I finished up a couple of books in the last few days (it's as though this blog is turning into Clay's personal reading recommendations...), Critical Mass by Phillip Ball, and On Intelligence by Jeff Hawkins. I also went to a play tonight, one Preludes and Fugues by someone named John Glore, and prior to that, I'd spent some time reading up on a current "documentary" called Expelled: No Intelligence Allowed (propaganda does not deserve a link).
Now, whether these things in particular (well, the first two mainly) have evoked some sort of connections in my mind or whether my current overall state had already been leading to it, I don't know. Regardless, a few things have managed to click lately, which for some reason I feel I need to divulge to the Web.
First, Critical Mass is generally about networks, and how certain general statistical properties reliably emerge from large numbers of agents, whether those agents be atoms in a magnetic material, people walking down a hallway, or stock brokers on the buying room floor. Nothing that arises is individually predictive, mind you, but certain things can be said about certain types of networks as a system. The part that interested me the most about this book was the section on small-world and scale-free networks (small-world networks being the topic of another book I like, Six Degrees by Duncan Watts). Small-world networks are those wherein the average path length connecting any two nodes is small and the "clustering coefficient" ("the average over all vertices, of the number of edges connecting a vertex v divided by the total number of possible edges in the neighborhood of v", Ball 367). It turns out that in certain circumstances, a "scale-free" network can be arranged such that its topography fits the small-world requirements as well, such as in the way networks of interconnect websites seem to have developed (seriously, if you are getting any of this--or want to--just read Critical Mass). In any case, that intrigues me--especially considering that I'm intending to get an advanced degree in networks and distributed systems. The thought that these scale-free, small-world networks can arise from an undirected addition of nodes based on some sort of criterion for survivability is definitely something I would like to pursue, which brings me directly to my next point, about Expelled.
Apparently (I haven't seen it, so I don't know) Expelled is a very thinly veiled attack by intelligent design proponents against the supposed tyranny of the ruling intellectual elite in the biological sciences--namely, the proponents of modern evolutionary theory (the Darwinists, according to the film). I'm not going to say one way or the other; I haven't seen it, so I don't know. If the reviews are correct, then it sounds like a horrible blight on the face of modern cinema and will actually harm the advancement of knowledge in this world. If not, more power to them. The reason I bring this up is because it got me thinking about evolution, about selection and mutation and about complexity can arise from simple entities. Self-organizing systems are a staple of modern evolutionary thought (check out The Origins of Order by Stuart Kauffman), and combining that with the idea of a scale-free network, with its characteristic robustness and redundancy could have some pretty interesting implications, some of which came up in the next book I read.
On Intelligence had been in the works for a while, since I got it in the same trip as some books on information theory. Still, in two sittings I managed to power through it, and given what I remember from my cognitive science days, Hawkins might be on to something about how the neocortex works. I haven't done any further research yet (although I intend to) but the "memory-prediction framework" (Hawkins 104) he espouses sounds like he's on to something. He gives some pretty detailed conceptual models of how the various layers of a column in the visual cortex might work, and ties it into the functioning of the whole visual processing chain, right up to the hippocampus at the "top" of the pyramid. In doing so, he talks fairly glibly about "connections" and "links", about axons and dendrites connecting from many nearby networks, as well as some elongated ones (specifically in layer 1) that come from the far reaches of the brain. I don't know; sounds like a small-world network to me (high clustering, potentially low path length). Also, given that he posits a hierarchical organization to the whole thing and points out that the plasticity of the brain shows that different parts are functionally identical at the biological level, it seems to me that with a bit more modeling (which may have already been done; I never said I was up to date) it may be shown that the organization of the cortex has some aspects similar to that of a scale-free network. On an interesting note, it seems that Hawkins defines creativity as a novel generation of links between similar parts of the network (he was more about applying partial pattern matches to similar situations, but maybe you can give me a little creative license).
I used to think that creativity was all about generating new content. I even coined a little phrase about it, in the hopes of explaining why I don't paint or write songs or something similar--"I can't create, I can only interpret and facsimilate." (wow, turns out that's not actually a word). The thing I've come to realize is that that interpretation and recombination that I had so often discounted in my own endeavors is creativity. Every painting, every screenplay, every song is composed of parts of the universe as experienced by its creator. The raw materials come from the same mundane sources that I have access to, yet are joined by a network of similarities and linkages that I do not have, because my experiences have provided me with a different set. In any case, I was wrong. Every day I create, by weaving a unique tapestry of familiar entities that itself can be fed back in as a raw material, whether it be something I code, write, say, or just think.
Which brings me to the last bit of raw material I had at the beginning of this post, Preludes and Fugues. I went to see the production primarily because two of the four people in the cast are pretty close friends who have already had a significant impact on my life recently (and one of the remaining two is also a less-connected part of my network) so I felt I should support them. As a brief review, I felt that the actors did a fair job of dealing with what I consider to be a poor script; I'd describe the writing to be the opposite of "subtle", whatever that may be ("blunt" doesn't really work in this case, nor does "trite").
In my own particular nerdy fashion, I started thinking about the play as a system (it didn't really work for me to think of it as a piece of music, as the author obviously intended; like I said, not subtle). The seemingly random story lines and intended (yet poorly executed) continuities of name, location, place, and actor became threads in the tapestry, links in the network (fine...melodies in the symphony). What I found was that the similarities he invoked (or tried to) did not support the piece as a whole; he spread the audience too thin, with too few of the nodes connected to the surrounding ones, yet tried to connect every node in the network to each other node. In a way, the script popped out at me as being a sort of densely-connected network, with too many redundancies and too little clustering to be believable. The overt attempts to tie everything back together became formulaic and boring, while the injection of forced absurdity and convention-breaking comedy simply seemed out of place and disconnected. The network of the script connected only a few nodes, yet there was such strong noise that the intended message got corrupted anyway.
It makes me wonder about what network science could say about such things as scripts; maybe, when modeled a certain way, there is an ideal topology for a narrative, just as there's an ideal topology for the brain, for a traffic system, for a power grid, for the Internet, etc.
In any case, it feels good to think about this. Maybe I can actually salvage something out of this random walk that I call a life; maybe it wasn't so random after all.
No comments:
Post a Comment