Science
Related: About this forumEngineers Solve a Biological Mystery and Boost Artificial Intelligence
Jan. 29, 2013 By simulating 25,000 generations of evolution within computers, Cornell University engineering and robotics researchers have discovered why biological networks tend to be organized as modules -- a finding that will lead to a deeper understanding of the evolution of complexity.
The new insight also will help evolve artificial intelligence, so robot brains can acquire the grace and cunning of animals.
From brains to gene regulatory networks, many biological entities are organized into modules -- dense clusters of interconnected parts within a complex network. For decades biologists have wanted to know why humans, bacteria and other organisms evolved in a modular fashion. Like engineers, nature builds things modularly by building and combining distinct parts, but that does not explain how such modularity evolved in the first place. Renowned biologists Richard Dawkins, Günter P. Wagner, and the late Stephen Jay Gould identified the question of modularity as central to the debate over "the evolution of complexity."
For years, the prevailing assumption was simply that modules evolved because entities that were modular could respond to change more quickly, and therefore had an adaptive advantage over their non-modular competitors. But that may not be enough to explain the origin of the phenomena.
more
http://www.sciencedaily.com/releases/2013/01/130130082300.htm
TheMastersNemesis
(10,602 posts)DreamGypsy
(2,252 posts)...makes things happen quicker and cheaper.
I found a preprint of the article (from July 12, 2012) here. Haven't had a chance to read it through yet, but the authors' Discussion and Conclusions provides a little more clarity than the popular article:
It is tempting to consider any component of modularity that arises due to minimizing connection costs as a spandrel, in that it emerges as a byproduct of selection for another trait17. However, because the resultant modularity produces evolvability, minimizing connection costs may serve as a bootstrapping process that creates initial modularity that can then be further elevated by selection for evolvability. Such hypotheses for how modularity initially arises are needed, because selection for evolvability cannot act until enough modularity exists to increase the speed of adaptation.
Knowing that selection to reduce connection costs produces modular networks will substantially advance fields that harness
evolution for engineering, because a longstanding challenge therein has been evolving modular designs. It will additionally
aid attempts to evolve accurate models of biological networks, which catalyze medical and biological research. The functional modularity generated also makes synthetically evolved networks easier to understand. These results will thus generate immediate benefits in many fields of applied engineering, in addition to furthering our quest to explain one of natures predominant organizing principles.
Lots of pretty graphics showing the results of their simulations, too.
phantom power
(25,966 posts)I would expect most modularity in evolutionary biology to stem from modularity in the gene regulatory networks that control organism development. So understanding how modularity in these networks is cheaper would be crucial.
Anyway, it's a really interesting result. I'll likely try and find a way to use it next time I'm doing evolutionary computing
phantom power
(25,966 posts)Paper:
http://rspb.royalsocietypublishing.org/content/280/1755/20122863.full.pdf+html
Data and src:
http://dx.doi.org/10.5061/dryad.9tb07
animated video:
http://dx.doi.org/10.5061/dryad.9tb07
AdHocSolver
(2,561 posts)As any competent programmer knows, modular code is much easier and faster to debug and correct.
The benefit is in reduced time and effort to develop a system that "works". It may not be the "best" system. However, aiming for a system that at least "works" is why there is so much variety in the universe.
Even as a metaphor, the term "cost" is misleading.
If the universe were designed on the basis of a least cost model (think cheaply made imported junk from low wage countries), our world would have disintegrated a billion years ago.
phantom power
(25,966 posts)Cost can really be any function of a structure. amount of materials, time, pain (e.g. SW that is easier and fast to debug costs less time and pain), etc.
"wiring cost" in their case refers to things like total length of the network graph. Or the energetic or raw material resources that an organism needs to grow and maintain itself.
AdHocSolver
(2,561 posts)Whether "fixing" a mechanical or electronic device (or getting a prototype to work), "divide and conquer strategies" work best.
Most systems are over-designed, over-engineered and overcomplicated. The worst software code is referred to as spaghetti code for a reason.
The key to understanding a system is to first not describe it in terms as muddled as economics terminology, which is misunderstood by most people including a lot of posters on DU.
From the article linked to in the OP:
**********
For years, the prevailing assumption was simply that modules evolved because entities that were modular could respond to change more quickly, and therefore had an adaptive advantage over their non-modular competitors. But that may not be enough to explain the origin of the phenomena.
**********
They should have stopped here.
However, they had to add:
**********
The team discovered that evolution produces modules not because they produce more adaptable designs, but because modular designs have fewer and shorter network connections, which are costly to build and maintain. As it turned out, it was enough to include a "cost of wiring" to make evolution favor modular architectures
**********
They assumed that their model mimicked nature, which doesn't follow. The only thing their adding a "cost of wiring" proved is that THEIR MODEL required such an addition, NOT that "nature" required it.
Javaman
(62,534 posts)"we tried for a Beagle but got a Hyena instead...oops".