Re: simple, plain-English introduction?
- From: Randy <joe@xxxxxxxxxxxxxxx>
- Date: Tue, 13 Dec 2005 13:29:15 -0600
Brian Janko wrote:
> Is there a simple, plain-English introduction to the limits of
> artificial life, consciousness, and the possible threshold where
> robots/computers could possibly cross over to being autonomous, having
> freewill and/or emotion? I'm not a scientist. I have some computer
> background, but it would be easier to understand the detailed technical
> info online if I could get a better grasp of the Overview first.
>
> Thanks,
> Brian
I think alife has little to do with strong AI. From my readings, alife
is a colloquial term that refers to any form of computing that can be
seen as analogous to biology, in essence, software that "behaves" in
some way that living systems do. IMHO, such an enterprise is much too
unfocused and insubstantial to explore the boundary to something as
subtle and elusive as the human level of intellectual competence.
What's the threshold for consciousness? That's a more precise question,
but the answer (and even the terminology of the question) has changed a
lot in the past 40 years. Frankly, I'm doubtful anyone can answer it
precisely, accurately, with any confidence. We've been unable to build
"smarter" software without explicitly embedding more and more knowledge.
We've accomplished even less in also integrating the necessary
cognitive mechanisms or learning facilities into a single
proto-conscious agent.
So I'd venture guess that the fundamentals to "self-realization" will
certainly involve: 1) developing a rational cognitive mechanism, 2)
implementing a flexible extensible knowledge representation framework,
and 3) lots of of reinforcement learning and autonomous discovery by the
agent.
Just like a real person. Is a human born conscious? If not, then when
does it arise? What prerequisites are needed to become conscious? Is
it only because, at some point in our development, each of us develops
sufficiently cognitively and learns enough about the world to realize
that we exist apart from the rest of it? If that's when we officially
become conscious of the self, our state-of-the-art agents have a long
way to go before achieving self-realization (satori :-).
I suspect alife will contribute little to developing strong AI, and I'm
no longer confident that symbolic AI ever will. IMHO, real biology is
likely to be our best source of revelation. It seems more productive to
base strong AI *not* on the emergent behaviors of living (or synthetic)
systems, but on the mechanisms that make real critters go (observing
what biology does, as well as how it does it). The cognitive machinery
probably doesn't *need* to be wetware, but until we have a better idea
of what constitutes a UHM (Universal Human Machine), we'd probably be
wise to pay closer attention to what does work and speculate less about
what could/should/might.
After all, strong AI does have an existence proof.
For a useful start in appreciating AI, you might look to the Wikipedia
entries for AI and alife. Then Peter Norvig's web site is a great place
to continue your journey:
AI on the Web
http://aima.cs.berkeley.edu/ai.html
Randy
.
- Follow-Ups:
- Re: simple, plain-English introduction?
- From: B1ackwater
- Re: simple, plain-English introduction?
- References:
- simple, plain-English introduction?
- From: Brian Janko
- simple, plain-English introduction?
- Prev by Date: Re: simple, plain-English introduction?
- Next by Date: Chat or Post
- Previous by thread: Re: simple, plain-English introduction?
- Next by thread: Re: simple, plain-English introduction?
- Index(es):
Relevant Pages
|
|
