Lotfi A. Zadeh was born in 1921 in Baku, capital of the former Soviet Republic of Azerbaijan, where his Iranian father was working as a journalist. When he was ten, the family moved to Teheran (Iran), and Lotfi was sent to an American Presbyterian missionary school. Due to an administrative oversight, he was enrolled in a class several years ahead of his supposed academic level, but this did not stop him from excelling in his studies, and obtaining the third highest score in the entire country in his university entrance exam.
Zadeh graduated as an electrical engineer from the University of Teheran in 1942, one of only three students to do so amid the disruption created by the Second World War. In 1943, he emigrated to the United States, where he took an MS in electrical engineering at Massachusetts Institute of Technology (MIT). Six years later, he earned his PhD in the same subject at Columbia University (New York), where he would teach for the next ten years. In 1959, he joined the Department of Electrical Engineering and Computer Sciences in the University of California, Berkeley, where he headed the Berkeley Initiative in Soft Computing since 1991.
He published his seminal work on fuzzy sets in 1965, and in 1968 proposed his theory of fuzzy logic. More than two decades later, in 1991, he introduced another new paradigm: soft computing, a hybrid methodology embracing fuzzy logic, neural networks, evolutive algorithms and probabilistic reasoning. Author of 245 papers, his research has been cited on more than 90,000 occasions, according to Google Scholar, and the vitality and influence of the field he created is evidenced by the number of papers – around 253,000 – that include the word ‘fuzzy’ in their titles. Zadeh held editorial posts at 75 specialist journals and though fuzzy sets and logic have given rise to tens of thousands of patents, not a single one was in his name.
In recent years, he was concerned with the computational scenario where data and operations may be specified in a natural language, or what he called ‘computing with words.’
Zadeh was a U.S. citizen but retained his Iranian nationality. He died in September 2017, in Berkeley, California (United States).
“In my life I have never hesitated to defend my vision. Above my desk at the University of Tehran I had a sign with the word ‘alone’.” If electrical engineer Lotfi Zadeh’s mind-set had been any different, computer science might never have experienced one of its great conceptual leaps, and consumer electronics would be lacking one of its most original tools.
For Zadeh was the inventor of fuzzy logic, a methodology which initially met with mixed reactions, but had since enabled machines to operate with imprecise concepts, as humans do, and generated more than 50,000 patents in the last 50 years in Japan and the U.S. alone. It is for this contribution that Lotfi Zadeh received the BBVA Foundation Frontiers of Knowledge Award in Information and Communication Technologies.
Fuzzy logic is “a revolutionary concept and methodology that created a new field of research, and has proved powerful in many application domains,” in the words of the prize jury. “(…) it is now an essential component in a wide range of consumer products including video cameras, washing machines, and medical instruments, as well as in more complex systems such as automatic trains, wastewater treatment plants, and intelligent control of car engines.”
Zadeh, since 1959 a full professor in the University of California, Berkeley (United States), said he knew from the start that “fuzzy logic would be useful,” but could never have predicted such a host of applications. Because fuzzy logic grew out of something as impractical as an idea, a reflection on the nature of precision: “As a systems engineer, I have always been a strong believer in the power of mathematics to solve all kinds of problems, but I also came to realize that such a precise conceptual framework imposed constraints. Psychology, linguistics and so forth are not precise. I have always wanted to narrow the gap between the real world, with all its imprecisions, and classical mathematics. That was the origin of fuzzy sets and fuzzy logic.”
Taking humans as his model, in view of what he calls our “remarkable ability” to take decisions based on inexact, subjective information, Zadeh began searching for a way to teach machines to operate with imprecision and words rather than measurements and calculations.
The first step in this paradigm shift were the so-called ‘fuzzy sets’; articulated by Zadeh in a 1965 paper that would come to be among the most cited of the 20th century, with over 35,000 mentions. Fuzzy sets, as their creator put it, are classes “with unsharp boundaries.”
Human language is full of them: tall men, fast cars, pleasant towns… Conventional logic works by defining which pa-rameters decide an element’s membership or otherwise of a demarcated set: being more than 1.80 meters tall; exceeding 250 km/hour; having lots of green spaces…, but there are invariably times when the picture gets muddied – a tall man in Spain may not be tall in Sweden, for instance. Fuzzy logic, created as such in 1968, admits degrees of belonging to a given set – so the man in question could be ‘a bit tall’ – allowing systems to operate with imprecision.
Zadeh at first had to contend with the skepticism of many colleagues. Part of the controversy, he thought, may have started with the name: “The word ‘fuzzy’ has a pejorative connotation in English, and this turned out to be a handicap when it came to gaining the acceptance of the scientific community. In Asia they don’t have problems with the word ‘fuzzy’, so they were more receptive to my work; they don’t share the western vision that everything is either black or white.” Now, however, the usefulness of fuzzy logic is beyond question. It is what allows machines to “understand” instructions like “brake smoothly” or “refrigerate until the air is cool.”
And we are not just talking about the efficiency of small appliances. The Japanese city of Sendai’s automated subway network, opened in 1987, uses fuzzy logic to such effect that speed changes are appreciably smoother, while the system’s energy consumption is 10 percent less than that of conventional subways.
Zadeh defined himself as “obstinate and tenacious, and not afraid to get embroiled in controversy”; traits, he mused, that have “probably been beneficial for the development of fuzzy logic.” Beneficial too for other daring but ultimately successful propositions that he has worked on in his career: the concept of ‘linguistic variable’, created in 1975 and which decades later would lead to ‘computing with words’ – in which words are used in place of numbers for computing and reasoning; and ‘soft computing’ – a hybrid of different techniques, including fuzzy logic, to address problems where information is incomplete or inexact. “My papers are among the most highly cited in their field,” noted Zadeh, “consequently… my skin is thick: if somebody tells me I’m stupid, I say thank you very much for the compliment.”
Zadeh continued working on ways to bring computers and systems closer to natural language. His influence is evidenced by the number of papers – around 253,000 – that include the word ‘fuzzy’ in their titles. He held editorial posts at 75 specialist journals and though fuzzy logic has given rise to tens of thousands of patents, not a single one was in his name.
The question of his nationality, broached in numerous interviews, invited parallels with fuzzy sets: Zadeh, who emigrated to the United States in 1942 after earning a BA in electrical engineering from the University of Tehran, was a U.S. citizen but conserved his Iranian nationality. As he said in an interview: “The question really isn’t whether I’m American, Russian, Iranian, Azerbaijani, or anything else. I’ve been shaped by all these people and cultures and I feel quite comfortable among all of them.”