On 2012/07/21, at 17:13, Mike Tintner wrote: > But Plato didn't say what it was, Hawkins hasn't said what it is, in fact no > one has said it ... and it doesn't exist.
OK, I can admit that invariant representation doesn't exist. Then how about thinking pattern recognition function to be invariance, while there would be no such thing as invariant *representation*? Of course, you can deny this too, but then any epistemology might not hold, or if you are that radical, I would not add words for it... -- Naoya A. On 2012/07/21, at 17:13, Mike Tintner wrote: > Arakawa:When a neural system learns some pattern, say that of a line segment, > it recognizes line segments regardless of their orientation or length > (hence 'invariant"). > > What Sergio was saying was : "wait a moment, do I really understand this - > what Hawkins is saying...?" > > To do that, you first have to say as you have done: "ok what say would be an > invariant representation of a line...?" > > At this stage, you can casually float over the problem, and think: " oh well, > there must be an invariant representation of a line... stands to reason" > > But if you take the second step, and actually start thinking about different > kinds of line, and what could possibly be an invariant representation of them > all - that could be transformed into them all - you will find there is no > such invariant representation - and Hawkins has neither posited one in > relation to any object, incl. Jennifer Aniston (or a line), nor explained how > it could be universally transformed into any variation of a given object. > > Similarly, people are thinking : "of course the world consists of patterns.. > stands to reason.. I know that this is a "street" and by god it looks like > that "street" to me, and that other one - I recognize all these "streets" > without a problem - therefore there must be a common "pattern",invariant > representation" which enables me to recognize them all. > > But people never go to the second step, and start thinking about what form > those patterns (of things like a street) could take. If you do, you'll find > there are neither such patterns in the world, nor in your mind. > > That's not how the brain achieves its magic feat of recognizing the > similarity between diverse forms, objects and scenes. > > Everyone here tends to be Platonist - "of course there is an essential idea/ > invariant representation/ pattern for objects - an essential "chair".." > > But Plato didn't say what it was, Hawkins hasn't said what it is, in fact no > one has said it ... and it doesn't exist. > > -------------------------------------------------- > From: "ARAKAWA Naoya" <[email protected]> > Sent: Saturday, July 21, 2012 3:10 AM > To: "AGI" <[email protected]> > Subject: Re: [agi] Re: How the Brain Works -- new H+ magazine article, by me > >> On 2012/07/21, at 4:59, Mike Tintner wrote: >> >>> Sergio: I noticed that Jeff Hawkins in On Intelligence writes about >>> "invariant representations," which are hierarchies, but never >>> explains how they come into existence. I am just a little confused. >> >>> I wonder whether you have an outstanding point there. Everyone >>> *talks* about "invariant representations". Does anyone anywhere >>> have any AI-worthy explanation of their nature/origin whatsoever? >>> >>> (Of course, invariant representations overlap with concepts. There >>> are psych/phil. explanatory theories of concepts, but that's why I put >>> in "AI-worthy". I suspect they are all v. vague). >> >> I interpreted "invariant representations" in the writing of Hawkins as >> learned patterns. >> When a neural system learns some pattern, say that of a line segment, >> it recognizes line segments regardless of their orientation or length >> (hence 'invariant"). >> "Invariant representations" in a neural network would be distributed >> so that one cannot point out saying, for example, *this* is the >> representation of a line segment... >> >> * The Gibsonian invariance might be a different notion while he may >> have made the term popular among cognitive scientists (?). >> -- >> Naoya ARAKAWA ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968 Powered by Listbox: http://www.listbox.com
