On 22 Sep 2012, at 15:29, Evgenii Rudnyi wrote:

On 22.09.2012 14:58 Bruno Marchal said the following:

On 21 Sep 2012, at 21:27, Evgenii Rudnyi wrote:

On 19.09.2012 00:57 meekerdb said the following:
On 9/17/2012 11:27 AM, Evgenii Rudnyi wrote:
Do you mean that the meaning in a guided missile system
happens as by-product of its development by engineers?

To me, it seems that meaning that you have defined in Mars
Rovers is yet another theory of epiphenomenalism.

And your quote and question are yet another example of
"nothing buttery" and argument by incredulity.

Brent


I am not sure if I understand you. I am not saying that I am
right but I really do not understand you point. You say

"Consciousness and computation are given their meaning by
their effecting actions in the world."

and it seems that you imply that this could be applied for a
robot as well. My thought were that engineers who have design a
robot know everything how it is working.

But they don't a robot, even one as simple as a Mars Rover
perceives and acts on things the engineers don't know.  A more
advanced robot will also learn from experience and become as
unpredictable as a person from the engineer's standpoint.


Okay, let us take more advanced robots. I guess that

Dario Floreano and Claudio Mattiussi, Bio-Inspired Artificial
Intelligence: Theories, Methods, and Technologies

should be perfect here. You will find in the book about learning in
behavioral systems. Yet, the authors do not use the term
consciousness at all. They even talk about intelligence just once

Conclusion, p. 585 : “A careful reader have noticed that we have
not yet defined what intelligence is. This was done on purpose
because intelligence has different meanings for different persons
and in different situations. For example, some believe that
intelligence is the ability to be creative; other think that it is
the ability to make predictions; and others believe that
intelligence exists only in the eye of the observer. In this book
we have shown that biological and artificial intelligence manifests
itself though multiple processes and mechanisms that interact at
different spatial and temporal scales to produce emergent and
functional behavior. The most important implication of the
approaches presented here is that understanding and engineering
intelligence does not reduce to replicating a mammalian brain in a
computer but requires also capturing multiply types and levels of
interactions, such as those between brains and bodies, individual
and societies, learning and behavior, evolution and development,
self-protection and self-repair, to mention a few”.

Hence, again let us imagine that a robot with artificial neural
networks developed as described in the book can learn something
indeed. In the book there are even examples in this respect. Yet,
the engineers developing it have not even thought about
consciousness. Hence, in my view, if consciousness happens to be in
such a robot, then we could talk without a problem about
epiphenomenalism. Why not?

I have seen a project where engineers at least talk about a module
QUALIA

http://www.mindconstruct.com/

“MIND|CONSTRUCT is developing a ‘strong-AI engine’, a so called
AI-mind, that can be used in (human-like) robotics, healthcare,
aerospace sciences and every other area where ‘conscious’
man-machine interaction is of any importance.

The MIND|CONSTRUCT organization is the culmination of many years in
AI-research and the so called ‘hard-problems’, and the application
of elaborate experience in knowledge-management, for the design and
development of a ‘strong-AI engine’.“

If consciousness happens here, then we could at least find that it
was planned this way.

It is part of what a machine is that we cannot know what we are doing
in building them, so human might as well build a conscious machine
without knowing it; except later, when the machine complains or fight
for its right. Comp is rather negative on the idea of programming
consciousness. We can only let consciousness manifest itself, or not.
Or we can copy intelligent machine, partially or completely.

Then, I am afraid, comp is of no help to the AI community as it seems cannot guide engineers on how to develop an intelligent robot.

It can help to see how to not develop intelligent robot, like programming them. It helps in the negative, like the proof of irrationality of 2 helps to not search for two such numbers with ration equal to sqrt(2).

The whole of theoretical AI is full of such negative, provably non constructive, theorems.

It helps to understand that "intelligence" is not a normative notion, and that if machine becomes intelligent, it might not been known in advance, and it is more a question of letting them explore, than obeying specific instructions.

Comp helps to be less naïve both on machine and human intelligence, and it separates also the notion of consciousness from the notion of intelligence.

Then our goal here is to get a TOE, not to help engineers in making an intelligent machine. But still there is that important negative help described above. Eventually comp can help humans to develop some more respect for machines ... and humans.

Bruno

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



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

Reply via email to