YKY, This paper was written for a particular audience that may disagree with my comments. Hence, consider your audience before answering any of my comments in your paper.
My primary problem is with your introduction. Most human reasoning, and indeed the examples you used throughout your paper, all are based on particular models (implicit in the "taxonomy"). Unfortunately, ALL MODELS ARE WRONG, at least when you examine them in enough detail, and hence all taxonomies are wrong. When you emerge from the mundane (just where DO I plug in?), taxonomy errors become a substantial fraction of all human error. Unfortunately, your Bayesian approach does NOT allow for taxonomy errors, except where (unknowable?) error terms are incorporated into present models. Interpolation may be pretty safe, but effective action in the world is often based on extrapolation, and it is extrapolation where model errors most often assert themselves. No, I don't have the answers to this puzzle, but maybe I can "blow some smoke" to inspire you to fill this void in a bit: Thinking of the Jovial language, if rules somehow also included the known ranges of validity, then at least results could be tagged as extrapolations and hence remain suspect beyond their Beyesian computations. Perhaps some reasonable heuristic could be found to amplify the uncertainties of extrapolations? Of course, this also applies to the very mundane, e.g. we know that there is a wall extending North, but we don't know how far North it extends. It is a pretty fair bet that it doesn't circumnavigate the globe. How rapidly does our expectation of a continuation of the wall deteriorate as we continue along the wall? Perhaps we should expect as much more wall as we have already seen? It seems pretty absurd to expect every rule to contain the (presently unknown) equations to handle unknown uncertainties in their extrapolations, so it would seem obvious (to me) that the mathematics and/or engine that handles the mathematics should incorporate whatever PFM is needed to amplify the uncertainty of extrapolations. Of course, a taxonomy of the world would take the population of the world just to maintain, with no idea who would ever write such a thing to begin with, so this entire discussion is academic. Also, your paper is a bit too long for most publication forums. Steve On 8/30/08, YKY (Yan King Yin) <[EMAIL PROTECTED]> wrote: > > This is a draft of my paper.... =) > > I need some feedback before releasing it officially.... > > Comments, suggestions, are welcome! > > YKY > > > > ------------------------------------------- > agi > Archives: https://www.listbox.com/member/archive/303/=now > RSS Feed: https://www.listbox.com/member/archive/rss/303/ > Modify Your Subscription: > https://www.listbox.com/member/?&; > Powered by Listbox: http://www.listbox.com > > ------------------------------------------- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244&id_secret=111637683-c8fa51 Powered by Listbox: http://www.listbox.com
