|
In the first installment of this exclusive interview with Lotfi Zadeh, the father of fuzzy logic, Zadeh discusses the philosophical underpinnings of fuzzy logic, how it relates to fractals and AI, and his youth in the USSR and Iran.
July 01, 1994
URL:http://www.drdobbs.com/architecture-and-design/lotfi-visions-part-1/184409272
Even at 73 years of age, Lotfi Zadeh, the father of fuzzy logic, has an energetic stage presence. Nowhere is this more evident than in the rapt attention he commands when presenting his paper, "Fuzzy Logic: Issues, Contentions and Perspectives," at the 22nd Annual ACM Computer Science Conference in Phoenix, Arizona on March 8, 1994.
It is during this session that, once again, Zadeh's colleague and friendly personal gadfly, Professor William Kahan, rises to challenge, for three full minutes, Zadeh's lifework as an assault of illogic upon the scientific foundation of control engineering. "A scientific idea is one which contains within it the germ of a refutation," says Kahan. "A test can then be posited whereby the hypothetical refutation can be proven or disproven. Fuzzy logic has no scientific content because it doesn't assert anything upon which we can model such a hypothetical refutation."
Zadeh's lips tighten into a tolerant smile. He toys with the lapels of his suit coat, he lowers his eyes and rocks back and forth slightly at the podium as he again hears the familiar arguments.
Afterwards, Zadeh (referred to as "LZ" in the following interview) and I chat on subjects ranging from the theory and economics of fuzzy logic, to AI, fractals, philosophy, and Zadeh's boyhood in Stalin's Soviet Union. We are joined by William Kahan (WK) who, like Zadeh, is a professor at the University of California, Berkeley, and John Osmundsen (JO), associate director, public affairs, of the ACM.
In the first installment of this two-part article, Zadeh examines the philosophical underpinnings of fuzzy logic, how it relates to disciplines such as fractals and AI, and his youth in the USSR and Iran. Next month, we'll discuss fuzzy applications, such as Japan's Sendai train, and hear in detail what Professor Kahan thinks of fuzzy logic.
DDJ: What you said in your lecture today that struck me the most is that fuzzy logic is a means of presenting problems to computers in a way akin to the way humans solve them.
LZ: There is one way of expressing that, which I use sometimes: The role model for fuzzy logic is the human mind. If you examine the way the human mind functions, you find that the human mind has this remarkable capability to deal with information which is incomplete, imprecise, uncertain, and so forth. Computers do not have that capability to any significant extent.
Classical logic is normative. Classical logic in effect tells you, "That's the way you should be reasoning." There is a big difference in that sense between the spirit of classical logic, which is prescriptive, and the spirit of fuzzy logic, which is descriptive. That is, it merely asks the question "How do you reason about this or that?" It's like translation. The translator does not take responsibility for what he or she translates.
DDJ: And the analogy of translation to fuzzy logic is that fuzzy logic is simply a translation mechanism?
LZ: Well, there are many facets to fuzzy logic, so you cannot summarize the whole thing in one sentence. I'm talking here about one particular facet of fuzzy logic. That facet has to do with most practical applications today in the realm of consumer products and in many other fields, where what you do is you use the language of fuzzy rules. You start with a human solution and you translate it into that language. But it doesn't mean that that is all there is to fuzzy logic, because there are many other things that fall within the province of fuzzy logic that would not fit this description. The essence of fuzzy logic is that everything is a matter of degree, including the notion of subsethood.
DDJ: Is this a philosophical point of view that finds itself translated into computer logic--the notion that everything is a matter of shades, and varyings, and degrees? Is there a personal philosophical viewpoint that is finding its expression in computer science in your work?
LZ: Not yet, but consider the following: The real world that we live in is very fuzzy, very imprecise, very uncertain. The theories that we have constructed are, on the other hand, very precise. We have mathematics, we have all kinds of things which are very precise in nature.
Now, these theories have proved to be very successful in many respects, but their ability to come to grips with the analysis of complex systems--I mean complex not just in terms of number of components, for example, a chip that has two million resistors_. When I say "complex," I mean "complex economic systems" and things of that kind, systems with many components, with relationships that are not well defined. The successes of classical techniques, in connection with certain kinds of systems, have led us to believe they can be successful also in dealing with the other types of systems, such as economic systems.
There is no question about it, classical mathematics has proved to be very successful in astronomy, where you compute the orbits of stars and planets. But people then conclude from that that you can apply mathematics equally successfully to laws of economic systems. And that's where I question this thing.
I say "No, these systems don't fit. You need a concept of classes which don't have well-defined boundaries." And if you do, as fuzzy logic attempts to do, construct such a framework, then you enhance your ability to model economic systems and other systems of that kind. It doesn't mean you'll be able to solve all the problems, but at least you'll be able to do much more.
This applies, for example, to natural languages. Notice that we didn't make that much headway in machine translation, and essentially nothing in machine summarization. If I ask you to write a program that will look at a book and summarize it, I think you'll say that we cannot do that. Not only can't we do it today, there's no way that we can conceive of doing that in the foreseeable future.
The situation is this, that there is this tradition of believing that conventional, traditional techniques have the power within them to solve some of these problems. My position is that this is not the case.
DDJ: It sounds like you are making a point analogous to that of Benoit Mandelbrot in The Fractal Geometry of Nature, in which he pointed out that mathematicians at the turn of the twentieth century, including his father…
LZ: His uncle…
DDJ:_his uncle, had examined fractal forms and had been roundly criticized in the math world for studying these "monstrosities," and asked why they did not go back to studying genuine geometrical forms such as the sphere. Mandelbrot points out that there are no spheres in nature, no squares, nothing which has perfect form to it. It's just a question of to what degree our measurement instruments are able to penetrate the form and discover the imperfections that nature has placed there--the "sacred error," as it were.
LZ: I know Mandelbrot quite well, and hold great respect and admiration for him. Mandelbrot stays within the traditional framework. He does consider objects, fractals, and he did raise questions of the kind that other people somehow did not raise, for example: What is the length of the coastline of Brittany? To me, that's a very good question, because somehow people took it for granted that there is an answer to that question. But he pointed out that there isn't an answer to that question, because it depends on the degree of resolution.
DDJ: It ends up being a question about your measuring instruments, and not about the problem you're trying to solve.
LZ: That's right. So I think it was a very incisive observation that questions like that cannot be answered within the traditional framework. But what Mandelbrot tried to do--and I think it's a very significant accomplishment, but different from what you do in fuzzy logic--he attempted to come up with a reasonably precise theory of this sort of thing. So he talks about fractional dimension. Basically, Mandelbrot is a mathematician by training, and he has not abandoned his home, so to speak.
So to me, the theory of fractals is an important theory, and it helped to focus attention on issues that were not really properly formulated before. But by itself, it stays within the traditional paradigm. In other words, you're still committed to the goal of mathematicians: to come up with theorems. I'm not saying that that goal is not a worthwhile goal, I'm merely saying that in many cases it is unattainable.
DDJ: When you come down to the field of practical engineering, especially in problems of embedded control, the problem of making machines that can in real time make, if not perfect decisions, then reasonable decisions…
LZ: Certainly.
DDJ: My father is 74 years old, and he doesn't know which end of the computer you hook up the airhose to, but he knows what fuzzy logic is because he has been an amateur photographer for 60 years, and his Japanese camera can determine the illumination of dim objects against bright background light--using fuzzy logic. "Fuzzy logic" is also part of the advertising.
LZ: Minolta uses fuzzy logic very extensively.
DDJ: Do you derive any royalties from this?
LZ: Zero. The thought of applying for a patent did not even occur to me at the time I did the work.
DDJ: Is this an oversight which you regret?
LZ: No, not at all. Perhaps I would be a rich man, but so long as I can live in reasonable comfort, that's enough.
DDJ: Is it possible that if you had patented this technology it would not have been so readily adopted?
LZ: It's difficult to predict what could have happened, because sometimes things evolve in [an] unpredictable fashion. Just to give you an example, the first consumer product [to use fuzzy logic] was a Panasonic showerhead which came out in 1987. If someone had asked me in 1986 what sort of applications would you expect, it wouldn't have occurred to me that there would be all these applications in the realm of consumer goods.
DDJ: What did you envision?
LZ: Industrial applications, yes. Industrial control, traffic control, applications in linguistics, yes. But not washing machines, not microwave ovens, none of those things would have occurred to me. So it shows that even a reasonably well-informed person in that field may find it very difficult to predict how things will evolve.
I could see that control was going to be an application, and in 1972 I wrote a paper called "A Rationale for Fuzzy Control." But my colleagues in the area of control didn't share that feeling, and to this day, most of the control-systems community is very antagonistic towards fuzzy logic, as is the AI community.
DDJ: I can understand the control-engineering community's wariness, which possibly is based on unfamiliarity, since in the classic control theory, proportional-integral derivative, and the like, so much time has been committed to working out these theories, and so much of the training of the people involved has dealt with this. You don't want to teach an old dog, in the form of a well-established school of engineering, new tricks.
In the AI field, I wonder if there is professional jealousy that AI didn't quite take off as a commercial proposition, whereas fuzzy logic has done so.
LZ: The issue is somewhat complex. Incidentally, many of the people within the AI community were very hostile toward fuzzy logic. Among them are good friends, so there is nothing personal about it.
I think that what happened is that, for historical reasons, AI embraced classical logic, classical predicate logic, symbolic logic. And so, the intellectual leaders of AI, who were deeply committed to that kind of logic, took the position of prescriptive logic, telling you "that is the way you should be reasoning." Since computers are symbol-manipulation machines, it appeared that computers could implement that kind of reasoning and go beyond what humans can do.
AI has hitched itself to symbol manipulation. As a result, most people in AI dislike numerical computation. They dislike not just fuzzy logic, but they dislike probability theory, neural networks, anything that involves numerical computations. Things are changing, and I think that gradually you'll find the number of people within AI who use numeric computations, [and] the number of papers at AI conferences in which numerical computations are used in one form or another, are increasing. It's a gradual process. If you go back five or ten years you will find that AI was almost entirely symbol-manipulation oriented.
Fuzzy logic is not symbol-manipulation oriented. It's computationally oriented. Because of that, it simply did not sit too well with the AI community.
Also, it's a human thing: You sit at a table and the pie is sliced, and the more people who sit at it, the smaller your slice of pie gets. It has happened in a number of fields. There is resistance to something that may result in a smaller piece of pie for yourself and perhaps, more importantly, may depreciate the value of your knowledge. This is what is happening in AI now, because many people will use neural networks; they use, to a lesser extent, fuzzy logic; they use things that were not in the mainstream of AI. Many people in AI see that as a threat. They see that as an intrusion of people whose thinking is different from their own.
In that respect, it's not very different from many of the sociological phenomena that we observe. Take the way people dress. People who are committed to a classical-logicalesque reasoning are people who dress very properly. They have a tie and a shirt and the colors match, shiny shoes, starched shirt, and so forth.
The fuzzy-logic people dress informally. Their shirt may have colors that don't match perfectly, they don't worry too much if the shoes are this kind or that kind. These people do have more traditional clothes somewhere in their closet, so that if they have to go to a party where that's what's expected, they have something to put on. But not the other way around! The traditional people would not have the other kinds of clothes in their closet, because they would never stoop to wearing such a thing.
People who dress very conservatively, when they look at people who dress very informally, they don't like it. They feel that "these people are not my people." And the people who dress very informally, when they look at people who dress formally, say, "these people are old-fashioned, they are not my people."
DDJ: So you subscribe to the theories advanced by Thomas Carlyle's fictional Herr Teufelsdroeck, who said in Sartor Resartus that clothes really do make the man.
LZ: (Laughs) So you do really have different philosophies. There is a defense mechanism that I have observed which is part of all of us, that if there is something that is unfamiliar to you, you convince yourself that it is not worth learning, because if that were not the case, then you would have to learn it. But if you convince yourself that it is garbage, or uninteresting, then that absolves you from the need.
DDJ: As long as we are on the subject of predisposition to certain viewpoints, may I ask you a personal question?
LZ: Sure.
DDJ: Are you a Moslem?
LZ: I am not practicing. Let me explain something about my background. I am of Iranian descent, but I was born in the Soviet Union, not in Iran. My father was a correspondent for Iranian newspapers. So I was born there but was not a Soviet citizen. When I was ten, my family moved back to Iran.
I went through the first three grades of elementary school in the Soviet Union. That was at the height of antireligious propaganda. I was born in 1921, and we left for Iran in 1931. At that point in the Soviet Union, no one would dare admit that he or she believed in God. That would have been sort of, sort of…
DDJ: Suicide?
LZ: Suicide. No one would talk to you, no one would associate with you. That was the sort of environment in which I grew up in those particular years.
Then when my parents returned to Iran, they placed me in an American Presbyterian missionary school where we had chapel every day at 10 o'clock, if you can imagine the transition from the one environment to the other environment!
Then, after studying there for a few months, I had to leave the school because the Iranian government at that time was nationalistic, and a law was passed to the effect that you couldn't go to a school run by foreigners without first completing an elementary school [education] in Iran. So I had to move from that school to an Iranian school.
DDJ: This was in the time of Mohammed Reza's father?
LZ: Exactly. Now, at that time Iran was anticlerical. That's what many Americans don't realize because of the Iranian Revolution. But before the revolution, Iran was highly anticlerical country--many of the imams were in prison--but antisecular at that same time. In Iran, Islam was even then the state religion.
It was assumed that you were Moslem unless you declared otherwise. So, in that Iranian school, if a Christian were to visit the school, all the walls had to be washed. I experienced these extremes of fanaticism.
When I was at the University of Tehran later, there was only one professor who was known to go to a mosque. That was Bazargan, who later became premier. If you said that you were going to a mosque, people would laugh at you. That was essentially the spirit of the times.
As a result of being subjected to these different influences, I became tolerant of different points of view. Because these people believed passionately one thing, and these people believed passionately another, and they couldn't all be correct. I began to realize that people have these passionate beliefs in something because they are insulated from people who believe in something else, so it sort of feeds on itself.
DDJ: My question had been aimed at this: It seems that what you have been teaching for the last two decades, and then some, has been more accessible to people from certain backgrounds than from others. It doesn't seem to be merely a question of what programming discipline one comes from. Fuzzy logic seems to be more accessible to people of different cultures, certainly more accessible to the Japanese, and to the coterie of Iranian students and former Iranians that has formed around you in this country that seems to find these ideas very easily accessible. Do you think that there is anything in cultural makeup that would make fuzzy logic acceptable?
LZ: In part. But some people try to explain the whole thing from that particular angle. I would not go that far, but there is some truth to it, in the following sense: When you talk about such cultures, you are talking about old cultures. When you talk about western cultures, you are talking about young cultures. Young people tend to be more dogmatic in their views than old people, because when you grow older, you become more keenly aware of the fact that, well, this is true, and that is true. You become more willing to concede that truth is not a monopoly of some particular way of thinking.
The thing to remember is that all traditions have a certain amount of validity, and beyond that point, they may become wrong or counterproductive. In this quest for precision, which is very characteristic of western culture, that's where fuzzy logic enters the picture. I say "Stop, maybe you should get off the train. Maybe it's not taking you to the solution to your problem. You have to reexamine things." And that is why some people become upset about this.
DDJ: You're introducing a degree of uncertainty into their life.
LZ: That's right. You're saying their goals may not be realizable. These views were expressed by Professor Kahan during my lecture: "We need more precision, more logic." He's a mathematician. I have very great admiration for him. You can see a forceful expression of that kind of thinking in him, and I agree that we need this expression, because it is precisely there that I disagree. I say that that kind of thinking has its place, but it also has its limits.
DDJ: It seems to presuppose that man has a godlike power to determine, incontrovertibly, the truth or error of a proposition.
LZ: What I'm very conscious of is that there are many things people accept without question, even though those things don't make any sense. They accept it because it's part of the tradition. I have many transparencies that I use in my lectures that illustrate how people come up [with] very precise conclusions on the basis of data which are completely unreliable.
Just to give you an example, there was a study made on the effect of removing lead from gasoline. The conclusion: "The removal of lead from gasoline will result in 1223 fewer cases of high blood pressure and 237 fewer cases of this_." It doesn't make any sense. How can you determine these things with such accuracy? Predictions of AIDS that by 1998 there will be so many cases of AIDS. How can you say this? People accept these things, because that is the tradition. You're supposed to swallow such assertions.
DDJ: Our society respects statistics as the ancient Babylonians respected the stars.
LZ: Given that it is respected, that's what people accept. Even if it doesn't make any sense.
JO: Scientists agree with Professor Zadeh. It's journalists who publish these numbers.
LZ: By the way, one of the interesting things we have looked at lately is establishing an "MIQ," a machine intelligence quotient.
DDJ: IQ is a measure that is determined by statistically analyzing those facts which are the common property of the civilization, and ranking people on the basis of what percentage [of] those cultural ideas the examinee is in possession of. How would one begin to establish a quotient on machines?
LZ: You have a standards committee. This standards committee considers various products and establish certain criteria. So there are dimensions to intelligence.
DDJ: Chess computers are rated these days.
LZ: So we would have to come up with a rating system agreed upon by a standards committee which can put its imprimatur upon the system. Many things are rated. Consumer Reports rates things.
DDJ: Once again you are proposing something that has tremendous commercial implication. Advertisers would love to say "Our washing machine has an MIQ of 190."
LZ: That is why I registered MIQ as a trademark! I hope to make some money out of this! I'm pretty sure that at some time in the future when you open Consumer Reports and read a report on washing machines, you'll see a column labeled, "MIQ." It's basically a measure of user-friendliness.
DDJ: You are reeducating America and Japan's young control engineers to a whole new paradigm of control engineering, yet you are in your seventies and show no sign of diminishing your activity.
LZ: Thank you. It's one of those situations where today I'm okay, and tomorrow I may be dead! Professor Kahan gave me this, which reads, "I hope you enjoy good health long enough to be invited to the White House to receive from the President a medal for so successfully distracting the Japanese for so long."
DDJ: Does Professor Kahan truly believe that you are leading the control industry astray?
LZ: I think so. I think he's very serious.
DDJ: Yet, in view of the vast commercial success of these products, how can an American argue with money?
LZ: We're very good friends. He is very sincere in his feelings that this whole thing is pernicious. Even if it became so ubiquitous that every product in Japan and the United States used fuzzy logic, he would still hold onto the opinion that it is wrong.
JO: Perhaps if you reassured him you would make no fuzzy-controlled nuclear bombs_.
LZ: (Laughs)
JO: But what if you proved that fuzzy logic works?
LZ: That's not good enough, because some people, not Professor Kahan, say that people aren't really using fuzzy logic, that it's just an advertising gimmick. Professor [John] MacCarthy at Stanford University makes that argument. Other people say that, well, yes they do use fuzzy logic, but they have not shown that they could not have attained this same result using standard methodology.
My response is "The Japanese, for instance, are not that stupid. If they could have achieved these things using standard techniques, why would they go into fuzzy logic?" It's not just a merchandising gimmick. You can't explain the whole thing by a merchandising gimmick.
The fuzzy-logic Sendai train started running in 1987. It's very successful. Now they're going to use it in Tokyo. A merchandising gimmick?
In the next installment, Professor Zadeh discusses the Sendai train in detail and the Japanese approach to fuzzy logic. Then Professor Kahan arrives on the scene.
Jack is a frequent contributor to DDJ and can be contacted at [email protected].
Terms of Service | Privacy Statement | Copyright © 2024 UBM Tech, All rights reserved.