Artificial
Intelligence
The Greeks called this idea Theos Ek Mechanes, and in Latin, it was
Deus Ex Mechina. The concept of creating an artificial automated system
is almost as old as the written word. In the past sixty years, it
is closer to reality. Although the field of Artificial Intelligence
has made great strides, it still lacks a crucial concept - a solid
theory of intelligence that includes a definition of intelligence,
tangible goals, and a method for their measure. While not required
to construct an intelligent application, a theory of intelligence
is important for the AI field to progress. As an analogy, the Wright
brothers did not need a theory of lift to create an airplane, but
it allowed them to succeed by design rather than by chance. The following
theory enables the reader to view intelligent system problems from
a different prospective, and more importantly, consider intelligence
as a series of ever changing relationships, rather than just the analyzing
of facts and data.
The Goal
AI researchers indicate the ultimate goal of AI research is
creating a fully automated artificial human being. While sounding
fantastic, what does this statement really tell us? Let us look at
a similar hypothetical situation. What if we could go back a hundred
years and ask the world's top scientists of that time to design and
build a 747 jumbo jet, without supplying any additional information.
The scientists could not build it due to a lack of tangible goals.
The scientist might ask questions such as: what is a 747? What does
a 747 do? All of these questions attempt to make the goal more tangible.
This same problem plagues the field of AI. Stating the goal of AI
as the search for an artificial human means nothing without first
understanding what a human being is. As with the scientists and |
|
the jet, they would not
know if they built an airplane, if they did not first know what one
was. We must understand the human before we can create the artificial
human.
One approach to understanding the human, which
can
be simply stated but not so simply answered, is to ask, "Why
do we exist?" Asking this question of thirty people would likely
generate thirty different answers, ranging from serving God, to propagating
the species, to creating art, to paying taxes. While these responses
are partially correct, they are too specific to fully answer the question.
These responses, which resemble tasks rather than reasons for existence,
do have a common thread; they are all solutions to a problem. Paying
taxes solves the problem of government funding. The service to God
answers the need for a belief in a higher being. Creating art eliminates
boredom and satisfies the need for self-expression. Propagating the
species is the answer to human survival.
Life's daily activities are all a form of problem solving. Why we
wake up in the morning, "live" during the day, and eventually
return to sleep at night are all part of the solution to the current
problem. Everything we create or do helps us on our journey to solve
the problem of life. Every building, tool, machine, theory and device
is specifically created to make the journey through life easier and
more efficient. This of course leads to the question, "what is
a problem?"
What is a Problem?
Problems are commonly defined as a discrepancy
between the current and an optimal state or situation. The discrepancy
is what causes the problem, and is not actually the problem. The problem
solutions are the result of some action and most actions are in the
pursuit of a solution to a problem. If the reason we exist is to solve
problems, and those solutions |
