chrism
New Member
Posts: 4
Likes: 3
|
Post by chrism on Jul 16, 2020 15:03:55 GMT
Hi all, newbie here! I'm really interested to hear what you think about an recent essay called Life Through Quantum Annealing.It argues that our universe is created through quantum annealing - a form of quantum computing - and it connects it to observations in physics, cognition, and biology. It also predicts how it should behave in a series of "games," which even seems to align with human behavior. From my perspective, if this is true, then it can establish new grounds by which we understand the mind and various branches of philosophy. Interested to hear thoughts from you all. |
|
|
|
Post by karl on Jul 16, 2020 16:06:29 GMT
What the article is describing are brilliant ways to find solutions to complex problems. However, it all boils down to algorithms, as a logical system based on natural numbers. A Boltzmann machine couldn't answer a simple questions such as: "Do we have free will?" Or: "Why are there conscious beings in this universe?" Understanding through algorithms has its limitations, and one hasn't properly identified those limitations and figured out how to go beyond them. I suspect the only path towards artificial true intelligence would be something so unethical it should be banned. One would need to identify the processes in the human brain that produce consciousness, for then to re-engineer them in a laboratory, after which they could be fused with computer technology. That would create a biological computer capable of conscious understanding. Such an entity should be regarded as an individual with rights. If it instead is used to serve mankind, there would be a risk of a real life terminator scenario.
So to answer the original question: No, I don't think we exist within a quantum computer.
|
|
chrism
New Member
Posts: 4
Likes: 3
|
Post by chrism on Jul 16, 2020 16:23:01 GMT
I think a sophisticated enough Boltzmann machine might be able to "contemplate" those questions in the same ways humans do, though. They are not algorithmic in a linear sense like most classical computers; they are probabilistic, stochastic, and course-grain information/concepts in the same way that human do through analogy. But as you stated, the challenge lies in knowing at what point such a system is as conscious as we would consider a human to be. |
|
|
|
Post by karl on Jul 16, 2020 18:28:11 GMT
I think a sophisticated enough Boltzmann machine might be able to "contemplate" those questions in the same ways humans do, though. They are not algorithmic in a linear sense like most classical computers; they are probabilistic, stochastic, and course-grain information/concepts in the same way that human do through analogy. But as you stated, the challenge lies in knowing at what point such a system is as conscious as we would consider a human to be.
One problem for me with this article, is that it's a popular scientific presentation of a variety of immensely difficult subjects. I am sometimes not quite sure exactly what it is that I'm reading, and I suspect I'd only know if I had read an actual scientific paper on it. In one part of it, it brings in a subject I have looked into before with the following sentence:
"Physicists have noticed that the fabric of spacetime can emerge naturally from a complex network of entangled particles"
This gives the wrong impression. Replace "can" with "may" and it would be correct. That may sound like a small difference, but it's not. They do not know that the fabric of spacetime can emerge from entanglement.
Further it says:
"These findings come from an idea called the Holographic Principle, which states that all information about a “bulk” area is contained within a lower-dimensional “boundary” — like saying that the 3D contents of a room can be completely described by information encoded on its 2D walls."
When applied to the universe, the idea is that what happens inside the universe is described on a 2-dimensional surface infinitely far away. This model of the universe premises that the universe has negative curvature, and while initial observations suggest that it's completely flat, newer observations suggests that there is a slight positive curvature.
Another thing to note is that it's not quite correct to state that the 2-dimensional surface describes what actually happens in the universe. Instead, it's the information about all the probabilities for what could happen. If you're holding a Geiger counter over a piece of radioactive material, what you're hearing will not be coded on that 2-dimensional surface as anything but one possibility out of many. If you have a radioactive atom, the information on that surface will only tell you the probabilities for when it will decay, not which possible outcome will be realised.
When it comes to the quantum computing bit, because I know less about it, there might be a chance for that I've misread it. But what I was left with, was that while the calculations are done in a probabilistic manner, the actual outcome isn't probabilistic. It's given a problem that has one correct answer, and it's only the method it's using to find that answer that is probabilistic. But let me know if you think I got that wrong.
|
|
|
|
Post by joustos on Jul 16, 2020 20:50:57 GMT
What the article is describing are brilliant ways to find solutions to complex problems. However, it all boils down to algorithms, as a logical system based on natural numbers. A Boltzmann machine couldn't answer a simple questions such as: "Do we have free will?" Or: "Why are there conscious beings in this universe?" Understanding through algorithms has its limitations, and one hasn't properly identified those limitations and figured out how to go beyond them. I suspect the only path towards artificial true intelligence would be something so unethical it should be banned. One would need to identify the processes in the human brain that produce consciousness, for then to re-engineer them in a laboratory, after which they could be fused with computer technology. That would create a biological computer capable of conscious understanding. Such an entity should be regarded as an individual with rights. If it instead is used to serve mankind, there would be a risk of a real life terminator scenario. So to answer the original question: No, I don't think we exist within a quantum computer. I can neither agree nor disagree with your own conclusion, or with the original thesis, since I do not have a full understanding of the current terminology about computers, quantum, etc.. which was not part of my school education. You can guess how old I am. However, I grasp old-fashioned popular ideas in your thinking that are blended with the new technological terminology. One case: The brain does not produce consciousness. Already in the Renaissance, some philosophers intuited that all matter is self-sensitive or conscious; animal and human brains have specialized capacities for consciousness (which would take some pages to describe), and I admit that we may be able to construct artificial brains with some forms of consciousness. Another point: rights. social rights, do not exist because of consciousness or of free will; they are contractual in a society of free humans (in a Republic....) Finally: certain machines or computers may not be able to answer questions about free will, consciousness, God, etc.,because these words are not understood, and Understanding is not a computation process. (Sorry, it would take too long to expound my done Psychology.) |
|
|
|
Post by karl on Jul 16, 2020 21:03:43 GMT
Chrism: To clarify what I wrote above: In the article, one example given for a problem solved with quantum Annealing is the traveling salesman problem:
"Given a list of cities and the distances between each pair of cities, what is the shortest possible route that visits each city and returns to the origin city?"
Whatever method one uses to solve that problem, and whether it's probabilistic or not, it's a clear question with one precise answer. Let's imagine the following question:
"Will this radioactive atom with a half-life of 1 year decay within the first year?"
There is no yes or no answer, but if you've programmed the computer to understand probabilities, then maybe it would identify it as a trick question, and state: "The probability for that it will decay within the first year is 50%."
Then let's imagine that humans have free will, and that free will is made possible by quantum uncertainty. So that there is some way by which our consciousness may affect quantum processes in the brain. Then imagine someone asking you the question:
"Will I make a jump in one minute?"
You would probably guess that if you say "no", then the person will prove you wrong and make a jump. And if you say "yes", the person won't make a jump and prove you wrong. You've then encountered a question that doesn't have a precise, correct answer. The only reasonable answer would be: "It depends on what you choose." -Which I think is a concept you couldn't program a computer to understand.
|
|
chrism
New Member
Posts: 4
Likes: 3
|
Post by chrism on Jul 16, 2020 21:43:02 GMT
karl: I see what you're saying - I think that all actually fits the model, at least the way I see it. If we are in some "calculation" phase of a big computation, and predictions are made based on probabilities, then it could be reasonable to think a probability-based program understands changing parameters and can adjust its models accordingly.
In the travelling salesman problem, the parameters are well defined and don't change; it can come up with an answer with high (virtually 100%) probability. In half-life decay, all the parameters may be nearly intractable, so its probability can change over time as it observes the system. In a "will I jump" situation, it's known that the parameters will change based on its own answer to the question.
In machine learning there is the concept of a Markov chain where the probability space changes based on each step within a sequence of events. I think that fits right in with the model of a Boltzmann machine, and it seems to match how we reason things too.
|
|
|
|
Post by karl on Jul 17, 2020 19:59:30 GMT
What the article is describing are brilliant ways to find solutions to complex problems. However, it all boils down to algorithms, as a logical system based on natural numbers. A Boltzmann machine couldn't answer a simple questions such as: "Do we have free will?" Or: "Why are there conscious beings in this universe?" Understanding through algorithms has its limitations, and one hasn't properly identified those limitations and figured out how to go beyond them. I suspect the only path towards artificial true intelligence would be something so unethical it should be banned. One would need to identify the processes in the human brain that produce consciousness, for then to re-engineer them in a laboratory, after which they could be fused with computer technology. That would create a biological computer capable of conscious understanding. Such an entity should be regarded as an individual with rights. If it instead is used to serve mankind, there would be a risk of a real life terminator scenario. So to answer the original question: No, I don't think we exist within a quantum computer. I can neither agree nor disagree with your own conclusion, or with the original thesis, since I do not have a full understanding of the current terminology about computers, quantum, etc.. which was not part of my school education. You can guess how old I am. However, I grasp old-fashioned popular ideas in your thinking that are blended with the new technological terminology. One case: The brain does not produce consciousness. Already in the Renaissance, some philosophers intuited that all matter is self-sensitive or conscious; animal and human brains have specialized capacities for consciousness (which would take some pages to describe), and I admit that we may be able to construct artificial brains with some forms of consciousness. Another point: rights. social rights, do not exist because of consciousness or of free will; they are contractual in a society of free humans (in a Republic....) Finally: certain machines or computers may not be able to answer questions about free will, consciousness, God, etc.,because these words are not understood, and Understanding is not a computation process. (Sorry, it would take too long to expound my done Psychology.)
Whether the brain produces consciousness or not may be a matter of semantics. The brain consists of matter, and so the consciousness of the brain must, one way or another, come from matter itself. However, that we experienced one "I" (As in "I am") while our brain consists of 100 billion brain cells, implies that the exists a mechanism for that matter working in unison can produce one, rather an a multitude of consciousness. If one knows how to isolate that mechanism, then that could possibly be re-engineered in a laboratory.
That rights are, in practice, contractual is a bit besides the point, since I was uttering my own opinion about how things ought to be, not describing how they are or have been.
I agree with your last point: Understanding is not a computation process.
|
|
|
|
Post by karl on Jul 17, 2020 20:23:50 GMT
karl: I see what you're saying - I think that all actually fits the model, at least the way I see it. If we are in some "calculation" phase of a big computation, and predictions are made based on probabilities, then it could be reasonable to think a probability-based program understands changing parameters and can adjust its models accordingly. In the travelling salesman problem, the parameters are well defined and don't change; it can come up with an answer with high (virtually 100%) probability. In half-life decay, all the parameters may be nearly intractable, so its probability can change over time as it observes the system. In a "will I jump" situation, it's known that the parameters will change based on its own answer to the question. In machine learning there is the concept of a Markov chain where the probability space changes based on each step within a sequence of events. I think that fits right in with the model of a Boltzmann machine, and it seems to match how we reason things too.
What did you mean by: "In half-life decay, all the parameters may be nearly intractable, so its probability can change over time as it observes the system."?
I might have misunderstood what you wrote, but the probability for whether the atom will decay does not change. It remains a constant. If the half-life is 1 year and it doesn't decay within that year, the chance for whether it decays within the next year will still be 50%. Since there is no way to predict when it will decay, it means that the Boltzmann machine must be able to produce outcomes that are principally unpredictable. Our universe is not deterministic, and so a universe ran by a quantum computer cannot be deterministic either. One way or another it must be able to generate quantum randomness. I didn't find any reference to that in the article. Here, however, is an article on the matter.
If one incorporate quantum randomness in a computer simulation, then all bets are off, as far as I'm concerned, on whether it's principally possible to simulate a world in which conscious beings could actually exist. -Since I believe quantum randomness and consciousness are closely connected.
(I doubt that the complex process described in that article would actually produce randomness comparable to the quantum randomness in the brain that produces consciousness, but that's beside the point. What I'm asserting is that no simulation that doesn't include quantum randomness may simulate reality. It would be a world where particles wouldn't obey the laws of quantum mechanics, and a world without consciousness.)
|
|
