EFTA02489725.pdf

From: on behalf of Ben Goertzel Sent: Monday, September 7, 2015 2:45 AM To: Jeffrey E. Cc: Joscha Bach Subject: Re: On Mon, Sep 7, 2015 at 10:04 AM, jeffrey E. cjeevacation@gmail.com> wrote: > im aware of the goals . i have yet to see deliverables. . Fair enough.... But we have learned and understood a lot through our work so far, which is important.... I understand the problem far better than I did 15 years ago and my planned solution is sketched out in far more detail, and I have a better team at work on the problem... All this anticipation will make the victory even sweeter once it arrives ;-) > babys learn > any one of a multitude of languages. this is less about intelligence and > more about brain architecutre . Architecture is an aspect of intelligence ---- in essence, an architecture encapsulates a prior distribution over possible worlds, which helps guide action to be intelligent (assuming that the worlds where the actions occur, are reasonably likely ones in the distribution implied by the architecture) >how do brains take signals, minimize > noiise and make cohenrent mental objects. images. language. tounch > smell. temporal ccomponents. maybe music analogy. . structure of > language. , illusions in each sense. ambiguity in each. ? ( maybe an > insight). im going to clean my house. ( interior > exterior?. im going > to paint my house. exterior. . visual lady and vase. ambiguity might > lead to underlying process.. it is easy to choose which note does not > beling to a new musical piece. . heard for the first time.? odd. The particular cases may feel odd to us from our subjective views, but the general principles are not hard to see.... Brains recognize patterns in their inputs, actions and states; and coherent mental objects are complex, self-consistent, compact bundles of interrelated patterns.. The challenge is to get all this pattern recognition to work effectively given available computational resources, using hardware and software infrastructures that were basically made for other things. I believe my analysis of the problem, as presented in "Engineering General Intelligence", reduces AGI to a large-scale engineering problem plus a dozen or two PhD theses worth of focused research work. On the one hand, that's cool. On the other hand, it still leaves a lot of work. Speaking of odd -- last week I had a 3-hour meeting with the Prime Minister of Kazakhtan, in Astana, about the possibility of setting up an AGI research center there. He is personally very pumped about it, however their currency has been plummeting recently due to the down-trend of oil prices, so it may be the wrong time. It's cool EFTA_R1_01610388 EFTA02489725 that national leaders are now taking AGI seriously though. He's a very smart and knowledgeable guy, deeply into science of all sorts.... Aubrey de Grey has has helped them set up a center for regenerative medicine research there... on another note what happend to gino, he went dark!? Gino can never go dark, he is too enlightened for that !! ;D He is mostly in Shanghai lately, he seems to have found some well-connected people there who are interested to support consciousness research. They are setting aside a building for it, with a conference center and a bunch of "free for researchers" apartments, etc. He seems very psyched about it, but I haven't been for a visit or investigated it closely.... I guess he'll get more active online when he comes back to HK for a longer spell... -- Ben <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd"> <plist version="1.0"> <dict> <key>conversation-idgkey> <integer>105457</integer> <key>date-last-viewed</key> <integer>0</integer> <key>date-received</key> <integer>1441593888</integer> <key>flags</key> <integer>8590195713</integer> <key>gmail-label-ids</key> <array> <integer>6</integer> <integer>2</integer> </array> <key>remote-id</key> <string>540308</string> </dict> </plist> 2 EFTA_R1_01610389 EFTA02489726