PM, What you’re saying is “this is what we do.." [subtext: what else can I do?]
What I’m telling you is – it doesn’t apply to real world problems – 90-odd per cent of problems. And this is demonstrable – you’re not going to produce any real world problems to which space/set/frame reasoning actually applies or is applied. So what does apply? I have barely thought about this particular aspect of the/a brain– i.e. the organization of its information/knowledge. But clearly if you look at real world agents, what you get is something like “web formation” as opposed to “[organized] network formation” of knowledge – IOW facts and data are accumulated pell-mell by a real world agent living in the real world, in a very loosely connected, *unstructured* way as opposed to a perfectly organized, systematic way. ***And this is the actual visible external reality – and major, generally acknowledged issue – of the world wide web – (which is now a great external appendage of all our brains)***. You have the vast mass of information on the WWW in the form of AI-acknowledged “unstructured data” and only very small segments of information organized in neat, systematic local networks. Again the actual reality of dealing with the WWW - which all of you agree will ideally be something an AGI should eventually do – is that you have to deal with unstructured, disorganized data – find your way around – **explore** not do pre-organized, pre-guided, computational *searches*. So PM, I suggest, stop being infant-like and arguing: “but this is the only technique that anyone has given me! boo hoo"! “ Develop your own techniques. Man up. The WWW gives you an externally visible, externally analysable way of how unstructured data are both collected and “browsed” – and provides a pretty good analogy and model (if no doubt with limitations) of how an individual AGI agent, real or proposed, does and most form and collect and *explore* rather than search its information stores. Be a bit creative. AGI – real world reasoning – is about EXPLORING [“browsing”/”surfing”] unstructured information as opposed to SEARCHING structured information [strictly narrow AI] – bleeding obvious, no? From: Piaget Modeler Sent: Tuesday, February 05, 2013 11:39 PM To: AGI Subject: RE: [agi] Multiple Vantages Can Be Used to Find Multiple Observation Objectives Mike Tintner, as a philosopher I respect your attempts to characterize the nature of AGI, it's attributes, and properties. But being a philosopher is not the same as being an engineer, where you apply known methods, or create new methods to construct an actual AGI. So, if in your experience, when you think about thinking you do not "form" problem spaces. That's one matter. But you overstep your philosophical limits when you insist that Engineers should not use a known method such as the "Problem Space Hypothesis" in constructing an actual AGI. To say you don't think AGI works that way is strictly your opinion. Perhaps we are not attempting to build your Real World Reasoner, but instead buld something that is indeed an AGI which operates using a different set of principles. Your RWR principles and requirements are not the only ones. I on the other hand can see from my own experience that formulating new problem space may indeed be how humans resolve problems. I can see that people have a vast repository of concepts and transformations from which they may select relevant objects and operations to form a problem space to reduce the distance from an initial situation to a goal situation. Problem space formulation and search may well be significant components to an AGI, in my humble opinion. ~PM -------------------------------------------------------------------------------- From: [email protected] To: [email protected] Subject: Re: [agi] Multiple Vantages Can Be Used to Find Multiple Observation Objectives Date: Tue, 5 Feb 2013 21:52:32 +0000 PM: Perhaps formulating the problem space is part of the solution. There is no problem space in a real world problem. This is a fundamental and near universal misunderstanding. ***The *world wide web* is the “problem space”.*** IOW more or less everything that exists in the world – the world wide web of real bodies - *may* pertain (and to some extent will pertain) to a real world problem – and even Jim and Ben with their far-ranging minds cannot define “everything that exists in the world.” (past, present and future). In fact, neither can our entire cultural organization (e.g. sciences/scientists). Why is a woman attracted to a man in an office? One factor could be butterflies flapping wings in Brazil that changed the weather in New York to sunny that changed her disposition. Another factor could be her period and the supply of tampons. Another factor could be the news on TV. Ultimately everything in the world around her has some bearing however limited on her attraction. And everything in her mind – her entire mental organization – has some bearing on it. And everything in the history of the world has some bearing. Plus – every situation is different and non-formulaic to some extent. Why she is attracted today will not follow exactly the same principles as her attraction last month, or other women’s attraction in other offices. Psychologists, historians and artists analysing her attraction will be able to endlessly produce new analyses citing new factors – just as artists can endlessly produce new versions of the story of Romeo & Juliet’s attraction.. AI has metacognitively failed to understand the real world reasoning principles of science and scientific models – and indeed of logic and mathematical models. It is a fundamental principle of science that scientific models are artificial selections of a set of factors affecting a situation – from a world – a world wide web - of possible factors. A real world reasoner trying to explain a problem situation, asks: “what in the world could have brought this about?” “What on earth could have...”? Not “what in my handily pre-prepared set of options could have caused this?” An AGI must be able to search the real world wide web for solutions just as any human can search the information world wide web.for solutions. (Try defining “the world wide web” that is the internet in terms of a problem and solution space – and remember that we all now do indeed search the world wide web/net daily for solutions). AGI | Archives | Modify Your Subscription AGI | Archives | Modify Your Subscription ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
