4 ms·

"How can consciousness be cast in a mathematical formalism without losing something essential? The mathematics does lose something essential, viz., consciousne

by based2 11y ago

"How can consciousness be cast in a mathematical formalism without losing something essential? The mathematics does lose something essential, viz., consciousness itself. Similarly, mathematical models of weather also lose something essential, viz., weather itself. A mathematical model of hurricanes won't create rain, and a mathematical model of consciousness won't create consciousness. The math is not the territory. But, properly constructed, mathematics reveals the structure of the territory." http://www.pnas.org/content/105/7/2745.full http://www.pnas.org/content/105/7/2745.full Bayesian learning of visual chunks by human observers

Lambdanaut 11y ago
The eerie thing is that as far as we can tell, our brains are no more than models of consciousness, and yet they are conscious. The line is very blury between a model of consciousness and the point where it can not only be described as conscious but truly has that ethereally, hard to define quality that we "feel" as consciousness. Would the first truly general AIs simply be working models of consciousness, or would they also feel consciousness as we do? When does the viz creep into our model?
- geophile 11y ago
  > The eerie thing is that as far as we can tell, our brains are no more than models of consciousness, and yet they are conscious. Wow. That is an extremely good description of the problem of consciousness. Is this your own formulation? If not, where did you run into it?
  - Lambdanaut 11y ago
    I've been thinking about it a long time and hadn't heard it from anyone else. I'm glad to get feedback :)
- orasis 11y ago
  Right, which is exactly why our brains do not directly perceive the "truth" of physical reality. It's models all the way down.
  - escherplex 11y ago
    Bingo. In common philosophical usage 'consciousness' is conflated with what is evinced in consciousness, or process contents (cogito ergo sum). Back in 1988, Bernard Baars in 'A Cognitive Theory of Consciousness' separated self-iterative 'awareness' from process and posited 'awareness' as a 'global workspace' interfacing with parallel subliminal neural subprocessors which collectively manufacture an interactive model of 'reality'. By extension this would reduce all human intellectual modeling of ascertained 'reality' to projections of native intellectual predispositions hard-wired into the neural architecture. But what's basic? Here you're still faced with the same problem which still confronts experimental physics: to study a phenomena you have to employ instrumentation synthesized from the same building blocks of which the phenomena itself is constituted.
- icebraining 11y ago
  Isn't a brain an implementation of a model of consciousness, rather than a model itself? I don't see how they are "no more than models of consciousness". Of course, this is using "model" as "mathematical/scientific model", not as "reproduction".
deciplex 11y ago
> A mathematical model of hurricanes won't create rain What about a detailed simulation of global climate at a granularity of, let's say, individual quarks. Is there a point at which a map is so detailed that it becomes, in some sense at least, itself a territory?
- p1esk 11y ago
  I think there's an important distinction between a rain and a consciousness. Rain is a movement of physical objects, that's all it is. Consciousness, on the other hand, is not just movement of physical objects (e.g. electrons), it's more than that. The movement of electrons that gives rise to consciousness (or is interpreted as consciousness), creates information. This process is an abstraction from the physical world. A math model of a rain does not create rain, and a math model of consciousness does not create consciousness, but a computer simulation of consciousness might actually create the real consciousness, unlike a computer simulation of a rain.
  - deciplex 11y ago
    > The movement of electrons that gives rise to consciousness (or is interpreted as consciousness), creates information. This process is an abstraction from the physical world. It's not clear to me what you mean when you say consciousness is an abstraction - is this dualism? Consciousness seems pretty thoroughly grounded in real stuff. Intelligent agents might "create information" in some sense, but there is no indication that they do it by breaking the laws of thermodynamics.
    - p1esk 11y ago
      No, not dualism. I meant this abstraction in the same sense as moving from the level of transistors when talking about a computer, to the level of, say, C++ objects in the software that runs on that computer. I'm not sure I understand your reference to thermodynamics. Information is created when a certain meaning is assigned to a pattern. Who assigns that meaning? As you said - an intelligent agent. This intelligent agent is the abstraction I'm talking about. It can be implemented with electrons flowing between neurons, or with photons in some kind of an optical computer - the physical implementation is not important. The high level structure and operation are important, that's why I think that simulating a human consciousness on a computer will result in a real consciousness, unlike simulating a rain. It might not be exactly the same consciousness as the original one, but it can be arbitrary close. For example, we can simulate Atari 2600 console on a modern computer in sufficient detail to play games exactly the same as we would play them on the real, physical console. We don't need to simulate every transistor in the original console for that.

by based2 11y ago

Lambdanaut 11y ago

The eerie thing is that as far as we can tell, our brains are no more than models of consciousness, and yet they are conscious. The line is very blury between a model of consciousness and the point where it can not only be described as conscious but truly has that ethereally, hard to define quality that we "feel" as consciousness. Would the first truly general AIs simply be working models of consciousness, or would they also feel consciousness as we do? When does the viz creep into our model?

geophile 11y ago
> The eerie thing is that as far as we can tell, our brains are no more than models of consciousness, and yet they are conscious. Wow. That is an extremely good description of the problem of consciousness. Is this your own formulation? If not, where did you run into it?
- Lambdanaut 11y ago
  I've been thinking about it a long time and hadn't heard it from anyone else. I'm glad to get feedback :)
orasis 11y ago
Right, which is exactly why our brains do not directly perceive the "truth" of physical reality. It's models all the way down.
- escherplex 11y ago
  Bingo. In common philosophical usage 'consciousness' is conflated with what is evinced in consciousness, or process contents (cogito ergo sum). Back in 1988, Bernard Baars in 'A Cognitive Theory of Consciousness' separated self-iterative 'awareness' from process and posited 'awareness' as a 'global workspace' interfacing with parallel subliminal neural subprocessors which collectively manufacture an interactive model of 'reality'. By extension this would reduce all human intellectual modeling of ascertained 'reality' to projections of native intellectual predispositions hard-wired into the neural architecture. But what's basic? Here you're still faced with the same problem which still confronts experimental physics: to study a phenomena you have to employ instrumentation synthesized from the same building blocks of which the phenomena itself is constituted.
icebraining 11y ago
Isn't a brain an implementation of a model of consciousness, rather than a model itself? I don't see how they are "no more than models of consciousness". Of course, this is using "model" as "mathematical/scientific model", not as "reproduction".

deciplex 11y ago

> A mathematical model of hurricanes won't create rain What about a detailed simulation of global climate at a granularity of, let's say, individual quarks. Is there a point at which a map is so detailed that it becomes, in some sense at least, itself a territory?

p1esk 11y ago
I think there's an important distinction between a rain and a consciousness. Rain is a movement of physical objects, that's all it is. Consciousness, on the other hand, is not just movement of physical objects (e.g. electrons), it's more than that. The movement of electrons that gives rise to consciousness (or is interpreted as consciousness), creates information. This process is an abstraction from the physical world. A math model of a rain does not create rain, and a math model of consciousness does not create consciousness, but a computer simulation of consciousness might actually create the real consciousness, unlike a computer simulation of a rain.
- deciplex 11y ago
  > The movement of electrons that gives rise to consciousness (or is interpreted as consciousness), creates information. This process is an abstraction from the physical world. It's not clear to me what you mean when you say consciousness is an abstraction - is this dualism? Consciousness seems pretty thoroughly grounded in real stuff. Intelligent agents might "create information" in some sense, but there is no indication that they do it by breaking the laws of thermodynamics.
  - p1esk 11y ago
    No, not dualism. I meant this abstraction in the same sense as moving from the level of transistors when talking about a computer, to the level of, say, C++ objects in the software that runs on that computer. I'm not sure I understand your reference to thermodynamics. Information is created when a certain meaning is assigned to a pattern. Who assigns that meaning? As you said - an intelligent agent. This intelligent agent is the abstraction I'm talking about. It can be implemented with electrons flowing between neurons, or with photons in some kind of an optical computer - the physical implementation is not important. The high level structure and operation are important, that's why I think that simulating a human consciousness on a computer will result in a real consciousness, unlike simulating a rain. It might not be exactly the same consciousness as the original one, but it can be arbitrary close. For example, we can simulate Atari 2600 console on a modern computer in sufficient detail to play games exactly the same as we would play them on the real, physical console. We don't need to simulate every transistor in the original console for that.