This page has notes on my PICS model for exploring complex systems as a learning process with a network structure, applying to all forms of physical systems and events. Prepared for the July 7-12 2007 SASO Conference.
One application is to use it to isolate 'cybernetic body parts' from natural systems,... to study what they were embedded in and how to fit them back in!
Some 'sidebar comments' on the conference below
Physics of Happening - (my old
systems observation methods page)
08/29/07 - I two best new ideas from SASO were adding to the definition of 'system' as 'a thing in a box' the new definition of 'complex system' as 'a learning thing in a box', and secondly new paradigm of Self-org & Self-adapt as the two step process for designed learning systems of first emergence and then integration into the larger complex environment. Now if I could just persuade them that complex natural systems are a gold mine amorphous computing organisms of just that sort...!
08/29/07 - My best new idea in NetSci, that I have not presented yet, is the realization that the way we draw network maps hides how they are embedded in their underlying natural systems. Every 'node' is actually a 'hive' of activity on another scale of description, and whether your call these basic 'members' of the network 'nodes' or 'hives' depends on which direction you are looking, up or down, in the organizational hierarchy of the whole complex system. How network diagrams are drawn also hides the 'complete loops and circuits' of the network itself. Emphasizing the hierarchy of connections distracts from the 'democratic' function of accumulative distributed process that is the real vital collective working of complex natural systems. Both these ideas came up in puzzling over the difference between 5 degrees of separation (people) and 19 (web pages), means that the web must be grouped into many highly remote high dimensional environments, separate 'hives' of ideas, each of which works as an independent whole world unto itself.
07/07/23 - Cybernetic Body Part List: Starting with 'homing devices', one could begin a list of natural system cybernetic 'body parts' of engineering interest. Any 'rule' of behavior or 'cell' of relationships could be looked at this way, a behavioral element to study separately as well as for how it is imbedded in the larger systems from which it was extracted.
| Complex System | ||||
| 'Body Parts' | Behavior | Function to Emulate | Key Observation | |
| Homing Devices - tapping natural systems |
||||
|
|
-behavioral recognition, not global probabilistic model, locating a thread | -connecting w/ natural
feature -locking in on path of discovery, either of thing or type of event |
-dog's constant sniffing | |
| Mutual Homing Devices - engaging natural systems |
||||
|
|
-mutual behavioral recognition | -simultaneous connection between natural agents | -preparation, readiness | |
|
Complex Networks -Natural System |
||||
|
- integration of
complex system scales acting as a whole. - interior link community, neighborhood link community, remote connections, |
- Learning process - emergence and integration |
- Remote links
give the larger net it 'connectivity' - Local links it's creativity in combining with ... - Internal links from which the individuals emerge
|
Some things are easier to do than explain, like a handshake. Engaging with rather than controlling natural systems can be an enormous simplification! If anyone has other things to add to the list of things we're trying to emulate, please contribute!
07/19/07 - Jay Sussman's 'Body Forms' - There seems a strong similarity between Jay's idea of studying the cybernetic body forms of traditional engineering circuit design, and watching the development of natural cybernetic body form elements as they evolve in complex natural systems. If we could learn to read them, the design of nature's organizational 'cells' would provide a storehouse of new kinds of parts to play with, along with a map to how they hook up with the larger complex systems in which they are embedded. Any mapped whole network of relationships from any real complex system can serve that purpose, for example. Once extracted it can be simplified and then reconsidered as a part of a larger complex whole, helping to expose the interface with the larger system. Is there a catch? Yes, what you'll find are not 'control circuits' mainly. They're something else, and don't work the same way, and that may be the best reason to look!
07/16/07 - Links to the Softside...
re: http://norvig.com/adapaper-pcai.html - a nice overview understanding of complex software design environs, though note interesting error, in giving the date as 'today' as if 'today' would still be that when it's read tomorrow...:-) - some advertisement angle, but great overview perspective and linksre: http://paris.utdallas.edu/iwsc07/ - a July,24 07 conference in China discussing many of the same things
7/12/07 - Rigid structures to Fluid organization In a new subject here are bound to be many kinds of surprises. One of the ones I've been talking about is how the rigid structures we find in fluid organization are very useful as learning tools for exploring the fluid worlds around them. One of the related nice new observations that several people at SASO were making was about the relation between Self-organization and Self-adaptation as distinct processes. There may be additional meanings to the terms as people are using them, but one prominent one I see is that they capture most of what one would mean by 'emergence of new order' & 'adaptation of and to the larger order', or perhaps just invention and integration. These are time sequenced events, A then B. I've been watching the A part in some detail and it really helps to have the B part in focus too. As I see it the first is a growth process involving a non-linear accumulation of new relationships and the latter a meandering negotiation of more often successively smaller emergence events, perhaps distributed throughout the whole interaction medium. As with anything that occurs in time, there are limiting response development times.
7/11/07 - An engineering/observation exercise:
Mining natural systems for design ideas. A typical natural system observation exercise
is to juat watch a traffic intersection for at least a steady half hour at a
time, but mostly just as you'd watch a landscape or ocean waves, both passively
and attentively so your thinking about how the negotiations are taking place are
not erased by your thinking about a particular network relationships you notice.
When you see a pattern follow it a little to see what it seems connected to, but
then go right back to simply passively observing. The mechanics of
intelligent complex systems observation is not unlike meditation in the one part
of being careful to be receptive to what's there, so that you can then explore
it. The intersection of Mass Ave and Memorial drive at the bridge
during the AM or PM shift is a good one, but so would any other intersection
with several different kinds of traffic that needs to negotiate the intersection
differently.
The idea, which might actually take a few tries, is to observe the several scales of interactions working together, and then notice where things follow paths of easiest steering, and try to imagine what sort of 'Lagrange Points' (perhaps actually or metaphorically) they're using for it. That'll not be easy, of course, but a little experimentation may reinforce the conceptual possibility of doing so. The main purpose of the exercise is to extend one's own perceptual reach into the complexity of natural system design, using any simple net of relations you start with as a guiding question, and producing better questions about it. In canoeing the steering, for example, is all about making your small correction at the perfect steering points as they approach, avoiding oversteering with the least effort, I think, though a good engineer might read it differently. A canoeing control stroke has a critical timing, attack, duration and decay, with changing power and direction, executed with a nearly effortless single motion. It's much easier to do than to explain, but observing extends how intelligently one can use natural system ideas you can control, and negotiate with those your observations show you need to respect.
7/10/07 - Desynchronization A nice new result for natural system exploration from the SASO papers turned up in the paper by Ankit Patel on resource allocation and the desynchronization of flows as a strategy for self-regulation. That's also a measurable property of nature and it seems very likely that the onset of desynchronization in natural flows would often indicate the emergence of a collective behavior of cooperation in a physical world system. Used as a measure that could locate the physical process of that emergence in the physical context, desynchronization indicating the smoothing of flow, adds to the measures that can be used that way, similar to the development of derivative continuity in sequences of measures (disappearance of step noise), and the markers of growth phases indicating the explorable beginnings and endings of the emerging networks of relationships themselves. Each of these can be used to locate instances of the physical system models of the behavior of the simple feedback nets that digital models are based on, and the physical model then investigated for the population of other feedbacks (enabeling and disabeling , arousal and lubricating, etc. ) deeper in the complex natural system context responsible for engaging the behavior of the simple networks we can model.
