 |
|
October 31, 2005 Vol. 77,
no. 10F
 |
| FUZZY
AND SO CLEAR: “Coming to Berkeley
was the most important thing in my life,” EECS professor
Lotfi Zadeh says. “It offered new challenges, a new environment.
Without Berkeley, I wouldn’t be who I am today.” Zadeh
holds 23 honorary doctorates. In December 2004, there were 40,352
papers in the INSPEC database containing “fuzzy” in
the title. (Photo Credit: Peg Skorpinski)
|
The world
is a matter of degree
EECS professor reflects on his pioneering theory, Fuzzy logic
In the Soda Hall office of EECS professor Lotfi Zadeh, there are so
many books and papers stacked floor to ceiling that only a small footpath
remains. The contents represent a lifetime of work that began before
the age of computers and continues to proffer theories about them today,
55 years later.
At the center of it all is Fuzzy logic, a theory that challenges classical
logic’s belief in absolute true or false. Its applications can
be found in everything from cameras to car transmissions, elevator control
to medical instrumentation. Zadeh, known worldwide as the “Father
of Fuzzy logic,” will be at the center of the EECS department’s
November 2-5 conference and celebration commemorating the fortieth anniversary
of his pioneering theory.
Zadeh was born in Soviet Azerbaijan in 1921. When he was 10, his family
moved to Iran, where he attended an American Presbyterian missionary
school. “By six or seven, I already knew I wanted to be a scientist
or engineer,” he recalls. He studied electrical engineering at
the University of Tehran, graduating in 1942. He received his master’s
from MIT, his Ph.D. from Columbia, and, in 1959, was recruited to Berkeley
from a full professorship at Columbia.
By then, Zadeh was engrossed with systems theory. Nine years earlier,
he published the paper, “Thinking Machines: a new field in electrical
engineering.” Zadeh, it turns out, had foresight. (As chair in
1967, he played a leading role in changing the name of Berkeley’s
department from EE to EECS. Although controversial, it was the first
of its kind, and other universities worldwide soon followed suit.)
He also had courage. “I’m an electrical engineer by training,
and I’ve always been an admirer of mathematicians; but I began
to see a gap between the precision of math and the imprecision of the
real world.” In 1965, he published a paper introducing “fuzzy
sets,” which are sets with “unsharp boundaries.” (Thus,
fuzzy.) A precursor to Fuzzy logic, the concept was received with skepticism,
even hostility.
Zadeh persisted. “In Fuzzy logic, everything is or is allowed
to be a matter of degree,” he explains. “This is the way
human thinking is organized. In the real world, almost nothing is black
and white.” He wanted computers, too, to run on gradations, not
binary absolutes.
In 1973, he introduced “linguistic variable,” describing
how systems function not with numbers, but with words. It approximated
how humans describe things and allowed machines to more closely mimic
human decision-making. Soon after, the Japanese began to use the concept
of linguistic variable in their commercial technologies, and it determines
how many appliances and products function today.
Fuzzy logic is still generating controversy, but it has evolved into
its own field. Zadeh spends most of his time lecturing at conferences
on the topic. At 84, he’s still coming up with new theories, such
as a Fuzzy logic-based approach to computation with information described
in natural language. Altogether, it’s a brimming career worthy
of celebration.
|
|
