Joe Gallian's Interview with Doron Zeilberger February 20, 2007

发布时间:2021-11-16浏览次数:10

MAA Focus May/June 2007 issue.pdf

 JG: How do you describe your research area?   DZ: I work mainly in combinatorics and the theory of       special functions, but in the last 10-15 years I consider myself mainly as an      experimental mathematician, and combinatorics and the theory of special      functions are just sources for examples and case-studies       of a methodology that its aim is to train the computer to discover conjectures and      then try to prove them all by itself, without any human intervention.   JG: Are there now a number of people who are doing experimental      mathematics?   DZ: Experimental mathematics is a rapidly growing field, both explicitly and implicitly.      Explicitly, there is a very good journal by that name, an Institute in Simon Fraser      University, and at the recent annual meeting at New Orleans there was a special      session dedicated to it. This is a good start, but still at its infancy.      However, implicitly, more and more mathematicians, even pure ones, use the      computer daily to formulate and test conjectures, in a mode that George      Andrews calls "pencil with power-steering". However, more often than not,      the computer's often crucial contribution is not mentioned, or grossly      understated. Human beings are such ingrates.             Also, Jon Borwein and David Bailey wrote a very nice monograph on Experimental Mathematics that       I highly recommend. But their emphasis is more on the traditional experimental mathematics that      has been pursued by all the great, and less-great, mathematicians      through the centuries, using pencil-and-paper.      Of course, with computers you can do so much more, and you can be  very systematic, and the great       power of today's computers , guided wisely, can take you a very long way.      However,  their emphasis is still on using computers  to find interesting      conjectures and phenomena, but not to prove them.      The  proof itself (when feasible) is still done largely by human beings, although sometimes with      assistance of computers.       As for computer-generated proofs, there is  a whole community of Automatic Theorem Proving, that      is very successful. Their methodology is logic-based, and      they try to teach the computer formal reasoning, using axioms      and "laws of deduction". My approach is more akin to what      the great algebraic geometer Shreeram Abhyankar calls "high-school algebra",      and Richard Feynman calls "Babylonian mathematics", using      algorithmic frameworks, that I call "ansatzes". But since      the algorithms are symbolic rather than numeric, one can hope      to prove genuinely new general theorems, valid for infinitely many      cases.  In this kind of research, we try to teach the computer      to firs make a conjecture, all by itself, and then automatically prove its own conjectures,       also all by itself!      Of course, the only way we can do it, at present, is by focusing      on  narrowly-defined areas.   JG: Are you getting some resistance from the traditional math      community with your approach?   DZ: Definitely. Many people are not very comfortable with this      approach. First, they claim that you can't trust computers      (as though that humans are so trustworthy), and they also      feel that it's not fun to have a computer do your work, or that it is      "cheating". For me, teaching the computer how to discover and prove      new theorems is even more fun than discovering and proving them myself, and as for       the cheating part, this is science not sport, and besides we can always      change the rules of the game.   JG: The Wilf-Zeilberger algorithmic proof theory is contained in all major       computer algebra systems. What does it do?   DZ: It's a collection of algorithms that can discover, and then prove,      binomial coefficient identities, and identities involving sums or      integrals of special functions.      As you know, the binomial coefficient n choose k counts      the number of k-element subsets of an n-element set.      Since many objects in combinatorics boil down to      unions and cartesian products of such choosings, it turns out      that many enumeration problems can be expressed as such      binomial coefficient sums, and often we get surprising      identities. Before W-Z theory, every such identity required      its own ad-hoc proof, and any such new identity, or even a new      proof of an old identity, was publishable. But thanks to      Wilf-Zeilberger theory, any such identity can now be proved      automatically. No one today would submit a paper      stating and proving the "theorem" that "ten plus five equals three times five".      Analogously, an identity like       Sum((-1)^k*binomial(2*n,n+k)^3,k=-n..n)=(3*n)!/n!^3,       first proved by Dixon in 1904, is now completely routinely      and automatically provable thanks to the Wilf-Zeilberger      algorithmic proof theory.   JG: You seem  to have an evangelical-like faith that computers will      make obsolete the way mathematics has      been done for over 2000 years. Is that a fair assessment of your view?   DZ: Yes, but people who know me well know that they should not take me too      seriously (laughing). In the Talmud it says that if you have      a talit (garment) and one guy says that it's all his and the      other guy says that it's all his, then they each get half.      But if somebody only claims a half and the other guy claims a whole, then      the guy who claims a half will      only get a quarter and the other guy gets three quarters.      So if you want a half, you have to claim a whole. So you have to overstate      your case. Then again, sometimes overstating can backfire and turn people off.   JG: Tell me about your frequent co-author Shalosh B. Ekhad   DZ: It is a charming individual. Of course it is made of silicon, and it is      not really ONE body,  but it is definitely ONE soul (software).       The body has just been reincarnated many time. As we know, computers are very powerful,       but their life expectancy is much shorter than that of  humans,       since computers get better and faster so quickly, you have to get a new one every three years,       but you can always upload all the software from one Shalosh to the next, thereby guaranteeing      the immortality of its soul.   JG: Where did you get that particular name?   DZ: The original Shalosh B. Ekhad was actually a Hebrew translation of the first PC that I owed,      called AT&T 3B1.  At the time it was a very innovative machine,  the first UNIX PC, that was manufactured by       AT&T  in the 80s. The Hebrew translation of 3B1 is Shalosh B. Ekhad.   JG: Tell me about your book A = B.   DZ: It was written by Marko Petkovsek, Herb Wilf, and myself,   and is an elementary introduction to so-called       Wilf-Zeilberger algorithmic proof theory mentioned above . It also has a long chapter on      the very important Petkovsek algorithm for deciding whether a sequence is closed-form.   JG: It was quite a coup to get Donald Knuth to write the preface. How did      that come about?   DZ: First, and foremost, Knuth is a good friend of Herb Wilf. Also      Knuth loves  binomial coefficient identities. He dedicated      quite a few pages to them in his classic "Art of Computer      Programming", which is the "bible" of computer science,      and  later, in much more detail, in his beautiful book with        Ron Graham and Oren Patashnik, "Concrete Mathematics: A Foundation for Computer Science".      He still loves them. In the latest  issue of American Mathematical      Monthly (Feb. 2007) there is a problem proposed by him      that could be easily done using the WZ method.       I was a little taken aback. If somebody submitted      a problem to prove that 11*13= 12^2-1,      they wouldn't accept it, since there are nowadays (and have been      for the last 5000 years) algorithms that can routinely prove this.      Similarly, for Knuth's Monthly problem there are (and have been for      the last 15 years)  algorithms that can routinely prove Knuth's problem.      But then again, Knuth is Knuth, if he proposed as a problem      to prove that 1+1=3-1, it might be accepted.       Of course, Knuth is very much aware of WZ, and he      probably had some hidden agenda in proposing this problem.      Maybe he meant some cute combinatorial argument.   JG: I notice that the preface of your book "A = B" begins with Knuth's famous      quote  "Science is what we understand well enough to explain to       a computer. Art is everything else we do."      Was A = B the first place that quote appeared in print?   DZ: Yes, I think so.   JG: I was very surprised that a commercial publisher would agree to      permitting the book to be downloaded free over the internet.   DZ: Herb Wilf is a great advocate of free publishing and A. K. Peters       is not a typical publisher. Klaus Peters, and his wife Alice,      are very good people, who care more about mathematics than making      a quick buck. Commercially it seems to have been a wash,      some people who would have bought it, didn't, and vice versa.    JG: I notice that you prefer Maple over other software. Why is that?   DZ: Originally Maple was much cheaper and more accessible  than its      competitors. Mathematica was  much more commercial.      Unfortunately, now Maple is very commercial too, and      Mathematica's prices have gone down, and many people      prefer the latter for its elegant syntax and beautiful      graphics. So I use Maple for "old-time's sake" and       mainly because I am used to it.   JG: Describe your Experimental Mathematics course.   DZ: I really enjoy teaching it. It is a graduate class (with      one or two advanced undergraduate students) that is held      in a "smart" classroom where everyone is connected to  a terminal.      It is a hands-on course where the main point is to teach students      how to program in Maple, in order to explore new mathematics.      The actual topic changes every year, so that I won't get bored,      and sometimes I even learn a new subject myself, since the best way of       learning a new subject is by teaching it. Of course, in this class we      also learn how to program everything, and an even better way to learn a new      subject is by teaching it to a computer, i.e. programming. So both myself and      my students learn the substance of the course very well, and at the same time      my students learn how to program in Maple.   JG: You have an Erdos number (2), an Einstein number (3),       a Wiles number (3), and a Knuth number (2) but you      say on your website that you are most proud of your Garfield       number (2). Why is that?   DZ: Richard Garfield is one of the most talented and creative      combinatorialists alive, but he used his talents in a non-standard way. He is also a very nice guy,      and I am fortunate to have known him, although only briefly.      During the mid nineties, he was a teenage idol      thanks to his innovative and lucrative "Magic: the Gathering"      card game. He designed the game while he was Herb Wilf's      PhD student at Penn in the late eighties and early nineties.      He did good research, but not as much as he could have, because      he was so busy developing and testing his card game. After his PhD,      he got a one-year job in some small college in Washington State, with a salary of about      22K. When the year was almost up, and it was not clear what      his prospect for the future in academia would be, his card game      caught on, and the rest is history.    JG: There must be a story behind the "Who you gonna call" t-shirt.   DZ: This is a T-shirt that I am wearing in my picture on my website.      It features a certain binomial-coefficient identity, with the      caption "who you gonna call". The back of that T-shirt      has the few-lines of Macsyma code needed to prove it, and the caption      "the binomial-coefficient-identity-busters".      My kids told me that it is an allusion to "ghost busters".      This cute design was made my Herb Wilf's son, David, who is      a lawyer by profession, and Herb game me one of them.    JG: Something that I found surprising is that the narratives for your      grant proposals are posted at your website.      That certainly helps people learn about your work.   DZ: I think that it's a waste to write something just for 5 or 6 people      and for no one else to be able to see it.      I think that everything should be public; I don't like secrecy.      I also hate the tradition  of anonymous refereeing. I think that there should      be open refereeing. Also people should post their PhD theses, especially      the introduction, that often gives a very good overview of a field.    JG: Another thing that surprises me is that you slip a bit of humor in      your grant proposals. One is ``There is a      delicate balance and trade-off between the general and the specific, the      abstract and the concrete, the strategical and      the tactical, the sacred and the profane."  Another is the line ``My      particular shtick is experimental mathematics." A third is      The computer is a powerful tool, go forth and use it!" Your proposals      are less formal than I would expect.   DZ: That's my style, I don't like to be too serious. But I think that it has      cost me some funding.  In my last grant      proposal, I was hoping to get co-funded by the computer science division.      I had no problem getting funding from the      mathematics division, but of course mathematics doesn't have much funds,      so I was hoping to get some additional funding      from computer science.  So I also sent my grant proposal to the computer      science division. That was a disaster. One of the panelists      gave me a lecture, saying this might make a good essay, but it's      not what one would expect in a grant proposal, and      they refused to give me funding.    JG: You are known for your opinions on your web page, some of which are a      bit over-the-top, especially those on April 1.  What changes--if      any--are you hoping to encourage in the mathematics community by      posting your opinions online?   DZ: I'm not trying to change people's views, at least not consciously.      I just like to express my opinions. It's only the Internet,      which is a free-for-all, and one doesn't have to be too uptight.      So I don't really worry about whether or not my pieces always make sense.      Hopefully they do most, or least some, of the time.    JG: You have been known to celebrate both Valentine's Day and      April Fools day in the classes you teach. What is the most memorable thing      you have done to mathematically celebrate a holiday?   DZ: In my Calculus class, I assign a homework "project", due Feb. 14, to graph the      parametric equation for a cardiodid. Some      people came with very strange shapes, but some people realized what was      going on and just drew a regular heart shape.      But they made a pointy one, which is wrong, because a cardioid is more      rounded.  It was nice to get 150 valentines (some rounded some not),      and I could brag to my wife how much my students love me.      In my Computer Algebra class, I gave my students some extremely complicated differential equations that they had      to use Maple to solve. The solution happened      to be a cardioid. Then they had to draw it (using the Maple plot program), and      cut it out.   JG: You have been known to spontaneously hand out money to      students for solving problems or pointing out a correction in class.      What is the largest prize you have ever awarded, and what was it for?   DZ: For calculation  errors the most I've given out is $1. But for      conceptual  errors in a graduate classes, I think I have once given $10.        I also offer prizes for really challenging problems.   JG: Do you have any concluding words?   DZ: Yes, I believe that teaching is at least as important as research, and I put lots      of effort into my teaching , and take great pride when I do a good job.      Many research mathematicians dislike teaching and view it as an unavoidable chore,      but they are wrong. First, teaching is great fun, and secondly, it is very important,      since this is the future!       So already today teaching is at least as important as research. But in years to come,      when more and more original mathematical research will be conducted by computers,      the importance of the human research mathematician will diminish, and the importance      of teaching, at all levels, will increase tremendously. Also for a long time to come,      we still need humans to program the computer, but what is programming?, it is      teaching computers, and I am sure that being a good programmer and being a good teacher      (for humans) are strongly correlated. One of the reasons that I believe that      I am a good teacher is that I do so much programming, and am used to spelling      out each and every step. In short, the future of mathematics is in good teaching,      both to machines and to humans.