I am posting this article to provide an insight into the use of machine
intelligence methodologies (neural networks, and fuzzy logic) in the medical
field.

Neural networks are an information processing algorithm that is based on the
way the brain processes information.  It is made up of a large number of
interconnected processing elements (neurons) working together.  It learns
through adjustments in the synaptic connections that exist between the neurons
just as in biological systems.  Neural network's most improtant advantage is
in solving problems which are too complex for conventional technologies.  They
have proved suitable for problems such as pattern recognition, prediction, and
modeling.

Fuzzy logic is a paradigm which uses knowledge in a manner very close to the
way experts think about complex problems.  Fuzzy logic provides the power to
break through the complexity of traditional decision support and expert
systems.  In addition, fuzzy logic builds systems which use information from
conflicting, cooperating, and collaborating experts; a capability which is
generally beyond the scope of genetic algorithms, neural networks, or
traditional expert systems.

What do the capabilities illustrated above bring to the medical field?  The
machine intelligence approach will integrate itself into almost all areas of
medical research, diagnostics, classification, analysis, and modeling.  They
are already being successfully applied to various areas of medicine such as
diagnostic aides, chemical analysis, image analysis, classification of
x-rays, cancer, and AIDS drug development, e-ray spectral reconstruction,
tumor detection in ultrasonograms, and determination of skeletal age, to name
just a few.  

In the diagnostic area, systems are on stream for the early diagnosis of
myocardial infarction, heart murmur, epilepsy, coronary artery diesase, and
diabetes.

Classification applications include microcalcifications in mamograms, tissue,
and vessels.

Analysis applications include urinalysis, and {*filter*}.

Although most of the above applications utilize artificial neural networks, it
is believed that fuzzy logic will quickly become the pre{*filter*} artificial
intelligence technology.  Fuzzy logic is the only technique which has the
capability to actually code the most ambiguous of systems, including human
emotions, thoughts, and feelings.  This is because fuzzy logic is based on a
calculus of infinite states.  While conventional Boolean logic (true/false)
has only 16 truth states, and three valued logic (yes/no/maybe) has 19,683
truth states, fuzzy logic has an infinite number of truth states.

This ability to handle ambiguity, plus the extremely small memory requirements
of a fuzzy logic coded system, will in all certainty be leveraged into systems
which here-to-fore have only been dreamed of.  For example, work is ongoing in
the area of nanotechnology (machines the size of molecules) which could yield
a myriad of wonders, such as infinitesimal devices the automatically convert
energy into food, or roam the {*filter*}stream in search of desease; repairing the
body on a molecular level.  This could result in the dream of the ages;
keeping those molecules healthy, and slowing down the aging process to nil.
The practicality of this endeavor however is based on having a logic engine
chip (tells nano-robot what to do) which is small enough to fit on the
molecular sized unit.  Fuzzy logic is the only system at present which offers
this capability.

The above is only a sampling of the capabilities of machine intelligence in
medicine.  If you would like additional information, or specifics on whether
this technology would be of merit in your research efforts, please contact
me by phone, e-mail, or surface mail:

Richard M. Olson
President
Olson Logic
8720 Jeffrey Ave. No.
Stillwater, MN 55082
(612) 429-0191