Complex systems made learnable

by Carl Dyke

My friend and sometimes tennis partner David just emailed me this link to a story at titled “Through a sensor, clearly: Complex systems made observable.” It’s right up my alley, he thought, and right up our alley, I thought.

Now, I don’t have either the math or the graphical chops to get under the hood of this research. But I think I understand what they’re up to, and I think I know enough to spot a couple of places where questions might be asked. For example, if I understand correctly we’re talking here about describing a snapshot of a complex system; it’s my impression that once the system is actually complexing, the data-crunching becomes prohibitive. But if so, one moment of a dynamical system is of limited utility, since it captures the system but not the dynamical. If I’ve understood correctly, this is not a criticism, but an appreciation of where we are in the learning curve.

I also appreciate that there’s a devil in the details of observer design; that is, the sensors have to be able to tell the difference between information and noise, nonlinearity and randomness. In effect this means that the sensors have to be able to learn to discriminate intelligently, which most human brains are not that great at. But they’re just doing feasibility at this stage, and I gather they think if they can use graphical modeling to specify some system parameters, they can eventually walk-in the data-gathering to yield more satisfying descriptions.

Well, I bet about half of what I just said is at least a little bit wrong. What I hope is that I’m just wrong and not ‘not even wrong‘, that is, that I know at least enough to be worth talking to further by someone with a better understanding. And this brings me to the question for today, which is this. Given that the project here is to represent and understand complex systems, which explicitly include “biological systems [or] social dynamic system[s] such as opinion or social influence dynamics” – that is, to start with, citizenship and life itself – what responsibility does a university general education core program have to bring students up to a kind of elementary competence where they can participate responsibly in this kind of conversation? What and how would we have to teach to make that so? And what in the reverend paleo-disciplines and contents might need to retool or move aside to enable this development?

UPDATE: if nothing else comes of this post, at least I’ve learned what it means to be ‘fractally wrong‘.

8 Comments to “Complex systems made learnable”

  1. I need to look at this way more deeply before I comment, but a first glance tells me that your “snapshot” comment is probably on the mark.

    I’m looking for ways to visualize ANNs – specifically to give people an idea of the emergent “functions” they perform – so this is of great interest to me.

  2. I recently had a back-and-forth with an ex-studewnt who’s now a (crusading) lawyer in Denver or some place in that area. He asked me if I had any wisdom to contribute about a situation where old folks were being ripped off somehow (I can’t remember just how: it’s a growth industry). What emerged out of the convewrsation was that I said he needed to understand the situation as the rippers off understood it (and they obviously did, since they were extremely good at it) and stop thinking about it as a crusading liberal do-gooder lawyer. He, in effect, said that he couldn’t do that and still be a crusading do-gooder lawyer. Two competing models built on two competing sets of imperatives.
    Hypothesis: The geezers will continue to get ripped off.

  3. OK then. Maybe that’s what they’re for.

  4. Well, maybe; but they themselves might come to understand the grift the way the grifters do. The question here was, for example, capturing the system without capturing the dynamic. There seem to be three modes of access: the grifters, the geezers, and the lawyers. Without getting into the characteristics of the three, the important thing to notice is that the grifters do understand the dynamic. It’s practical knowledge for them; and only if you prioritize theoretical knowledge can you dis their understanding of what they’re doing. Or, you could ask the question ‘tother way round: What has to be withheld from the geezers in order for the grift to work? Can you see the dynamic in a snapshot? isn’t very far from Can you see the real game on the ballfield. In the latter case, Asher suggests that you can, and I agree. I also think you can see the dynamic in the snapshot. — not from the point of view of the virgin innocence of solipsism, but from the “how to do it” point of view of someone with relevant experience, and so on. I’m reading a book at the moment that wants me to understand mathematics in that praxical way, claims superior legitimacy for that way, and calls it an art withour giving up its cognitive authority usually assigned to math.

  5. Sure, I get it. But the sense in which we can see the game is that we can get a sense for how it works from watching it for awhile – in fact, that’s how we do, even if we’re also playing it. Same with the grift. There’s no photograph you could take that would tell you how it all moves around, any more than one frame of a tropical decline tells you how weather works. Now, if you’ve got that feel from long observation / participation, then you can read the dynamic back from the snapshot, but that’s because you can fill in the dynamic the snapshot can’t from an experiential archive.

  6. Exactly. From a philosophical point of view, this is another nail in the coffin of the Parmenidean dream of rational stasis — to some extent we’re all re-writing Dewey (as well as Hegel and Marx) along these lines. Lived experience is necessary for knowing, and can’t be reduced to the “rational moment.”

  7. Right. And as you correctly imply, this is also the subtext of the gen ed reform, and the moocs debate, and mind emergence. You get into these conversations and find out very quickly what the participants’ complexity tolerance is.

  8. From a different space on the Net. Have a look at

Leave a Reply!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: