That's true but it has to be done correctly or you end up solving the problem before it had time to compute it correctly or completely. Obviously it takes less time to look at a multiplication table than to actually work out the calculation.
-
Welcome to Religious Forums, a friendly forum to discuss all religions in a friendly surrounding.
Your voice is missing! You will need to register to get access to the following site features:- Reply to discussions and create your own threads.
- Our modern chat room. No add-ons or extensions required, just login and start chatting!
- Access to private conversations with other members.
We hope to see you as a part of our community soon!
Getting from cause effect to awareness
- Thread starter idav
- Start date
What is your evidence for this? Studies for the past 20 years say you are incorrect. Have you read the literature on developmental cognition and facial recognition in neonates?
I wouldn't say pre-programmed, but yes the brain is amazingly equipped to learn the way that no other system in the known universe is. It is the most efficient learning system there is and vastly outstrips any other that is currently known or even predicted (e.g., quantum computing is supposed to result in certain advantages but if this is one we have no idea how that would work).
It isn't. Because the categorization ability is equivalent to that of a plant or a cell. We can get computers to evolve (mathematically, not in the biological sense) over time to increasingly adapt the same procedure to classify input according to desired output. However, we can only do this for input that are e.g., linearly separable or otherwise mathematically divisible into classes. Most conceptual classes are impossible or too difficult (not computable in p-time) to do this. Computers have no shortcuts like brains. I can train my dog to associate the input of the word "food" with many things because like me her brain is composed of active nonlocal networks. It is easy for stimuli to link up to other encoded concepts because these concepts are always being encoded and linked and distributed among other types of pairings. Computers are built to store memories in discrete departments. Not only that, attempts to build computers that aren't haven't been very successful at all. We simply aren't near the level of Biocomputing needed to achieve the kind of flexibility required of concept-processing.
Programming suggests it is computable. We have no evidence to suggest that it is and currently the evidence suggests that it isn't. It's hard to say, of course, and even if living systems aren't computable that just means that no equivalent of a turing machine will be aware. However, calling something which is always active and that has no software and no hardware "programmed" is just going to mislead.
PolyHedral
Superabacus Mystic
Computing power is not synonymous with learning ability.
Quantum mechanics is computable. Therefore, to say that the brain is not is to deviate from standard physics quite severely.
It is when we don't have an effective learning algorithm.Computing power is not synonymous with learning ability.
This is so utterly unrelated to anything I said I have the feeling you misunderstood something.
PolyHedral
Superabacus Mystic
For all I know, once you decode the actual model that the brain is running/building, you could implement it on a machine from 2003. The fact that we don't have an effective learning algorithm means that we don't know how much "real" computing power is needed to run it, and what scale it can run at. That makes the difference between quantum and classical computing irrelavent - we don't know what to do with them yet.
"Computable" is a technical term, meaning that it can be answered if given an unbounded amount of spacetime. (There's no problem if this is larger than the amount of spacetime in the universe, it's still computable.) This property is transitive when you start composing computable functions together - if two questions are computable then the combination of the two is also computable. Similarly, if your question is about some finite number (e.g. how long does it take to find a solution to a boolean formula containing n terms) and it is computable, it will remain computable for arbitarily large inputs.
IOW, given that quantum mechanics is computable for a trivial system, like a single hydrogen atom, it doesn't matter how many atoms or molecules we add to the system, we will not get an uncomputable problem. If the brain is built of quarks, it is computable, because individual quarks are computable and the interactions between quarks are computable.
Last edited:
Although it wasn't at all relevant to my point, that doesn't make it not worthwhile. However, the above is a bit misleading to say the least:
"The possibility that quantum mechanical effects might offer something genuinely new was first hinted at by Richard Feynman of Caltech in 1982, when he showed that no classical Turing machine could simulate certain quantum phenomena without incurring an unacceptably large slowdown
Williams, C. P. (2011). Explorations in quantum computing. Springer.
PolyHedral
Superabacus Mystic
"unbounded amount of spacetime."
A slowdown factor of 2^n is irrelavent for the purposes of whether or not its computable.
We can't. That's because it is "building" a model. There is nothing to suggest that brains actually work according to algorithms. We tried to make this true for 60 years and it turned out very poorly for us.
We do have learning algorithms we know are NP-complete. In fact, some of the theoretically superior learning algorithms (ones we have to reject) are NP-complete. Every model neural network is a vast simplification because implementing even simple model ANNs at the level the brain does can't be done.
Tell me about it.
That's not really transitivity but composition. Computability is defined on a set A if the set's characteristic function is computable. As composition doesn't change the properties of two functions defined over any two arbitrary sets, necessarily A and A' are computable. However, this means that there exist programs T and T' s.t. T halts if x is an element of A and T' halts if x is not an element of A.
I don't recall asking a question.
Can you rephrase this to show how it is computable? As stated, it is completely ambiguous.
Last edited:
Where are you getting your definition of computable from? And what is it that you are stating is computable (e.g., computable model or computable set)?
EDIT: I'm not trying to imply that the above is wrong, but it is hard to talk computability when you don't discuss halting or arguments.
Last edited:
nazz
Doubting Thomas
Here's my problem with the above- It is ambiguous. For example, a computable number can have infinitely many decimal places so long as there exists a procedure for calculating in finite time any particular decimal place (i.e., if we treat some real number as an infinite sequence we can compute any nth term in the sequence through finitely many computations). Naturally, for any number with infinitely many different decimal places (i.e., not a rational number), computing the number is impossible: there are infinitely many decimals and so there is no answer. But for any nth step in the computation, there is an answer. Hence, computable.
However, a problem that requires an infinite time to compute is not answerable. A computable number, function, set, etc., must be able to halt after finitely many steps. Otherwise, it is not computable. That's one of the important reasons behind the infinite tape and the ordering of arbitrarily many machines. He showed that with any arbitrary number of Turing machines with infinite tape, the problem of determining whether a procedure to find the answer to a problem for which no known explicit procedures exist is not computable. Infinite space mattered because he required arbitrarily long tapes, but infinite time means, in general, that there is no answer. The question is whether or not a problem that requires infinite time can halt at any desired step with an answer the way an infinite number can be computed.
Qubits could help as an alternative since quantum computing can allow multiple states simultaneously. Of course there are limitation since the qubit gives probabilistic answers in that it gives possible answers.
The Economist explains: What is a quantum computer? | The Economist
Even that seeing in the dark does so with the use of eyes.
That was my point. We can't say that they can at all. In fact, we can prove that we can't say this as we cannot know whether a procedure is computable if we don't have a decision procedure already.
Quantum computing is unlikely to help us here. It will likely help us get some answers faster to special kinds of computation but quantum decoherence is too difficult to control to make quantum computers capable of helping here. Biocomputing might work eventually. And there may be other technologies. But classical computing and quantum computing are not going to get us there. We need something that can simulate things like M-R systems.
That's not exactly accurate. Probabilistic computation isn't about possible answers so much as it is possible starting places. We start from random places rather than using deterministic computation.
nazz
Doubting Thomas
No it doesn't. That's the point.
My point was to Legion, in that it matters little how we are able to perceive the world, however we have to give a machine some method of perceiving even if it is just doing echo location.
It matters absolutely: Embodied cognition. What we do not know is how cognition might be able to work in other environments. But embodiment is fundamental to cognition. That said, we can program environments for computers and have, and we have built robots according to embodied cognition theory to help them learn. But we're still at the level of modelling slug learning. We've had the knowledge of how to do this for decades and decades.
FranklinMichaelV.3
Well-Known Member
Lol Proprioreceptors.
Sure but I'm not going to assume it isn't possible and we certainly have some clues. I will reiterate the argument that, human cognition isn't the only way to have cognition. As I've stated, the redundancy of the human mind is very beneficial to us but it isn't beneficial or even necessary if your trying to create cognition.
Curious why it can't just be as simple as, machine sees car, machine receives several concepts of the car and understands it. This isn't what a human does? Sees car, brain interprets object based on what is in memory. All the behind the scenes stuff doesn't get to the root of the issues of simply seeing an object and interpreting that object "correctly". I say "correctly" because interpreting is ambiguous and sometimes that means there isn't a "wrong" answer per se. It's like your saying the computer is wrong unless it does it like a human, meh.
PolyHedral
Superabacus Mystic
And we've not tried even a fraction of algorithms, techniques and ideas.
Can you elaborate on why we have to reject them? Remember, the brain is not dealing with arbitary-sized data, so bad complexity in the limit of increasingly large parameters isn't necessarily a problem.
I'm talking about computability of function evaluations.
Once you construct a model of a hydrogen atom in quantum mechanics, all physically meaningful questions such as "What is the probability amplitude of the electron at position x,y,z?" are computable.