Sunday, March 11, 2012

Alan M. Turing

Moral: Wherein we get back to some technical issues that imply limits.


The motivation is partly celebration of the 100th of Alan M. Turing. The March ACM Communications has a series of articles. We'll use one of these. To be brief, Turing's work was seminal to computing; yet, some, like the quants in their race for profit, ignore some basic issues. Then, we, the populace, bear the costs. Too, economists have migrated toward 'modeling' as a key method; they, too, do not seem to have paid attention.


But, first, a comment is in order. What we have in the world are two things: those who use these limits to sustain their intake (Made-offs and those who game the system and more) and those who let themselves be led by the nose (as in, the stupid, the Made-off victims, et al).

You know what? This has always been the case. The problem now? Computation. It is not as robust as we might be led to believe.


One of the ACM articles is by Prof Cooper (of Leeds, UK): Turing's Titanic Machine? The Prof looks at the background and provides some insight into the current debate. But, we might ask, what debate? You see, many operational stances have been able to skirt around the issues. That ability, itself, can be seen within the contexts we obtained from Turing's work.

For now, let's just look at the categories of adaptation to the reality. That is, even if one has not read Turing's work, dealing with computation causes one to broach upon this subject, by necessity. The Prof has four categories (Note: the order does not imply ranking). We (perhaps, tongue in cheek) could add more.

  • Reductionists -- characterized by confidence, these apply what they know. Is this the most? Wait? Isn't this operationalism, at its core? Or, in economics, making money if one can. 
  • Impressionists -- now, this mindset has become aware of the limitations of science (ah, is this where quasi-empiricism first come to fore?) and knows that we need philosophical (just keep from becoming too meta) bridges. Wait? Perhaps, the economic schools, and the debates thereof, crop up. 
  • Remodelers -- ah, hypercomputationalism or perturbations, one might say. Or, there looks to be something parallel to what we see with the interpretations in the realms dealing with the quantum world. In economics, is this where we have the interminable experimenting with our lives? 
  • Theorists -- ..., sloughing off a whole lot of detail, let's just point to Chaitin (pdf), for now. But, one could ask? Where is our meta-economics? 

To reiterate the importance of the subject, we can use a phrase from the Prof's article: In 1970, the negative solution to Hilbert's Tenth Problem when combined with results from classical computability arising from contrived priority constructions, circumstantially pointed to a very rich infrastructure of incomputable objects.

Which means several things that ought to give one pause. Of course, we can (ought to) go ahead with innovation. But, letting the best-and-brightest experiment real time (and consideration of the side-effects, thereof) is not something to take lightly.


We'll get more into the Prof's notions. However, for now, contemplate the fragile state, please.


12/13/2012 -- Is it time to move beyond the Turing Test?

08/04/2012 -- Alan will feature in coming discussions.

05/01/2012 -- We'll get back on this theme: Technological singularity (note, not too late, page dates from late 2010). This is put here as Alan argued that computational intelligence might exceed human talent (true, in many cases). Yet, I need to ask him: Alan, how are computablity issues to be resolved within the computer? You see, this is what the role will be for humans: cut out of the fog, essentially. The argument will ensue at some future point, and memes will be on thing to consider.

03/16/2012 -- Computability's definition (yes, used in the context of this post) will need to cover some notion of the delta (difference) between expectations and delivery, implying (yes) the importance of the user's values. That is, the providers (ISPs) are servants, not the other way around. Utility, if you would.

03/12/2012 -- Related organization (CiE). They'll feature Alan this year at their meeting in Cambridge.

03/12/2012 -- Prof Cooper has an interesting mathematical pedigree. With two advisors, he has two 'grandparents' of note: Ludwig Wittgenstein (Wiki) and Alan M. Turing (Wiki). The blogger does not have an advisor/parent, but he claims a heritage via a cousin and an uncle: Carl-Wilhelm Reinhold de Boor (Wiki) and G.H. (Godfrey Harold) Hardy (Wiki).

Modified: 12/13/2012

No comments:

Post a Comment