists can agree on."
DE: Yeah, that's the million dollar question.
JJC: "If an artificial entity can continue to function independently long enough to reproduce several times with a proper amount of variation, then the entity will probably be able to evolve. Give it the ability to repair itself independently and it can probably improve itself during its operation (it's "lifetime")."
DE: We're on the same wave length here. I was posing a similar scenario above. Penrose however discards the use of neural networks and genetic algorithms as means of generating awareness/intelligence. I was just looking for some loophole around his objections.
When you get a chance, read the book and see what he says.
JJC: "If the offspring of this entity do evolve, then the I.Q. of some of the offspring will probably increase from generation to generation, or even within the operational "lifetime" of a single individual of this "species"."
DE: We know that technology evolves, but can a technological device intrinsically evolve as well? That's the scenario you're suggesting, and it's an exciting one to consider.
It would seem to be apparent that human beings had awareness from the beginning of human existence; perhaps intelligence as well. But we don't know that. These could very well be the pure product of evolutionary development. It's probably some combination of innateness and evolution, but again we really don't know.
An artificial device like a computer or robot wouldn't need to emulate human evolution or possess such innate human awareness, but would need to possess some means to acquire that awareness and develop some semblance of intelligence (i.e. pass through an evolutionary process that would generate it).
JJC: "Some people have suggested we define intelligence in terms of ability to survive."
DE: No, I think there's more to it than that. Problem solving is the core of intelligence, but solving problems doesn't necessarily lead to a higher probability of survival. I love solving online puzzles, but if I don't apply that knowledge in a meaningful way to better the condition of the human species, then it won't lead to an increase in survivability.
Ah phooey! Can't edit our own comments.
Regarding "humans, on the other hand, are not *aware* of this limitation!"
I obviously meant to say "humans, on the other hand, are *aware* of this limitation!"
Post a Comment
<< Home