| Subject: NeuroNetwork -- studying AI design and other interesting stuff |
| From: Clark Robinson <robinsonchicago@gmail.com> |
| Date: 8/15/10, 01:52 |
| To: Barrett Brown <barriticus@gmail.com> |
Ed Bullmore, psychiatrist
It turns out that the human brain and the internet have quite a lot in common. They are both highly non-random networks with a "small world" architecture, meaning that there is both dense clustering of connections between neighbouring nodes and enough long-range short cuts to facilitate communication between distant nodes. Both the internet and the brain have a wiring diagram dominated by a relatively few, very highly connected nodes or hubs; and both can be subdivided into a number of functionally specialised families or modules of nodes. It may seem remarkable, given the obvious differences between the internet and the brain in many ways, that they should share so many high-level design features. Why should this be?
One possibility is that the brain and the internet have evolved to satisfy the same general fitness criteria. They may both have been selected for high efficiency of information transfer, economical wiring cost, rapid adaptivity or evolvability of function and robustness to physical damage.Networks that grow or evolve to satisfy some or all of these conditions tend to end up looking the same.