Fractal Intelligence

Tuesday, May 24, 2005

Delbruck and his search for the physics of emergence

This Historical account is very fascinating. Delbruck, a Nobel prize winner, was a quantum physicist that was attempting to locate a fundamental paradox in biological chemistry / quantum chemistry whereby he could extract the basic laws of a new type of physical force. That physical force, he believed, was necessary to explain the complex mechanics in cells that allow cells to exhibit behaviors we ascribe to life.

This is very interesting to me, because in my studies I am reasonably sure that I have seen the shadow of the same phenomena although I wouldn't describe it so much as a force as I would a complex set of rules of emergence in interacting systems. Nevertheless, I do believe it is a physical phenomena that should be studied within the context of physics.

I have used the term "intelligence vector" to describe this set of physical laws in the past. The basic idea is this: Any time there is communication between systems (whether they be atoms, birds or galaxies), the system exhibits emergent behavior that cannot be isolated to it's constituent parts (the whole is greater than the sum of it's parts, 1+1>2). The nature of this emergence itself appears to be anti-entropic (often in direct contradiction to our assumptions from thermodynamics), providing a vectored set of interactions towards order. While this may seem anti-intuitive and perhaps even in violation of common sense (e.g. why should interacting particles be anti-entropic), even the most basic thought experiment can reveal the truth of it: If you take a system of atomic gas in a vacuum and provide enough time, the gravitic interactions alone will cause the atoms to clump into a solar system, perhaps even clumping enough in the central gravity well to pass the mass threshold required for fusion, igniting a star. This is a well understood and well accepted concept by the same scientists who live and die by thermodynamics and yet the process itself is clearly anti-entropic. That and many other similar phenomena are ignored by scientists everywhere, or handwaved away by claiming that the gaseous state of the system is more ordered than planets, moons and star providing energy that powers potential life on the planets (a claim that is difficult to stomach).

In life, this effect is overwhelming and in our faces. If we examine the most basic form of information processing wetware in terrestrial life, the biological neural net, we find very simple information processing systems (neurons, dendrites, synapses) that when linked together are able to do profound information processing. The pattern recognition and differential calculus that such nets are inherently capable of processing cannot be deconstructed down to the constituent components. The emergent capabilities of these systems comes completely and fully from the act of interaction between the parts.
A direct manifestation of the intelligence vector.

Drilling down to the level of the cell, it is clear that the machinery working within the cell are highly ordered, highly dynamic and energetic and participating in complex interactions and information exchanges both within the cell and outwards with other cells. Analysis of such interactions through the use of emergent ideas should hold promise of breaking the secrets of cellular automata (the living kind), but this is where I think Delbruck was probably obsessed; There is a boundary in living cells between the basic automation of the smallest machinery of life and the mechanics (quantum mechanics?) of the atom. If one could study that boundary, I think he believed, one could extract the essential laws of the force that automates life.

While I think he was wrong, because i don't think there is such a force, I also think that he was seeing the shadow of something real, in that there are in fact physical laws that manifest at that boundary to move from atom to goal-directed automata. But we can just as easily study those in mathematical systems using game theory, artificial life in simulations and whatnot. Here, the study of artificial intelligence, evolution, emergence and molecular physics merge into perhaps the most profound set of physical laws that mankind will ever embark on understanding. In this puzzle lurks basic physical laws that should show conclusively that the universe not only allows intelligence but is itself structured to ensure it exists...

Friday, April 29, 2005

Thoughts on Geometric analysis of Emergent Behavior

I was contemplating emergence within the context of physics the other day (yes, I know that sounds strange). What I mean by this is that I was contemplating communication between physical bodies over time that show emergent behavior. The reason that I was contemplating such a thing is that I have this nagging, strong belief, that physics is somehow at work in emergence. Ok, so any rate:

Such an exercise will quickly lead one to realize that when thinking of physical objects communicating (say billiard balls bouncing off of one another) over time, we have a stable and static four-dimensional system that can be analyzed (for billiard balls it is 3 dimensional, 2 spacial and one time). This much is obvious.

This thinking of billiard balls got me thinking. It is clear that one can analyze a billiard game using geometric techniques when the time dimension is considered a spacial dimension. But of course, billiards aren't so interesting in and of themselves, I bring the game up to provide a thought substrate (to calibrate my brain).

If we, instead, consider some simplified artificial life system instead of billiard balls, but using the same technique (e.g. time as a spacial dimension), then we have the makings for a geometrically analyzable system where we can draw out geometric relations in the artificial physical system that has been created.

The point of studying such a system with geometric tools would be to attempt to isolate self-organizing principals with geometric mathematics in such a way as to capture the essence of the emergent (self-organizing) behavior of the system. Once such relations are describable mathematically, one should be able to use them to make predictions.

This requires further thought...

Thursday, August 19, 2004

Emergence, Fractals, Self-Organization, ...

Designing for the Emergence
Some Links

Some interesting articles and links with relevant fractal intelligence references.

"Design Methods, Emergence, and Collective Intelligence", by Nikos A. Salingaros

"Design Methods, Emergence, and Collective Intelligence", by Nikos A. Salingaros

This is an interesting paper that touches on several concepts that seem relevant and interest me.

Wednesday, August 18, 2004

Synopsis: Fractal Intelligence

Fractal Intelligence (See my writeup on Everything2.com) : The idea that intelligent systems are defined by the emergence of the components that they are composed of, and can participate as components of higher-order intelligent systems that exhibit emergent behavior.

The term Fractal Intelligence is meant to transmit the basic idea of multiple levels of cognition in intelligent systems with the fractal property of being self similar over ranges of scale and location.

Synopsis: Intelligence Vector

Intelligence Vector (See my writeup on Everything2.com) : A hypothetical phenomenon and physical law, like gravity or field theory, that manifests itself in the physical world as emergence and fractal intelligence. The, as yet unspecified, models, mathematics, relationships and underlying phenomena in the physical world that allows emergence to occur.

The term Intelligence Vector is meant to transmit the idea of a set of natural laws or a force at work in the universe that enables emergent behavior. At the bare minimum, this idea can be explored through imagining any collection of communicating components. Any system that allows communication between components is capable of exhibiting emergent behavior (behavior that is not present in the components, only in the collection of interacting components).


 
Web Counters
Dell Canada Coupons Discount