From: Ben Goertzel (ben@goertzel.org)
Date: Mon May 06 2002 - 08:27:48 MDT
Hey James,
Looks like really interesting stuff!
I think that an efficient way of doing Solomonoff induction for modest
program sizes, can definitely form the core of an AGI.
We do something similar in Novamente, though with a different algorithm (our
design involves the use a hybridization of probabilistic inference and
evolutionary programming for program induction)
However, I also suspect there may be a long path from getting the individual
"model" (in your language) to work well, and figuring out how to piece
together a bunch of models into an overall functional mind....
Pei and I long ago made the observation that most of the algorithms
important for mind have exponential time and space complexity, but with a
manageably small exponent for realistic problem sizes. It's interesting
that you've come up with the same conclusion in your work.
This is one of those things that is potentially problematic for
superintelligent AI, of course .. though I expect the problem will be
solvable by the introduction of better (but probably still exponential)
algorithms....
I look forward to seeing the details of your program induction algorithm,
one day!!
-- Ben G
> -----Original Message-----
> From: owner-sl4@sysopmind.com [mailto:owner-sl4@sysopmind.com]On Behalf
> Of James Rogers
> Sent: Sunday, May 05, 2002 10:25 PM
> To: sl4@sysopmind.com
> Subject: So I'm going to start documenting *my* work
>
>
> Hi folks,
>
> As many people here know, I've been working somewhat secretively on my own
> AI framework which takes a fairly unique approach in a couple
> respects. In
> truth, it isn't so much that it is a secret so much as it is that I am too
> lazy to write anything up. :-) A number of people have been
> asking for more
> meat, and so I am going to start posting stuff to the web. I figure that
> eventually I'll have enough stuff up that I won't have to answer
> questions,
> just point to the web.
>
> Since the girlfriend is out of town and after 60 weeks of work at
> the office
> I don't feel like writing code, I've started writing things up and cutting
> and pasting from various sources. Note that the location is temporary;
> don't count on it being there in a year.
>
> So here is a somewhat meandering (and utterly improper) architecture
> overview doc:
>
> http://jamesr.best.vwh.net/ARCH.html
>
> Hopefully the formatting is okay, as it is obnoxious MS output (though it
> looks fine in both Nutscrape and IE). Someday I will re-hack it into a
> proper web doc. This will give people some additional stuff to
> chew on. I
> actually have several more pages of relevant stuff I wrote this
> weekend, but
> none of it (including the one above) have been edited or organized, though
> I'll probably redo it eventually.
>
> I'll also be putting up miscellaneous software libs and DSP plugs
> (for audio
> processing) that I've written that the world could use, so there
> may be some
> other neat (but unrelated stuff) up there soon.
>
>
> Cheers,
>
> -James Rogers
> jamesr@best.com
>
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:38 MDT