On 05 Apr 2013, at 11:17, Telmo Menezes wrote:

On Fri, Apr 5, 2013 at 1:09 AM, meekerdb <meeke...@verizon.net> wrote:
On 4/4/2013 3:50 PM, Telmo Menezes wrote:

On Wed, Apr 3, 2013 at 10:44 PM, Jason Resch <jasonre...@gmail.com> wrote:



On Sat, Mar 30, 2013 at 7:58 AM, Telmo Menezes <te...@telmomenezes.com >
wrote:




On Thu, Mar 28, 2013 at 1:23 PM, Craig Weinberg <whatsons...@gmail.com >
wrote:



Then shouldn't a powerful computer be able to quickly deduce the
winning
Arimaa mappings?


You're making the same mistake as John Clark, confusing the physical computer with the algorithm. Powerful computers don't help us if we
don't
have the right algorithm. The central mystery of AI, in my opinion, is
why
on earth haven't we found a general learning algorithm yet. Either it's
too
complex for our monkey brains, or you're right that computation is not
the
whole story. I believe in the former, but not I'm not sure, of course. Notice that I'm talking about generic intelligence, not consciousness,
which
I strongly believe to be two distinct phenomena.


Another point toward Telmo's suspicion that learning is complex:

If learning and thinking intelligently at a human level were
computationally
easy, biology wouldn't have evolved to use trillions of synapses. The
brain
is very expensive metabolically (using 20 - 25% of the total body's
energy,
about 100 Watts). If so many neurons were not needed to do what we do, natural selection would have selected those humans with fewer neurons and
reduced food requirements.

Yes but one can imagine a situation where there is a simple
(sufficiently-)general purpose algorithm that needs some place where
to store memories and everything it has learned. In this case, we
could implement such an algorithm in one of our puny laptops and get
some results, and then just ride what's left of Moore's law all the
way to the singularity. We don't know of any such algorithm.


But it doesn't follow from human brain complexity that no such algorithm exists. Evolution doesn't necessarily do things efficiently. Because it can't start-over, it always depends on modification of what already works. But I think there are other theoretical and evolutionary reasons that would
limit the scope of general intelligence.  Just to take an example,
mathematics is very hard for a lot of people. Mathematical thinking is not something that has been evolutionarily useful until recent times (and maybe
not even now).

Agreed.

What puzzles me the most is not that evolution hasn't found it
(although we're not sure, there's a lot we don't know about the brain
still). It's that the swarm of smart people that have been looking for
it haven't found it. I still have some hope that it's simple but
highly counter-intuitive.

All recursively enumerable class of total computable function is learnable. There is a simple algorithm: dovetail on that class, and output the programs which match the data. This will converge (in the computer science sense = eventually output something correct) to a program explaining the input-outputs given. That algorithm will already not work if they are strictly partial function in the class, and it will not work on non recursively enumerable class (like all total functions). Such an algorithm is of course not really a practical algorithm, but they can be accelerated, --- even more so, if we allow weakening of the identification criteria. Some previews knowledge and rule of thumbs can also accelerate them, with non computable (= huge) gain.

The field of theoretical learning is very rich, and leads to the idea that competence is something never really universal, quite domain dependent, and speedable when allowing the usual things that evolution exploited all the time: the making of error, randomness, team or swarm of machines, etc. It leads to a large variety of possible implementation of competence, and exploits maximally the high general intelligence which exists already in the any universal machine. But as I said, such high competence can also restrict that intelligence. Intelligence is needed to develop competence, but competence tends to make that intelligence blind. With language and culture, that blindness can pass from a generation to another, so that babies can became more quickly stubborn than without, until the next paradigm shift.

Bruno





Telmo.


Brent


--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to everything-list+unsubscr...@googlegroups.com.
To post to this group, send email to everything- l...@googlegroups.com. Visit this group at http://groups.google.com/group/everything-list?hl=en .
For more options, visit https://groups.google.com/groups/opt_out.



--
You received this message because you are subscribed to the Google Groups "Everything List" group. To unsubscribe from this group and stop receiving emails from it, send an email to everything-list+unsubscr...@googlegroups.com.
To post to this group, send email to everything-list@googlegroups.com.
Visit this group at http://groups.google.com/group/everything-list?hl=en .
For more options, visit https://groups.google.com/groups/opt_out.



http://iridia.ulb.ac.be/~marchal/



--
You received this message because you are subscribed to the Google Groups 
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to everything-list+unsubscr...@googlegroups.com.
To post to this group, send email to everything-list@googlegroups.com.
Visit this group at http://groups.google.com/group/everything-list?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to