[ Boolean Logic | Fuzzy Logic ]
ne strong AI approach to AI is by using logic as the basis
of all development and research. This school of thought originates from the view
that AI can be achieved through systematic exploration that science demands.
Another reason why logic would be applied to AI development is because the system that
inherits and tests AI, generally computers, are machines operating on both the logic of
the physics that govern its inner workings as well as programming the computer to run the
AI programs requires logic as well.
Invented by George Boole who developed an entire
aspect of algebra on this topic, boolean logic is a simple and systematic
way of thinking and reasoning. In mathematics and philosophy, boolean logic can be
applied to assess the validity of a mathematical expression or a sentence. Let's
start with a set of statements followed by a "true" if the statement is valid
and "false" it is invalid.
- Apples are red--true.
- Violets are blue--true.
- Three plus three is equal to seven--false.
If the statements were linked together with "AND" or "OR" or negate
the statement with "NOT" in combinations of pairs, then the validity of
the entire sentence can be established based on the validity of the sub-statements.
Thus,
The combination of statements allowed a sentences or a mathematical expressions to
clearly express their meanings simplistically yet be powerful and effective. Thus,
this was an attempt to formalize thought and reasoning at the mathematics level.
Computer scientists don't use "true" or "false" to perform
calculations. Instead, they use a more abstract method involving the binary system with true represented as a "1" and false is
represented as "0". In the case of computers, "1" is considered
"on", i.e. the existence of an electric current in a particular area of the
computer, while "0" is the absence of electricity. The routing of
electricity by the computer and its interpretation through boolean logic that allows a
computer to calculate and make decisions to serve the user.
Seemingly as rebellious as the hippies were in the 1960s, fuzzy logic was an attempt to
break away from the traditional and inflexible thinking of logic programming and create a
machine that acted more "organic." What this means is that while a
computer operates on logical principles, what it actually does does not have to be
straightforward logic. This kind of operation would allow a computer to make
intelligent decisions based on incomplete information just as real organisms, especially a
human being.
Fuzzy logic actually relates to boolean algebra in that instead of
having only two values like "true" and "false", it has a range of
values such as "always true", "sometimes true", "sometimes
false", and "always false" to address statements that express something in
general terms like "some apples are red."
In real life, most things cannot be entirely expressed accurately in a definite
"either-this-or-that" statement. This point is what separates computers
from people: computers are digital machines that cannot definitely capture or represent
something that is analog while a person, whether his/her
brain is digital or not, can process analog information
to a degree of accuracy good enough to make precise decisions. In AI, fuzzy logic
typically applied in pattern-recognition in which a picture or a sound recording does not
have a clear representation of an object, but enough of it is there for the computer to
recognize it using fuzzy logic. Also, fuzzy logic can prevent computers from
becoming lost in decision-making as it journeys from one decision to another without
making much progress. The special logic could allow computers to identify the faulty
train of thought that cause the never ending loop, terminate the process, and still make a
decision with a few pieces of information that could be processed.
