> Here is a conjecture on which I very much would like to hear the opinion
> of members of principia cybernetica.
Hello, Walter !
Considering your proposition that any intelligent system must have a
sensorial
aparatus similar to human aparatus, I would like to comment on it.
First of all, I follow the thought of Cliff, which said that before
concerning
reproducing the type of intelligence an human is able to perform, we must
first try to mimic the lower levels of intelligence we are able to find in
other
types of animals. After we are able to reproduce this type of knowledge,
them we MAY propose ourselves reproduce human-type of intelligence.
I agree also with Sharov, when he says that the best way of developing
such systems is using virtual worlds, where we may grow intelligent
capabilities and test them until we are convicted that they are OK.
Von Uexkull has a very interesting idea that he calls "Umwelt", or in
other words "sensible environment". Strictly speaking, the "Umwelt" is
the world a simple animal is able to sense, and so it is the world that
effectively exist for such animal. All its internal knowledge is related to
such a world, and not to the "real world". Even we, developed animals,
also have our "Umwelt", that is, this world we believe do exist around us.
If
the world is more than this, we are not able to tell, because our sensorial
devices are only able to sense this "Umwelt". So, corroborating your
hypothesis, it seems that any type of intelligence must be grounded on
a species of "Umwelt". One of the main problems in Artificial Intelligence
so far is that there is no such ground. This is what people use to talks as
the "Symbol Grounding Problem".
I agree with you that any intelligent system MUST ground its knowledge
in information collected from its "Umwelt".
I only would not suggest this requirement that the sensorial devices need to
be akin
to human sensors. You may develop different levels of intelligence,
depending
on how our "Umwelt" is the same as the intelligent system's "Umwelt". I am
talking about sharing experiences. If we are able to share experiences, then
we are able to communicate. The point is ... to start becoming intelligent,
it
has to sense and actuate on the world. Depending on the types of models it
is able to build from sensor data, different types of semiotic
interpretations it
will be able to develop and so the intelligence it is able to show us.
Ricardo
--
//\\\
(o o)
+-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-oOO--(_)--OOo-=-=-+
\ Prof. Ricardo Ribeiro Gudwin /
/ Intelligent Systems Development Group \
\ DCA - FEEC - UNICAMP | INTERNET /
/ Caixa Postal 6101 | gudwin@dca.fee.unicamp.br \
\ 13081-970 Campinas, SP | gudwin@fee.unicamp.br /
/ BRAZIL | gudwin@correionet.com.br \
+-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-+
\ URL: http://www.dca.fee.unicamp.br/~gudwin/ /
/ Telephones: +55 (19) 788-3819 DCA/Unicamp (University) \
\ +55 (19) 254-0184 Residencia (Home) /
/ FAX: +55 (19) 289-1395 \
+-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-+