Selected Newsgroup Message

Dejanews Thread

From: "Sergio Navega" <snavega@ibm.net>
Subject: Best Names in AI, CogSci
Date: 03 Dec 1998 00:00:00 GMT
Message-ID: <3666f816.0@news3.ibm.net>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 3 Dec 1998 20:44:06 GMT, 200.229.242.158
Organization: SilWis
Newsgroups: comp.ai.philosophy,comp.ai

End of year is approaching and the time to do that
recapitulation is arriving. I decided to list the
names of significant researchers and thinkers that
were influential on me. It does not mean I agree
with all of them, just that their work is important
in my vision.

May I invite you all to do the same? I would be
glad to know other opinions about other names.
Following is my list (I'm sure I forgot important
names, but that can be one excuse for you to
remind me).

Sergio Navega.

--------------------------------------------------

Michael Gazzaniga
One of the coiners of the term Cognitive Neuroscience,
Gazzaniga gave us a lot of important texts. One I
particularly appreciate is his recent textbook
"Cognitive Science, The Biology of Mind".

Daniel C. Dennett
Dennett is for me the best kind of philosopher: one that
cares to look at scientific explanations for any
armchair speculations. I have all his books but my
preferred is "Darwin's Dangerous Idea".

Aaron Sloman
Prof. Sloman belongs to that rare list of thinkers
with a significant amount of work, much still waiting
for recognition. Among his huge work spanning a good
area of AI, I highlight his theories of emotion in
robots.

Marvin Minsky
Prof. Minsky is, perhaps, the best known name in AI and
one of the most influential figures. He gave us splendid
ideas (like the frames knowledge representation) and several
captivating articles. The Society of Mind is probably his
best work so far, a constant source of inspiration.

Risto Miikkulainen
DISCERN is a connectionist system able to treat natural
language processing entirely at the subsymbolic level.
Prof. Miikkulainen's (no, it's not a typo) contributions,
besides a vast number of papers, span to his graduate
students, such as Joseph Sirosh's RF-LISSOM model of
the primary visual cortex. Connectionism alive and kicking.

Noam Chomsky
I may have some reservations with his innateness ideas but
no one can deny his extraordinary importance, perhaps the
legitimate father of Cognitive Science. The downfall of
behaviorism owes a lot to him.

Jerry Fodor
Important and polemic, reading Fodor's ideas are a must for
everybody barely involved with AI, CogSci or Philosophy of
Mind. His latest books are "Concepts, Where Cognitive
Science Went Wrong" and "In Critical Condition, Polemic
Essays on Cognitive Science and the Philosophy of Mind".

Robert Sternberg
Prolific author, Sternberg edited dozens of books about
intelligence and cognitive science. Other important works
include papers and book editions about creativity.

Douglas Hofstadter
If I had to name the single most influential author for me,
Doug would be it. I'm not talking only of GEB and Metamagical
Themas. I'm talking mainly of "Fluid Concepts and Creative
Analogies", in a way the center of everything I think is
relevant to intelligence.

Paul Thagard
Prof. Thagard wrote the single most readable and interesting
little introductory book about CogSci: "Mind - Introduction to
Cognitive Science". But I like him also because of his focus
on inductive reasoning, which is the main topic of a book he
wrote with Holland, Holyoak and Nisbett ("Induction, Processes
of Inference, Learning and Discovery").

Allen Newell
Newell is among the most influential figures of AI and CogSci and
together with Herbert Simon will be remembered because of the
important PSSH (Physical Symbol System Hypothesis). Too bad he
is not with us anymore.

John Anderson
Anderson's ACT-R model of cognition is among the most important
works on unified cognition, comparable only to Newell's Soar.

Stevan Harnad
I have two reasons to list Harnad here. He have produced one of
the gems of this decade (The Symbol Grounding Problem) and he
is conducting the academic and publishing industries to find a
way to publish electronically, on the Internet, favoring millions
of researchers around the world.

Alan Turing
Definitely a genius, with a sad personal history, the world would
not be the same without him.

Margaret Boden
Prof. Boden is listed here because she wrote a lot in the intersection
between AI and psychology. However, she is probably one of the most
important researchers dealing with AI and creativity.

David Chalmers
Together with Dan Dennett, another example of how philosophers should be.

William Calvin
With a lot of published books, a very influential author. I particularly
like "The Cerebral Code" and "Conversations with Neil's Brain". He have
probably the best and most complete personal web page so far.

Jeffrey Elman
One of the most cited authors when the subject is recurrent neural nets.
His book "Rethinking Innateness" must find a place in
the shelves of anybody with a slight interest in CogSci.

Annette Karmiloff-Smith
She is a co-author of "Rethinking Innateness" but she is in my list
because of her book "Beyond Modularity". As a disciple of Piaget, she
managed to disagree with her master in some aspects and produce
a valuable work.

George Lakoff
Prof. Lakoff is another of my preferred. "Metaphors we live by" is one
of his books that show how important metaphors are to our cognition.
But my preferred book is "Women, Fire and Dangerous Things", a must
in everybody's shelves.

Philip Johnson-Laird
A psychologist with a lot to say to AI researchers, I appreciate his
books "The Computer and the Mind" and "Mental Models".

Roger Schank
Roger Schank is, for me, second only to Hofstadter. His vision that
cognition is very linked with story understanding/representation is
among the most influential ideas I carry. But there's another gem
produced by Schank, his Conceptual Dependency knowledge representation,
something I think is a very important idea to flourish.

George Miller
Two reasons for Miller in this list: he was influential and author of
the seminal "The magic number seven, plus or minus two" but he also
started one of the most important projects in AI and natural language
processing: Princeton's WordNet.

I have some other names, but the list is good at this size. Now,
it's your turn!

From: rickert@cs.niu.edu (Neil Rickert)
Subject: Re: Best Names in AI, CogSci
Date: 03 Dec 1998 00:00:00 GMT
Message-ID: <7471n4$61@ux.cs.niu.edu>
References: <3666f816.0@news3.ibm.net>
Organization: Northern Illinois University
Newsgroups: comp.ai.philosophy,comp.ai

"Sergio Navega" <snavega@ibm.net> writes:

>End of year is approaching and the time to do that
>recapitulation is arriving. I decided to list the
>names of significant researchers and thinkers that
>were influential on me. It does not mean I agree
>with all of them, just that their work is important
>in my vision.

>May I invite you all to do the same?

Hmm.

My list might be close to empty.  I obviously have read many of these
researchers.  But I have taken a fresh approach, so that I can't say
that anyone stands out as particularly influential.

However, I'll comment on your list.

>Michael Gazzaniga
>One of the coiners of the term Cognitive Neuroscience,
>Gazzaniga gave us a lot of important texts. One I
>particularly appreciate is his recent textbook
>"Cognitive Science, The Biology of Mind".

I haven't read that book.  I did read "Nature's Mind: The Biological
Roots of Thinking, Emotions, Sexuality, Language, and Intelligence,"
but I thought he reached some seriously wrong conclusions (about what
is innate).

>Daniel C. Dennett
>Dennett is for me the best kind of philosopher: one that
>cares to look at scientific explanations for any
>armchair speculations. I have all his books but my
>preferred is "Darwin's Dangerous Idea".

I like a lot of what Dennett writes.  But DDI is not my favorite.  I
thought he was altogether too negative about Gould (in DDI), and too
committed to the Dawkins view of evolution.  In fact, DDI seemed more
of a polemic that I have come to expect from Dennett.

>Aaron Sloman
>Prof. Sloman belongs to that rare list of thinkers
>with a significant amount of work, much still waiting
>for recognition. Among his huge work spanning a good
>area of AI, I highlight his theories of emotion in
>robots.

I have considerable respect for Sloman.

>Marvin Minsky
>Prof. Minsky is, perhaps, the best known name in AI and
>one of the most influential figures. He gave us splendid
>ideas (like the frames knowledge representation) and several
>captivating articles. The Society of Mind is probably his
>best work so far, a constant source of inspiration.

I have a lot of respect for Minsky, although I don't agree with all
of his ideas.  I agree that SoM is a good work, although I think it
can be criticized as not giving enough in the way of direction to a
researcher.

>Risto Miikkulainen

Sorry, I don't know his work.

>Noam Chomsky

I'll strongly disagree with you on this one.  His basic approach to
language is obviously mistaken, although I don't expect many people
to agree with me in this view.  Chomsky may be responsible for
sending more researchers on a wild goose chase than anyone else in
history.

>Jerry Fodor
>Important and polemic, reading Fodor's ideas are a must for
>everybody barely involved with AI, CogSci or Philosophy of
>Mind. His latest books are "Concepts, Where Cognitive
>Science Went Wrong" and "In Critical Condition, Polemic
>Essays on Cognitive Science and the Philosophy of Mind".

I disagree with most of Fodor's ideas.  On the other hand he is a
clear writer, and he doesn't pull any punches.  I do agree that he is
well worth reading.

>Robert Sternberg

I don't recall having read any of his work.

>Douglas Hofstadter
>If I had to name the single most influential author for me,
>Doug would be it. I'm not talking only of GEB and Metamagical
>Themas. I'm talking mainly of "Fluid Concepts and Creative
>Analogies", in a way the center of everything I think is
>relevant to intelligence.

Sorry to disagree.  I think of GEB and Metamagical Themas as two of
the most boring books ever written.

>Paul Thagard

Hmm.  According to my records, I've read a couple of his books.  I
must not have found them memorable, since I don't remember anything
about them other than the author's name.

>Allen Newell
>John Anderson

I think both of them are on the wrong track.

>Stevan Harnad
>I have two reasons to list Harnad here. He have produced one of
>the gems of this decade (The Symbol Grounding Problem) and he
>is conducting the academic and publishing industries to find a
>way to publish electronically, on the Internet, favoring millions
>of researchers around the world.

I agree with you on electronic publishing.  I'm not so sure of the
importance "The Symbol Grounding Problem."  It is often connected
with Searle's CR argument.  But I'm inclined to think that the symbol
grounding problem is being looked at backwards.  We don't ground
symbols.  Rather, we symbolize the ground.  Indeed, that is what is
wrong with Newell, Anderson, Schank and others.  They start with
symbols, and run into the problem of grounding.  We have to start
with reality, and the way we symbolize it.

Part of Harnad's research is related to perception, which does have
something to do with symbolizing the ground.

>Alan Turing
>Definitely a genius, with a sad personal history, the world would
>not be the same without him.

Agreed.

>Margaret Boden
>Prof. Boden is listed here because she wrote a lot in the intersection
>between AI and psychology. However, she is probably one of the most
>important researchers dealing with AI and creativity.

She writes clearly, and raises many good questions.  I agree that she
is well worth reading.  Incidently, she also wrote a nice little book
on Piaget.

>David Chalmers
>Together with Dan Dennett, another example of how philosophers should be.

I think he gets too carried away with questions of qualia.  But apart
from that, I would agree.

>William Calvin
>With a lot of published books, a very influential author. I particularly
>like "The Cerebral Code" and "Conversations with Neil's Brain". He have
>probably the best and most complete personal web page so far.

I haven't given more than a quick once over to his books.  His work is
a little outside my areas of interest.

>Jeffrey Elman
>One of the most cited authors when the subject is recurrent neural nets.
>His book "Rethinking Innateness" must find a place in
>the shelves of anybody with a slight interest in CogSci.

I liked "Rethinking Innateness".

>Annette Karmiloff-Smith
>She is a co-author of "Rethinking Innateness" but she is in my list
>because of her book "Beyond Modularity". As a disciple of Piaget, she
>managed to disagree with her master in some aspects and produce
>a valuable work.

Agreed.

>George Lakoff
>Prof. Lakoff is another of my preferred. "Metaphors we live by" is one
>of his books that show how important metaphors are to our cognition.
>But my preferred book is "Women, Fire and Dangerous Things", a must
>in everybody's shelves.

I agree that WFDT is valuable.  But I think he gets too carried away
in his work on metaphors.

>Philip Johnson-Laird
>A psychologist with a lot to say to AI researchers, I appreciate his
>books "The Computer and the Mind" and "Mental Models".

I too appreciate those books.

>Roger Schank

I mostly disagree with his work on AI.

>George Miller
>Two reasons for Miller in this list: he was influential and author of
>the seminal "The magic number seven, plus or minus two" but he also
>started one of the most important projects in AI and natural language
>processing: Princeton's WordNet.

I support your first reason.

You might want to look at Gerald Edelman's "Neural Darwinism: The
Theory of Neural Group Selection".  Some of Putnam's writings are
worth reading.  Perhaps "Representation and Reality" is a good place
to start, and "Reason, Truth and History" also says some interesting
things.

From: "Sergio Navega" <snavega@ibm.net>
Subject: Re: Best Names in AI, CogSci
Date: 04 Dec 1998 00:00:00 GMT
Message-ID: <3667d09c.0@news3.ibm.net>
References: <3666f816.0@news3.ibm.net> <7471n4$61@ux.cs.niu.edu> <36672DC4.DAA9A27A@clickshop.com>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 4 Dec 1998 12:07:56 GMT, 129.37.183.228
Organization: SilWis
Newsgroups: comp.ai.philosophy,comp.ai

Seth Russell wrote in message <36672DC4.DAA9A27A@clickshop.com>...
>Neil Rickert wrote:
>
>> I'm inclined to think that the symbol
>> grounding problem is being looked at backwards.  We don't ground
>> symbols.  Rather, we symbolize the ground.  Indeed, that is what is
>> wrong with Newell, Anderson, Schank and others.  They start with
>> symbols, and run into the problem of grounding.  We have to start
>> with reality, and the way we symbolize it.
>
>You may be right if your are talking about building a stand alone robot
that
>is designed to duplicate human natural intelligence.  But I think you are
dead
>wrong where it comes to practical robust artificial intelligence.  Such
>intelligence sits in a different position visa vie symbols and their
grounding
>... or rather the human mind sits in a different position visa vie its
symbol
>ground than will robust AI sit visa vie its symbol ground.  They are in
>different basic predicaments - don't forget us and our culture and the eons
of
>evolution of our human minds and how *that* must sit in relationship to
Robust
>AI.
>

Seth, Neil is, unfortunately, right. I say unfortunately because most (if
not all) work on symbolic AI so far is suffering from the same fatal flaw:
they are not able to work by their own, only in conjunction with a human
(which stands as the intelligent part of the circuit!). This is the case,
for example, of CYC.

CYC is, IMHO, an "idea freezer". Lots of intelligent humans stand in front
of the machine and put his best ideas on the freezer. It is indeed a very
nice way to conserve ideas for a long time! But books do the same thing
for much, much less. I agree that it is a special kind of freezer, one
able to combine the foods you put inside to make a new recipe.
But it is not able to come up with the raw material by itself,
and this happens because it is not able to recognize *what is significant*
in the world to be incorporated as valuable resources. It is that
recognition
of significance that is the core of intelligence and this is missing in CYC
and all symbolic systems I know today.

Regards,
Sergio Navega.

From: =?iso-8859-1?Q?Mat=EDas?= Giovannini <matias@k-bell.com>
Subject: Re: Best Names in AI, CogSci
Date: 04 Dec 1998 00:00:00 GMT
Message-ID: <36688442.4E5DF099@k-bell.com>
X-NNTP-Posting-Host: 208.192.4.30
Content-Transfer-Encoding: 7bit
References: <3666f816.0@news3.ibm.net> <7471n4$61@ux.cs.niu.edu> <36672DC4.DAA9A27A@clickshop.com> <3667d09c.0@news3.ibm.net>
X-Accept-Language: en
Content-Type: text/plain; charset=us-ascii
X-Trace: 4 Dec 1998 19:54:39 -0500, 209.14.117.161
Organization: Script S.A.
MIME-Version: 1.0
Reply-To: matias@k-bell.com
Newsgroups: comp.ai.philosophy,comp.ai

Sergio Navega wrote:

(deletia)

> CYC is, IMHO, an "idea freezer". Lots of intelligent humans stand in front
> of the machine and put his best ideas on the freezer. It is indeed a very
> nice way to conserve ideas for a long time! But books do the same thing
> for much, much less. I agree that it is a special kind of freezer, one
> able to combine the foods you put inside to make a new recipe.
> But it is not able to come up with the raw material by itself,
> and this happens because it is not able to recognize *what is significant*
> in the world to be incorporated as valuable resources. It is that
> recognition
> of significance that is the core of intelligence and this is missing in CYC
> and all symbolic systems I know today.

That's exactly what I feel is at the same time appealing and appalling
about CYC: a lot of knowledge coded and stored in a handy format. Handy,
that is, for computer processing (I was about to write: "automated
processing", but the question begging to be asked is: could AI ever be
called "automated processing"? I think it is a distinct possibility :-):
if only we had an AI, the wonders it could do with such knowledge!

On the other hand, it is not immediate that the common-sense type of
knowledge CYC tries to capture necessarily requires "strong" or
"fundamental" AI capabilities; it may very well be the case that
sophisticated applications could arise from its application to some
"limited" form of AI (for a somewhat inane example, a "Did you knonw
that..."-like tip generator applying "random" inference chains to the database).

Just my 2 neurons' worth.

From: "Sergio Navega" <snavega@ibm.net>
Subject: Re: Best Names in AI, CogSci
Date: 05 Dec 1998 00:00:00 GMT
Message-ID: <36692f9b.0@news3.ibm.net>
References: <3666f816.0@news3.ibm.net> <7471n4$61@ux.cs.niu.edu> <36672DC4.DAA9A27A@clickshop.com> <3667d09c.0@news3.ibm.net> <36688442.4E5DF099@k-bell.com>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 5 Dec 1998 13:05:31 GMT, 129.37.182.131
Organization: SilWis
Newsgroups: comp.ai.philosophy,comp.ai

Matías Giovannini wrote in message <36688442.4E5DF099@k-bell.com>...
>
>Sergio Navega wrote:
>
>(deletia)
>
>> CYC is, IMHO, an "idea freezer". Lots of intelligent humans stand in
front
>> of the machine and put his best ideas on the freezer. It is indeed a very
>> nice way to conserve ideas for a long time! But books do the same thing
>> for much, much less. I agree that it is a special kind of freezer, one
>> able to combine the foods you put inside to make a new recipe.
>> But it is not able to come up with the raw material by itself,
>> and this happens because it is not able to recognize *what is
significant*
>> in the world to be incorporated as valuable resources. It is that
>> recognition
>> of significance that is the core of intelligence and this is missing in
CYC
>> and all symbolic systems I know today.
>
>That's exactly what I feel is at the same time appealing and appalling
>about CYC: a lot of knowledge coded and stored in a handy format. Handy,
>that is, for computer processing (I was about to write: "automated
>processing", but the question begging to be asked is: could AI ever be
>called "automated processing"? I think it is a distinct possibility :-):
>if only we had an AI, the wonders it could do with such knowledge!

If I was obliged to use the word "automated" to describe AI, I would say
"automated learning". This is not what CYC looks like to me.

>
>On the other hand, it is not immediate that the common-sense type of
>knowledge CYC tries to capture necessarily requires "strong" or
>"fundamental" AI capabilities;

To store that knowledge, indeed, one does not have to possess
intelligence. But that is just the problem with the whole project.
We are not after something that can store knowledge, we're after
something that can *produce* knowledge when fed with raw data.
*That's* what makes one entity intelligent: to grasp what is important
about one's world and to use it to solve problems (like improving
its life condition or serving its human masters).

>it may very well be the case that
>sophisticated applications could arise from its application to some
>"limited" form of AI (for a somewhat inane example, a "Did you knonw
>that..."-like tip generator applying "random" inference chains to the
database).
>

The state of current computers is so ridiculously underdeveloped that
I believe CYC will have a good impact in several innovative applications.
But that's a drop of water in the ocean, compared to what a truly
intelligent machine could do.

Regards,
Sergio Navega.

From: "Sergio Navega" <snavega@ibm.net>
Subject: Re: Best Names in AI, CogSci
Date: 04 Dec 1998 00:00:00 GMT
Message-ID: <3667d09a.0@news3.ibm.net>
References: <3666f816.0@news3.ibm.net> <7471n4$61@ux.cs.niu.edu>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 4 Dec 1998 12:07:54 GMT, 129.37.183.228
Organization: SilWis
Newsgroups: comp.ai.philosophy,comp.ai

Neil Rickert wrote in message <7471n4$61@ux.cs.niu.edu>...
>"Sergio Navega" <snavega@ibm.net> writes:
>
>>Michael Gazzaniga
>>One of the coiners of the term Cognitive Neuroscience,
>>Gazzaniga gave us a lot of important texts. One I
>>particularly appreciate is his recent textbook
>>"Cognitive Science, The Biology of Mind".
>
>I haven't read that book.  I did read "Nature's Mind: The Biological
>Roots of Thinking, Emotions, Sexuality, Language, and Intelligence,"
>but I thought he reached some seriously wrong conclusions (about what
>is innate).
>

I don't have "Nature's Mind", perhaps I can leave it to another time.
The textbook I mentioned does not suffers from this problem, it is a
very complete (and very well illustrated, with lots of 4 color photos
and diagrams really meaningful, not just fancy) of all aspects of
evidences in cognitive neuroscience. I think it is the best way to
know about this subject. It stands among the best books I have
on the subject.

>>Daniel C. Dennett
>>Dennett is for me the best kind of philosopher: one that
>>cares to look at scientific explanations for any
>>armchair speculations. I have all his books but my
>>preferred is "Darwin's Dangerous Idea".
>
>I like a lot of what Dennett writes.  But DDI is not my favorite.  I
>thought he was altogether too negative about Gould (in DDI), and too
>committed to the Dawkins view of evolution.  In fact, DDI seemed more
>of a polemic that I have come to expect from Dennett.
>

Indeed that's what the book seems: a long attack to all
those who think differently. I would appreciate if it came with an
appendix with the replies of the other authors. But then, the
volume would weight more than an arm could handle. I would
buy such a book only if Dennett had a chance to reply to
the comments.

>>Douglas Hofstadter
>>If I had to name the single most influential author for me,
>>Doug would be it. I'm not talking only of GEB and Metamagical
>>Themas. I'm talking mainly of "Fluid Concepts and Creative
>>Analogies", in a way the center of everything I think is
>>relevant to intelligence.
>
>Sorry to disagree.  I think of GEB and Metamagical Themas as two of
>the most boring books ever written.
>

And so is "Fluid Concepts". The initial chapter "To Seek Whence
Cometh a Sequence" is awfully painful to read, an exercise in
determination for the reader, although constructed with a
somewhat captivating writting style. But the ideas contained (and
in particular the principles of Copycat) are worthwhile.
The book spans for 500 pages, but could be rewritten in 250
without much loss. Those who find Hofstadter books readable are,
generally, those who appreciate literature in general, which
is *not* my case and does not seem to be yours.

>>Allen Newell
>>John Anderson
>
>I think both of them are on the wrong track.
>

So do I, but I keep their ideas fresh because, one way or another,
they raised valid points that must be taken care by any model that
have the same goals.

>
>>Margaret Boden
>>Prof. Boden is listed here because she wrote a lot in the intersection
>>between AI and psychology. However, she is probably one of the most
>>important researchers dealing with AI and creativity.
>
>She writes clearly, and raises many good questions.  I agree that she
>is well worth reading.  Incidently, she also wrote a nice little book
>on Piaget.
>

Didn't know about that, will look for it.

>>William Calvin
>>With a lot of published books, a very influential author. I particularly
>>like "The Cerebral Code" and "Conversations with Neil's Brain". He have
>>probably the best and most complete personal web page so far.
>
>I haven't given more than a quick once over to his books.  His work is
>a little outside my areas of interest.
>

Most of his books are for general readers and even "The Cerebral Code",
which he mentions having written to fellow scientists, is not representative
of the tipical "academic" work. I like Calvin but I would prefer to see him
digging a little bit deeper.

>>Jeffrey Elman
>>One of the most cited authors when the subject is recurrent neural nets.
>>His book "Rethinking Innateness" must find a place in
>>the shelves of anybody with a slight interest in CogSci.
>
>I liked "Rethinking Innateness".
>

It's one of my bibles, although I don't share his enthusiasm for
connectionism.

>You might want to look at Gerald Edelman's "Neural Darwinism: The
>Theory of Neural Group Selection".  Some of Putnam's writings are
>worth reading.  Perhaps "Representation and Reality" is a good place
>to start, and "Reason, Truth and History" also says some interesting
>things.
>

Thanks for the names. I owe Putnam a chance, I haven't read nothing of
him. About Edelman, I have on my reading list "Bright Air, Brilliant
Fire", there's a chapter on Neural Darwinism. I guess I can't postpone
reading this anymore.

Regards,
Sergio Navega.

From: "Sergio Navega" <snavega@ibm.net>
Subject: Re: Best Names in AI, CogSci
Date: 04 Dec 1998 00:00:00 GMT
Message-ID: <3667d096.0@news3.ibm.net>
References: <3666f816.0@news3.ibm.net> <3666fe56.2543438@news-server>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 4 Dec 1998 12:07:50 GMT, 129.37.183.228
Organization: SilWis
Newsgroups: comp.ai.philosophy,comp.ai

Jerry Hull wrote in message <3666fe56.2543438@news-server>...
>On Thu, 3 Dec 1998 17:38:31 -0200, "Sergio Navega" <snavega@ibm.net> wrote:
>
>>Daniel C. Dennett
>>Dennett is for me the best kind of philosopher: one that
>>cares to look at scientific explanations for any
>>armchair speculations. I have all his books but my
>>preferred is "Darwin's Dangerous Idea".
>
>"Dennet's Disingenous Idea".  Maybe if I ascribe my controversial opinions
to
>a revered historical figure like Darwin, I can win their greater
acceptance.
>Perhaps no one will note that Darwin himself never talked about algorithms.
>

Neither he knew about genetic mechanisms. I bet if Darwin met Mendel and
Turing he would find the perfect complements to his ideas.

Regards,
Sergio Navega.

From: "Sergio Navega" <snavega@ibm.net>
Subject: Re: Best Names in AI, CogSci
Date: 04 Dec 1998 00:00:00 GMT
Message-ID: <36681181.0@news3.ibm.net>
References: <3666f816.0@news3.ibm.net> <3667690a.0@mithril.niia.net>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 4 Dec 1998 16:44:49 GMT, 166.72.21.99
Organization: SilWis
Newsgroups: comp.ai.philosophy,comp.ai

Kyle Rhodes wrote in message <3667690a.0@mithril.niia.net>...
>
>Rodney Brooks?
>
>

Yes! Thanks for reminding us. I found his paper
"Intelligence without representation" to be among the
classics. Of course, his ideas puts him as antagonist
of both connectionists and symbolicists simultaneously.
Maybe that's his main merit, to look at the problem from
a totally different perspective.

Regards,
Sergio Navega.

From: "Sergio Navega" <snavega@ibm.net>
Subject: Re: Best Names in AI, CogSci
Date: 05 Dec 1998 00:00:00 GMT
Message-ID: <36692f47.0@news3.ibm.net>
References: <3666f816.0@news3.ibm.net> <749knr$kcj$1@paul.rutgers.edu> <74aa0k$r5d$1@amenti.rutgers.edu>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 5 Dec 1998 13:04:07 GMT, 129.37.182.138
Organization: SilWis
Newsgroups: comp.ai.philosophy,comp.ai

Luke Kaven wrote in message <74aa0k$r5d$1@amenti.rutgers.edu>...
>
>How about a list of philosophers whose work has the most important
>implications for Cognitive Science and AI, but whose work is largely
>too difficult to understand without a good background in philosophy,
>and thus is largely still unread in Cognitive Science and AI:
>
>Saul Kripke
>_Naming and Necessity_, _Wittgenstein on Rules and Private Language_
>
>Sydney Shoemaker
>_Identity, Cause, and Mind_, _The First Person Perspective_
>
>Ruth Garrett Millikan
>_Language, Thought, and Other Biological Categories_, _White
>Queen Psychology_
>
>Carl Ginet
>_On Action_
>
>Fred Dretske
>_Knowledge and the Flow of Information_, _Explaining Behavior_,
>_Naturalizing the Mind_
>
>Peter Godfrey Smith
>_Complexity and the Function of Mind in Nature__
>
>Ludwig Wittgenstein
>_Tractatus Logico-Philosophicus_, _Philosophical Investigations_
>
>Jaegwon Kim
>_Supervenience and Mind_
>
>William Lycan
>_Consciousness_, _Judgement and Justification_
>
>also, John Perry, Bernard Williams, P.F. Strawson, D.H Mellor,
>Hilary Putnam, Richard Boyd, Lynn Rudder Baker, John Heil,
>John McDowell, Michael Devitt, Kim Sterelny, David Armstrong, etc., etc.

Nice list! Gave me lots of names to play with! With the exception
of Wittgenstein, Putnam and Kripke, all others are unknown to me.

Regards,
Sergio Navega.

From: "Sergio Navega" <snavega@ibm.net>
Subject: Re: Best Names in AI, CogSci
Date: 05 Dec 1998 00:00:00 GMT
Message-ID: <36692f44.0@news3.ibm.net>
References: <3666f816.0@news3.ibm.net> <749knr$kcj$1@paul.rutgers.edu>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 5 Dec 1998 13:04:04 GMT, 129.37.182.138
Organization: SilWis
Newsgroups: comp.ai.philosophy,comp.ai

Omar Haneef wrote in message <749knr$kcj$1@paul.rutgers.edu>...
>Sergio:
>
>    An excellent list and an excellent idea. I appreciate both, so
>thanks. I agree with many, but not all (thank goodness), of your
>suggestions. Here is another look at the cog. sciers who influence
>me. You will notice the heavy Rutgers influence because a lot of this
>I've read in the course of courses:
>

>Philosophers:
>    Fodor, Stich, Dennet, Goldman (and, dare I say, Foucault?)
>    Almost the most popular figures in philosophy, and I hope its not
>taken as a dirty word.
>

Thanks for the names, Stephen Stich and Alvin Goldman are two I don't
know much of. I'll try to keep one eye on them.

>AIers:
>    Minsky, McCarthy, Simon, Newell, Mclelland, Quinlan
>    Justa who's who of symbolic AI, I think, with neural nets and
>Machine Learning thrown in. Obviously classic Comp Sci influences
>everthing too with Church, Turing, Knuth etc.
>

I should have mentioned John McCarthy in my list! James McClelland is
also one I should have put in.

>Psychologist:
>    Rumelhart, Tversky, Kahneman, Shepard
>    Shepard's work on categorization has been the seed for a lot of
>what I do now. Tversky and Kahneman, however controversial their work
>is, is my idea of what psychology should be. In my mind it represents
>the best of what psychology has to offer. If it weren't for PDP, I
>would be in some other field.
>

Shepard is a new name for me and in a subject I'm very fond. On the
same lineage of psychologists, perhaps I should add Smolensky.
Tversky & Kahneman are decidedly influential.

>Linguists:
>    Chomsky (goose chase or not, he did INFLUENCE us), Prince (There
>is so much optimality theory around here that I can't help be more
>influenced by Prince than by, say, Pinker)
>

I'm just beginning in Optimality Theory, much to learn, thanks for
the name of Prince.

>Neuroscientists:
>    I researched this stuff for a year, and frankly don't understand
>how the tools are alligned with the project at hand. Maybe if you
>consider Marr a neuroscientist, I would call him in.
>

On the list of neuroscientists perhaps I should have added Kosslyn.

>For important works:
>    Modularity of Mind
>    Parrallel Distributed Processing
>    Others...
>
>    -Omar

This is a good idea, a list of the most influential works.
I'll try to tackle this next week, if there's some spare time left.

Regards,
Sergio Navega.

From: "Sergio Navega" <snavega@ibm.net>
Subject: Re: Best Names in AI, CogSci
Date: 14 Dec 1998 00:00:00 GMT
Message-ID: <36753f60.0@news3.ibm.net>
References: <3666f816.0@news3.ibm.net> <19981214022859.06064.00001184@ng53.aol.com>
X-Notice: should be reported to postmaster@ibm.net
X-MimeOLE: Produced By Microsoft MimeOLE V4.71.1712.3
X-Complaints-To: postmaster@ibm.net
X-Trace: 14 Dec 1998 16:40:00 GMT, 166.72.21.90
Organization: SilWis
Newsgroups: comp.ai

MRFerrier wrote in message <19981214022859.06064.00001184@ng53.aol.com>...
>I'd like to add Stephen Grossberg, from Boston University, to this list of
>people influential to AI.  He's doing amazing work with neural networks, a
>technology that he stuck with even while it was least popular.
>

Good name! He really belongs to that list.

>>Risto Miikkulainen
>
>I've never heard of him (her?), but the name sounds Finish.  Do you know if
>he's at Helsinki University?
>

No, he is at University of Texas, Austin. If you like Grossberg for his work
with NN, then take a look at Risto's homepage. Truly amazing work.

http://www.cs.utexas.edu/users/risto/

His publications (along with other members of his group) can be found
online at:

http://www.cs.utexas.edu/users/nn/pages/publications/abstracts.html

Impressive work.

Regards,
Sergio Navega.

Back to Menu of Messages               Sergio's Homepage

Any Comments? snavega@attglobal.net