Re: Hawkins ideas on building AI's
From: bkaz (bkaz__at_hotmail.com)
Date: 10/22/04
- Next message: Curt Welch: "Re: The Hard Problem for Behaviorists"
- Previous message: James Landle: "Antonio Damacio"
- In reply to: dan michaels: "Re: Hawkins ideas on building AI's"
- Next in thread: dan michaels: "Re: Hawkins ideas on building AI's"
- Reply: dan michaels: "Re: Hawkins ideas on building AI's"
- Messages sorted by: [ date ] [ thread ]
Date: 22 Oct 2004 16:47:14 -0700
> > Differences in connectivity don't have to be 'preprogrammed', they're
> > learned: Hebbian 'fire together, wire together' principle. Different
> > inputs could probably stimulate even subtle differences in neural
> > composition of the columns.
> Yes, we've had several discussions on this in the past on c.a.p. There
> are some on both sides of this issue. For some of us, it's a little
> hard to imagine that 30 visual areas with different operations, and
> arranged in a more or less hierarchical fashion [regards function],
> could have wired up this way, and the same from individual to
> individual organism. Such as the following, for instance ...
Hawkins gives a good example: people can 'see' with the electrode grid
implanted in their tongue, - bypassing visual cortex altogether.
What happens to all those '30 specialized preprocessing' areas?
Noise is an absense of patterns, you must recognize them to filter it
out.
> Interestingly, while hunting around, I found an earlier thread
> discussing this issue from December 2001, and with several of the
> locals to c.a.p. ... Louis, Curt, Rick, plus "another" Boris.K :) ....
Yeah, that was me. I've given up on c.a.p. since.
Boris
- Next message: Curt Welch: "Re: The Hard Problem for Behaviorists"
- Previous message: James Landle: "Antonio Damacio"
- In reply to: dan michaels: "Re: Hawkins ideas on building AI's"
- Next in thread: dan michaels: "Re: Hawkins ideas on building AI's"
- Reply: dan michaels: "Re: Hawkins ideas on building AI's"
- Messages sorted by: [ date ] [ thread ]