Re: [agi] My proposal for an AGI agenda

2007-04-09 Thread Jeff Rose

Philip Goetz wrote:

On 3/23/07, Samantha Atknis [EMAIL PROTECTED] wrote:

 8,Fast where most of the processing is done.

 In the language or in things written in the language or both?  Lisp has
been interpreted and compiled simultaneously and nearly seamlessly for 20
years and has efficiency approaching compiled C in many problem domains.


Samantha, you need to provide me with references if you want me to
believe this.  No LISP compiler has ever been optimized to any serious
degree AFAIK.  The nature of the language makes it difficult to write
efficient code in the first place.  And I suspect that these many
problem domains don't include any that involve numeric calculations.

-
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/?list_id=303



Here are some benchmarks with SBCL lisp and GCC compiled C:

http://shootout.alioth.debian.org/gp4/benchmark.php?test=alllang=sbcllang2=gcc

Compared to other high level languages lisp is pretty impressive, but 
it's not at the level of C for sure.


-Jeff

-
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/?list_id=303


Re: [agi] Why evolution? Why not neuroscience?

2007-03-25 Thread Jeff Rose

Eugen Leitl wrote:

Why evolution? Why not neuroscience?

...

I think that the ultimate source of intelligence is the process that
generates algorithms. That process is evolution, in all its self-adaptive
recursive evolvability-evolving glory. Biological evolution is a
process-generating process; an algorithm that generates algorithms (that
generates algorithms that...).


How about another metaphor.  Imagine that we find a faster than light 
space ship one day.  Of course everyone will want to understand how it 
works so that we can try to build some of our own.  Now you could take 
it apart piece by piece, watching various bits and pieces while 
functioning, and then analyzing them in isolation.  Recreate pieces, 
sub-units, assemblies etc., until finally arriving at functional units 
that can be put together into a new ship.  Another strategy would be to 
say, ok, we know the basics of physics and all of these materials we 
have found, now lets explore the space of every possible arrangement of 
these things things until we arrive at something that kind of does the 
same thing.


I think it was an Baum's book, where he discusses the immense number of 
parallel trials that have been occurring on earth since life arose. 
Evolution takes such an incredibly large number of iterations, it seems 
like the wrong way to get somewhere if you already know where you want 
to go.  Even in the history of life on earth, we haven't even scratched 
the surface in terms of exploring the total space of possible proteins, 
and working within the constraints of a limited set of possible base 
pairs, codons etc. is already cutting the space considerably.


If you have a working model right in front of you, without a doubt it 
makes sense to understand it as well as you can.  Evolution is a good 
exploration strategy that is well worth putting lots of effort into 
further research and development, but I don't think it is likely to 
stumble upon AGI unless it is being used to search a highly constrained 
parameter space where it is really just tuning an already engineered design.


Ciao,
Jeff

-
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/?list_id=303