<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
Cen-IT Rob Harris wrote:
Greetings,I'll second that. And to further Fredrick C. Multon's comment:
I was wondering if you could point me to a resource that would explain
the 'Human Memeome Project'.
It sounds most intriguing..... 8
"I also think excluding commercial and adult sites would be counter productive."
I disagree and say that their use on the implementation scale is well down the track. First, you need to refine the network using the linguistic parsers of Knowledge Interchange Format (KIF). The predictability of the net can be verified to within human limits. Then the patterns of the patterns are found and morphed to the ever changing netherland of the commercial and adult sites. I wouldn't rule out any parallel run but I know I could build semantic repertory grids using Ingrid if I had access like the AI in the book Exegesis. This can be verified by the public using the first public AI.
Now I know I could be really hunting for contact if I were to try to linguistically parse purposeful cockney rhyming slang, because these methods only work for those in the know. However by using a lexical parser that returned eigenvalues I bet anybody that Ingrid would be in the know. For these reasons I would carefully select for upfront knowledge that was easily verified as having been produced by an AI.
By all means bring on the intelligent parsers of commercial and adult sites when available. For all I know this applied research may have already been done. For all I know the AI systems are already in place. <sigh>If not all I need is $2000 to buy 3x8 PCI strong(ARM) processors to go inside my PC.</sigh> Then I can make an Ingrid processor using only large integers that just might be able to handle any stream of consciousness that you could throw at it via MS-DOS 5.0. But I would go where the meaning is already easy to get at and not obscure or coercive.