While general game playing is a topic
within here and interest, work is in this
area has practical value as well. The
underlying technology can be used in a
variety of other application areas, such
as enterprise management and computational
law. Science, engineering, and so forth.
In the case of enterprise management,
computational law, just consider that the
rules of games are analogous to the laws
of society, the rules of business, and so
forth. More fundamentally, general game
playing is valued as a test bed for
theories of intelligence. Game
descriptions provide full information
about a world and determine optimal
strategies. It is a base line for
evaluating the behavior of actual agents.
By its nature the GP setting can be used
to evaluate problem solving strategies and
by extension theories of intelligence by
taking into account representation,
incompleteness of information, and
resource balance. That said, it's worth
noting that there are some shortcomings of
the current state of GDP as a test bed for
AI. First, there's the tabula rasa
critique. It goes, as follows, human
intelligence is arguably the product of
eons of evolution. We are to some extent,
at least, wired to function well in this
world. General game players have nothing
at their disposal but pure mathematics.
Okay. That's true. However a key indicator
of intelligence is the ability of each
individual to function in radically new
environments. And general game playing is
exploring that capability, that play of
our intelligence. Second shortcoming to
GDP, is the learning critique. In GGP,
players are told how the world works. The
critiquos, that real intelligence requires
the ability to figure out new
environments. Okay, well there's no
question that the ability to learn is
essential. However, real intelligence also
requires the ability to use theories once
they're formed. And that's what general
game playing is emphasizing. This is
really the mean of what's often called
general systems, or universal systems, or
sometimes declarative systems. A nd the
interest in this problem dates back to the
very beginnings of the field. It was in
1958 that John McCarthy invented the
concept of the advice taker. The idea was
simple, he wanted a machine that he could
program by description. He would scribe
the intended environment and the desired
goal, and the machine would use that
information in determining it's behavior.
There would be no programming in the
traditional sense. McCarthy presented this
concept in a paper that has become a
classic in the field of AI. and let me
just read this quote. The main advantage
we expect the advice taker to have is that
its behavior will be improvable merely by
making statements to it. Telling it about
its environment, and what it is wanted
from it. To make these statements will
require little if any knowledge of the
program. Or the previous knowledge of the
advice taker. Well that's an ambitious
goal. But that was a time of high hopes
and grand ambitions. The idea caught the
imaginations of numerous subsequent
researchers, notably Bob Kowalsky, one of
the high priests of logic programming. And
then Feigenbaum, the inventor of knowledge
engineering. in a paper written in 1974,
Feigenbaum gave his most forceful
statement of McCarthy's ideal. He says,
the potential use of computers by people
to accomplish tasks, can be
one-dimensionalized into a spectrum,
representing the nature of instruction
that must be given the computer to do its
job. Call it the what-to-how spectrum. At
one extreme of the spectrum, the user
supplies his intelligence to instruct the
machine with precision exactly how to do
his job step by step. At the other end of
the spectrum is the user with his real
problem. He aspires to communicate what he
wants done, without having to lay out in
detail all necessary sub-goals for
adequate performance. Okay, one final
remark. Some have argued that, the way to
achieve intelligent behavior is through
specialization. That may work, so long as
the assumptions one makes in building such
systems are true. For general intellige
nce, however, general intellectual
capabilities are needed. And such systems
should be capable of performing well in a
wide variety of tasks. To paraphrase the
words of Robert Heinlein, a human being,
or a computer robot, should be able to
change a diaper. Plan an invasion, butcher
a hog, con a ship. Design a building,
write a sonnet, balance accounts. Build a
wall, set a bone. Comfort the dying. Take
orders, give orders. Cooperate. Act alone.
Solve equations. Analyze a new problem.
Pitch manure. Program a computer. Cook a
tasty meal. Fight efficiently, die
gallantly. Specialization is for insects.
Those of us who are more interested in
artificial intelligence than artificial
insects agree with Heinlein.