Contra “deep learning”

Post #424,455 by rcareaga 6/23/18 12:57:17 AM 6/23/18 12:57:17 AM Reply	Contra “deep learning” NYT article about a different approach. I don’t know that I’m competent to evaluate it. cordially,
Post #424,456 by drook 6/23/18 7:23:35 AM 6/23/18 7:23:35 AM Reply	Combination of marketing and motivated reasoning "That method of so-called AI isn't real AI," say the owners of a company developing a different method of what they say is real AI. -- Drew
Post #424,457 by a6l6e6x 6/23/18 10:08:21 AM 6/23/18 10:08:21 AM Reply	There's more than one way to skin a cat. Prolog goes back to the days when processors were slow and memory limited. I never had the need to get into it although I did do some LISP programming as a student. The "deep learning" approach relies on huge processing power, huge storage and sample data availability. Apparently, once a solution to a given problem is found and understood it's also possible to undermine it. Alex "There is a cult of ignorance in the United States, and there has always been. The strain of anti-intellectualism has been a constant thread winding its way through our political and cultural life, nurtured by the false notion that democracy means that "my ignorance is just as good as your knowledge." -- Isaac Asimov
Post #424,462 by Ashton 6/23/18 9:19:19 PM 6/23/18 9:19:19 PM Reply	OTOH.. you can have all the Gigaflops/sec and make dbases accordingly ..and your 'intelligence' will be pseudo; no more human-like than: an AI-BookkeeperAccountant. Having no Persona the collection can never answer any of the larger-Why??s (and lots more.. but..) tl;dr: until there is developed AE aka Artiificial Emotional-intelligence--as the absolutely Necessary corollary to all the AI-Wonderfulness being tauted --you Won't have much more than .. ... {see above}. Now as-to ARTIFICIAL CONSCIOUSNESS being right around the corner: I no words on that (..prolly for my duration within this Chicken-outfit. Or a Species-change?)
Post #424,463 by boxley 6/23/18 10:22:51 PM 6/23/18 10:22:51 PM Reply	example scan every known breed of cat in every known color then scan pictures of cats the machine can identify them. scan a picture of a pink house cat. Chances are the system will not recognize it as a cat. Do the same with humans, a guy will look at a woman in the room and wonder if she has a shaved pussy, he will then notice a different woman with hair died pink then correctly identify the pink picture as a pink pussy cat. Until they devise a way for non linear synapse burns, AI will just be an efficient data set compiler. "Science is the belief in the ignorance of the experts" – Richard Feynman

Welcome to IWETHEY!