Memetic warfare

[Reginald Ret]

Emergence.


let me say that again(i love this word) emergence.

"so you're saying that whenever you reach a certain complexity threshold new properties just appear? thats crazy!"
"don't look at me man, i didn't do it."

http://www.amazon.com/Emergence-Connected-Brains-Cities-Software/dp/0684868768
good book before reading this i never would have thought it would apply to cities.

[Triple Zero]

"so you're saying that whenever you reach a certain complexity threshold new properties just appear? thats crazy!"

i've been going on about this for years, it's one of the most awesome things out there.

in case you haven't checked out the wikipedia article on Emergence, you should, it's got some pretty good stuff in it.

tell me, i dunno about your background, but have you ever ran computer simulations of systems or networks that display emergent properties? personally i have the feeling that one cannot quite grasp the amazingness of the concept of emergence until one has created and caused it himself. otherwise it seems slightly too much like a "trick" or a "that's obvious" kind of thing.

[Reginald Ret]

yeah i ran some simulations of simple ecosystems at uni, something like 1 primairy producer(example: photosynthetic algae), 1 'grazer' and 1 top predator (that eats both grazer and algae)

It weas pretty easy to tweak the system so that it became unpredictable, within certain parameters we managed to create chaos.

If this sounds like cheating just remember that having a consumed food to biomass convertion efficiency of near zero for one of the beasties is the same as removing said beasty. And the something like that goes for all other traits too.


I have a simple example somewhere you can try with you calculator but i have to dig up the book(something something chaos i think)

[Triple Zero]

cool!

was it by any chance the Volterra-Lotka simulation?
http://en.wikipedia.org/wiki/Lotka-Volterra_equation

or perhaps the logistic map?
http://en.wikipedia.org/wiki/Logistic_map

what kind of study do you do? especially if it includes books named "something something chaos"?

i'm a Computer Science student myself, mastering in Computational Science and Machine Learning.

[Reginald Ret]

i study biology, and the chaos book was a gift from my father (or maybe i borrowed it(that would explain why i can't find it))

Lotka Volterra sounds familiar but this was more then a year ago, and i'm not very good at remembering silly details like facts, methods or models(how i manage to stay in the university is beyond me too)

It was really fun to learn to think multidimensionally and F-ing hard to explain it to my fellow students(the modelling thing was one of many diverse projects).

How good are you at understanding/'visualizing' systems with 10+ variables?

[Triple Zero]

in my head? not so good, cause my head's not made for that kind of stuff. highest meaningful bandwidth is attained via our eyes, which are 2D in nature (which is why Virtual Reality keeps being such a failure at this sort of thing, you can represent stuff in 3D, but in the end it projects onto a 2D retina anyway, so you don't gain very much).

i do, however, have a whole bunch of scientific visualisation toolsets at my disposal for slicing and cutting up N-dimensional spaces, datasets and timeseries into meaningful representations so that i can interpret the outcomes of my computer simulations, not as a shitload of numbers, but as pretty graphs and pictures

it mostly comes down to a bit of creativity and practice in order to turn high dimensional data into a meaningful 2D representation. funny thing is, in Scientific Visualisation (part of my masters curriculum), there are a lot of parallels with Graphics Design theory (hence my interest in both), even though they don't really go into that very much, i noticed that my brother (who studies Industrial Design) got the same examples of "good visualisations" for some course in his study as i was presented.

okay maybe i can hold a vague conceptual idea in my head of a high dimensional space. but in my mind it usually ends up looking like a 3D space (which is the highest you can imagine, unless you train, a LOT), with the added mental note "actually it looks a lot more complicated than this"

for example my bachelors research project was about classifying 35x19 pixel images of boar spermatozoid head cells into "good" (healthy) and "bad" (damaged), that's 665 pixels, and therefore 665 dimensions. at one point, my professor theorized that, because all the good cells looked the same, but the bad ones could be deformed in a lot of different ways, the distribution of our dataset (in 665 dimensions!), where every instance of a cell is a point in 665D space (of all possible 35x19 images), might sort of look like a central cluster of good cells, with on the edges farther away from the center all the variations, the bad cells. in my head this ends up looking like a 3D swarm of points, with the mental note "actually this is 665D". it's not entirely accurate, but at least it's workable

so, you're a biology student that voluntarily picked a very mathematical topic for a project? that is very cool, we need more biology students like that
my master's research (which is currently on hold, but i shall hopefully start again soon) is partly in cooperation with the biology center of my university, and when i visit them to discuss plans and such, it's like we're coming from two different worlds sometimes the CS guys want to say "just give us the numbers we don't care where they come from" and the bio guys start explaining about cells and transcription factor binding sites and whatnot (which is really interesting btw, don't get me wrong), but especially when it comes to statistics and other mathematical oriented subjects, i sometimes wonder how they get any proper scientific research done at all (the answer: hard work, blood sweat and tears).

[Reginald Ret]

ok, you did see where i put visualize in '' right?
You were lucky with the 665D space cause it was just a binairy choice(healthy or other right?) imagine using gradients like we did(ok the value of variable X was somewhere between 0 and 1 but i accidentily didn''t write it down or save it somewhere... )

Changing the data into a good visualisation was hell! (we ended up using a 2D chart with 2 of the relevant variables on the axi , and a color coding for the different areas for wich the resulting system was judged on stability and presence of species. The only reason we got somewhat stable and sensible data was because we used an theoretical ecosystem that was so extremely limited that it won't happen irl ever.

Don't ask me to do any statistics right away btw, if i concentrate adn someone is willing to explain it to me, i could understand it but without preparation i'm as useless as most biologists :P

Oh i found the chaos book: "Does god play dice? The mathematcs of chaos" by Ian Stewart
and the trick i referenced to was this:
pick a value between 0 and 1, put into this formula   2X² -1  and iterate(put the result back into the formula)
try this for as long as you want, it wont settle down to something predictable, also try changing the beginning value just the tiniest bit and the result wil still be completely unpredictable but also very different from the previous result.

and here's a prog you can try if you know a bit about computers(i assume you know how to do this i have no clue, i would love tips or the finished product btw)

10 INPUT k
20 x = 0.54321
30 FOR n = 1 TO 50
40 x = k*x*x-1
50 NEXT n
60 FOR n = 1 TO 100
70 x = k*x*x-1
80 PRINT x
90 NEXT n
100 STOP

chaos enters if you use a k-value of about 1.5, the higher k the more chaotic. Ofcourse there are lots of exeptions, apparantly using 1.75 leads to order again(after 60 iterations you get a cycle of 3 values)
My favorite bit is were the results may vary depending on the computer you are using.



P.S. sorry 000 i also typed this for those without your elite mathematical skillz