----- Original Message ----- From: "Ben Goertzel" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Wednesday, December 11, 2002 9:08 PM Subject: RE: [agi] TLoZ: Link's Awakening.
> > The idea of a 2D UI in which the user and the AI can both create stuff, is > part of our idea for a "Novababy" teaching/training environment, and also > part of Peter Voss's A2I2 teaching/training environment. It's a > "natural".... > > The inclusion of games in this environment is also a good idea... I am eager > to see what Michael Roy Ames will come up with in terms of a series of > progressively more & more difficult games for young AGI's. > > -- Ben Think of it as convergent evolution in action; the ai architecture I have been working on for the past month uses a similar idea, in my case it's a sort of simplified simulation of real life (something like the Halo physics engine would probably work great for this). my design is 'embodied' within its world; all input is through simulated sensors, all out put through somatic effectors. At initialization, the ai will be much like a baby; almost totally helpless, but with built-in components which direct it's mental growth towards the development of intelligence and self awareness. Things we have like curiosity and the innate need for attention and parental "love" I've just uploaded some stuff to my site, so now all the basic components of my design and a rough-draft spec. of it's system-level architectue are on my site. The neural processing stuff discussed above: http://users.rcn.com/standley/AI/neuralnet1.jpg 367K http://users.rcn.com/standley/AI/neuralnet2.jpg 210K http://users.rcn.com/standley/AI/neuralnet3.jpg 303K and the architecture overview: http://users.rcn.com/standley/AI/sysarchitecture.jpg 296K now 'all' I have to do is put all this stuff together into a single, coherent document, then I can start fleshing out the system architecture and some of the poorly explained/illustrated basic components, as well as draw up the specifications for the symbolic knowledge representaion system I've designed for the ai (though it could be used for almost anything) also, has anyone seen this yet? http://www.inform.umd.edu/EdRes/Colleges/ARHU/Depts/English/englfac/MTurner/ cin.web/cin.html Conceptual Integration Networks - I haven't finished reading the paper as i just got this link yesterday, but what I have read intrigues me. IMHO anyone working on an ai project should at least give it a quick look J Standley http://users.rcn.com/standley/AI/AI.htm ------- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED]
