Re: [agi] The Smushaby of Flatway.Hi Ed, (and Matt too) If you'll be patient with me, I'm going to work longer and truly formalise my reply here - distinguishing among other things between Rational Logic (as we know it) and a new Creative Logic, (which computers are currently, though not necessarily, incapable of).
As you can gather, I let myself be goaded by Richard's/Mark's stuff, and replied before I was quite ready. If you want a quick, dirty reply, what I maintain and what you must therefore provide counter-evidence against is that no AI or AGI to date, or any formal tool of theirs, has any capacity to deal with any creative problem, high- or low-level - all of which demand the readiness to switch domains. (Creative problems, at a lower level, by the way include at least half the problems in education, and the majority of everyday practical problems that you face). So no rational logic or existing formal mode of thought, I maintain, can solve firstly the creative, engram problem Richard was dealing with of: "how does the brain record information?" Just start on it - which was what I tried to get Richard to do at a secondary, metacognitive, self-aware level - and you'll quickly find that no structured approach to creative problems is possible. More specifically, no pre-structured form of thinking can tell you how to proceed systematically to:find and choose - which evidence (from the vast diversity of available sources) to gather -which possible *sources* of evidence to consider -which ways to identify available sources of evidence, -which ways to experiment to find new evidence - which existing, on-the-table engram ideas to consider, or whether any of them are worth pursuing, (say goodbye GA's), -- which, existing non-neuronal forms of recording information that might possibly yield clues, to consider - which of the vast number of ways possible, should you choose, to order, and cut between, these different sub-problems let alone: - how to develop altogether new ideas about the creative problem, of how engrams might work - and which parts of cells/brains etc might be involved There are simply no reliable premises about any of these that you can confidently start from in any given creative problem . And if you do try to solve any creative problem, you will keep getting stuck, as you encounter metaphorical blank pages. And remember that the *same* approach (which might include adaptive principles) must be generally applicable to solving other and new creative problems. If you have a rational method, it must apply not only to the problem of engrams, but also "how does the binding principle of consciousness work?" or "find an unknown cause of any kiind of cancer - , or how, say, do tumours grow so fast". or "why do some advanced tumours disappear without treatment?" (Or how should Matt go back and rework his last paper, or develop a new compression program?) And no existing approach to AGI is capable of finally coming up with a creative solution to such problems - one that crosses unexpected domains. No rational approach could come up with a creative (if probably completely wrong) idea like that of Sheldrake about the engram problem, that actually it isn't the brain at all that records information, but rather some ethereal domain "out there" - and even produce some new kind of evidence for that idea, that is at least worth considering.for a brief,momen or two Put all this another way - there is, as Ben says, no formally definable scientific method. Well, that's what I maintain. If you can propose any half-way reasonable, formal method of approaching creative problems, (or part thereof that shows the slightest promise), you will have provided some evidence that current approaches to AGI, have even the slightest chance of working. But don't break your head - as De Bono, and thousands of pyschologists of creativity will also tell you - "Logic is the tool that is used to dig holes deeper and bigger, to make them altogether better holes ... lateral (creative) thinking is trying again elsewhere" - in a new, non-rationally-predictable/deducible hole/domain." Kauffman will also tell you that in Inventing the Sacred, over and over. But AGI-ers aren't basically interested in the nature of creative problems and where their holes are going, only in what logical, digging instruments they should use. As De Bono also says: "No matter how obvious this may seem to every digger, it is still easier to go on digging in the same place than to start all over again in a new place." Hence, a vast number of failed projects ----- Original Message ----- From: Ed Porter To: [email protected] Sent: Friday, January 09, 2009 7:39 PM Subject: Re: [agi] The Smushaby of Flatway. Mike, What is the evidence, if any, that it would be difficult for a sophisticated Novamente-like AGI to switch domains? In fact, much of valuable AGI thinking would involve patterns and mental behaviors that extended across different domains. Human natural language understanding is believed to use multiple domains of knowledge in the brain when necessary, such as visual imagination, to help in understanding what is being said or what is to be said. The OpenCog WikiBook describes multiple procedures for controlling, and automatically learning to control inference, activation, and procedure execution. This could be used to accomplish sophisticated interaction of knowledge from different domains, much as the human brain does. Brain studies conducted by Wolf Singer indicate that the brain synchronies can be used to interconnect portions of the brain with different areas of expertise when performing a job that requires one of those areas to tune into information coming from the other. It would be easy to make an AGI do something equivalent. In fact, in routine tasks, the synchronies are often set up in advance of there being an activation in some of the regions connected by them, because the brain has learned from prior inferencing patterns to expect such activations to arise given the task being performed. Ed Porter ------------------------------------------- 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=123753653-47f84b Powered by Listbox: http://www.listbox.com
