RE: Suggestion

Paul S Prueitt (Paul@htech.com)
Fri, 6 Mar 1998 18:59:47 +0100 (MET)

Lotfi's suggestion, regarding the use of terms describing certain
professional associations, is important. I feel that the importance has
a social dimension as well as a scientific dimension.

The social dimension has to do with harm that accrues to the scientific
community, in general, when terms such as "Intelligent Systems" are used
to indicate a hoped for, but not actual, label to research activity and
applications. In the case of "Intelligent Systems" the hoped for label
has come to have very little behavioral correspondence to the class of
natural living intelligent systems. Moreover, the incorrect use of the
label effectively blocks its correct use by competing view points. The
yearly Information Systems and Semiotics conferences at NIST is one
forum where this issue should be addressed more fully.

One may make an observation that a class of "scientific" terms are used
in a marketing fashion; to pull together coalitions designed to find
funding, publish papers and form social groups. Marketing is a social
behavior that produces strong localized self and professional interests.
These localizations of self interest are concomit with mechanisms that
absorb individual's in narrow belief systems and single inflexible and
entrenched scientific paradigms. In the Information Systems community,
mismatches between the understanding of theoretical foundations and
program management use of buzz words has caused chaos and terrible
squandering of national resources. These resources are used to build
useless and very costly "information systems" when the basic research
and architectures for a new generation of systems is pushed to one side.
As a consequence, the government and the public have not had the benefit
of the technology that could have been funded and built, if these buzz
words where not so vacuously used.

If the marketing of science is not fully honest, there are consequences
such as a decreased political capital with the American public, and a
sense of disbelief from policy makers regarding information
technologies.

A similar problem exists with respect to the reinforcement of an
interpretation of the term "neural network" that occurred when DARPA
defined this to be "nodes and connections between nodes" and pumped a
lot of money into that small community that would go along with this
definition. The selective funding occurred at a point in history when
members of the neuroscience and mathematics community might have come to
a deeper common understanding of what is right and what is wrong about
the so called "neuronal doctrine." Had this deeper understanding
occurred, then we might have avoided the publication of tens of
thousands of research papers, on artificial neural networks, that have
very little scientific value or lasting influence.

There are many similar examples. A mismatch between the common use of
the term "information" and professional use of the term in the
"Information Retrieval" (IR) community has lead to an acknowledged
failure of the TIPSTER Text Retrieval and Extraction Conference (TREC)
community to find support of a renewal of funding. Moreover, this
failure is at the core of reasons why no "intelligent" machine
"computational argumentation" system exist today for government use
(including intelligence use.) Thus, individuals responsible for
reaction to treats to our national security do not have the tools that
they would otherwise have, if the science community was not creating
shallow buzz words and the information industry was not marketing these
buzz words without a clue regarding theoretical foundations.

We should all look closely at the mismatch between what our "scientific
language" promises and what it actually delivers. We must not only
restructure the foundations of professional paradigms but also force the
terms to have some common sense and "transparency."

This is difficult for many reasons, but there should more serious
discussion about the issue.

Dr. Paul S. Prueitt

> -----Original Message-----
> From: tosh munakata [SMTP:munakata@grail.cba.csuohio.edu]
> Sent: Wednesday, March 04, 1998 4:45 PM
> To: bisc-group@diva.EECS.Berkeley.EDU
> Cc: munakata@grail.cba.csuohio.edu
> Subject: Suggestion
>
>
> 3/4/1998
>
> I do not support Lotfi's recent suggestion that the labels
> "Intelligent Systems," "Information Systems" and "Intelligent
> Information Systems" be replaced by Information/Intelligent
> Systems.
>
> The commonly accepted concept of the term "Information Systems" may
> be that it is one of the major computer related disciplines based
> on their primary interests and application domains. There are at
> least three such major disciplines: "Computer Engineering,"
> "Computer Science," and "Information Systems." As ACM/IEEE
> prototype curricula suggest, "Information Systems" refers to a less
> technical, more application oriented, often business flavored
> discipline.
>
> On the other hand, a common perception of "Intelligent Systems"
> would be that it is one of the major computer based techniques.
> Other techniques include databases, software engineering, and so
> on. Taking the union of the two terms, "Intelligent Systems" and
> "Information Systems," does not make much sense to me. I think
> they should be left as they are.
>
> Incidentally, as I wrote in the Guest Editor's introductions in the
> CACM AI Special Issues (March 1994 and November 1995), I think the
> term "intelligent" is misleading. The current state of
> "Intelligent Systems" or "Artificial Intelligence" is not a
> realization of human intelligence. The current state of the field
> is still best characterized as advanced computing. However, once
> a term is widely accepted, it is hard to change. We recall that
> the movement (during the 1980s?) to change the name of the
> Association of Computing Machinery, since "Machinery" is no longer
> dominant, was voted down.
>
> Needless to say, this letter is a matter of opinion and I do not
> intend to continue arguments and counter-arguments of this matter,
> nor is this a personal attack.
>
> Tosh Munakata
>
>