Joel Spencer

Professor, Computer Science and Mathematics Depts

Email: {lowercaselastname}@cims.nyu.edu


Department of Computer Science and Department of Mathematics
Courant Institute, New York University


If Creativity were anything but Random, Someone would have figured out the Algorithm by now.
-- Dilbert

Contact

Room 729, 251 Mercer St.
New York, NY 10012, U.S.A.
212-998-3219 (voice) 212-995-4124 (fax)

CV/Papers/Talks

  • Vita--Selected Work (short form)
  • Vita--Publication List (long form)
  • Papers Description of selected papers. Vita and those papers in postscript.
  • Talks Slides for various talks.

    Course Information

    In Spring 2012 I am teaching

    (i) Random Graphs. This is be a graduate level course on the Probabilistic Method and Random Graphs. It will be crosslisted (one can register under either number) with Math and Computer Science.

    CSCI-GA 3033-007; Random Graphs: T 5:00 -- 6:50; ciww 317

    MATH-GA 2932.002 Advanced Topics in Probability (Random Graphs): T 5:00 -- 6:50; ciww 317

  • WebSite for Course

    (ii) Algebra II. Undergraduate V63.0344, MW 11-12:15.

  • Info
  • WebSite for Course

    Queries about either course should be emailed directly to me. These files will be updated as more information becomes available.

    The Erdos-Selberg Controversy

    Some 60 years ago a controversy erupted between Paul Erdos and Atle Selberg over the elementary proof of the Prime Number Theorem. Ernst Straus -- then a young mathematician at the scene -- wrote a description of the events. These have been published in Math Intelligencer, along with commentary by Ron Graham and myself, and a mathematical postscript by Carl Pomerance.

    click here for the article

    The Probabilistic Method -- Third Edition!

    Check out a sample material and table of contents of this book with Noga Alon.

    Click here for the frontcover front cover with a great picture of Uncle Paul

    Click here for a richgetricher new Probabilistic Lens on Preferential Attachment

    Click here for a phasetransition new chapter on the Erdos-Renyi Phase Transition, with particular emphasis on the Critical Window.

    Click for more, available from archives here

    Click here
    Hannah1 Hannah2 Hannah3 Hannah4 for our most beautiful and most attentive reader. She's a fox! And, in strong competition: johanna

    SPECIAL OFFER: Yes, your child (grandchild, niece, nephew, sibling) can have an Erdos Number of 1+\sqrt{-1}! Just send a photo of him/her absorbed in reading/playing/teething our book and I'll post it!

    Photos


    Photo by MA

    Click for family photos, or a photo of Paul Erdos.

    News Flashes!

    Alantha Newman and Aleksandar Nikolov have solved the Beck Three Permutaion Problem. I had offered a prize of 100 USD for its resolution. Here are Alantha and Sasha with the prize jpg and here is the arugment (in draft form!) itself pdf They found three permutations of 1,...,n=3^t so that for any coloring of 1,...,n with +1,-1 in one of the permutations there will be a prefix whose sum has absolute value at least ct. The conjecture had been that there was an absolute constant K so that for any three permutations of any size n there would be such a coloring where in each permutation all prefixes would have sum in [-K,+K], and so the conjecture was disproved.

    Robin Moser, a student of Emo Welzl (ETH), has given an algorithmic implementation of the Lovasz Local Lemma. A rough explanation:

    postscript LaTeX pdf Click for Moser's paper with Gabor Tardos .

    Nikhil Bansal (at TU Eindhoven (Holland)) has given an algorithmic implementation of my result that given any n sets on n points there is a 2-coloring with all discrepencies less than 6 \sqrt{n}. I had long conjectured that no algorithm would exist. Semidefinite Programming is a key ingredient. His preprint, Constructive Algorithms for Discrepancy Minimization, is available on arXiv. Click for Bansal's paper . or here
    postscript LaTeX pdf for my own rough explanation. Even more recently Shachar Lovett and Raghu Meka ArXiV link . have come up with another argument, using a restricted Brownian Motion.

    Cornell Probability School 2012

    [Note: Any further information will be added at this location.] I will be giving two lectures at the Cornell Probability School this July 16,18, 2012. (days still tentative). Much of the material is captured in the News Flashes above. The Moser-Tardos implementation of the Lovasz Local Lemma

    postscript LaTeX pdf Click for the Moser-Tardos paper

    My own "six standard deviation" result as done by Bansal Click for Bansal's paper . or
    postscript LaTeX pdf for my own rough explanation. Even more recently Shachar Lovett and Raghu Meka ArXiV link .

    The original proofs provide a good contrast and can be found in "The Probabilistic Method" by Noga Alon and myself. The Lovasz Local Lemma in Chapter 5 (numbering in the 3rd Edition, pretty much the same in previous editions) and Six Standard Deviations in Section 13.2.

    High School MathCamps

    This is a topic dear to my heart. I am former chair of an AMS committee that gives grants to High School Math Camps. Click for camp information or for information about donations to a worthy cause!

    Some Links

    My Wikipedia page
    Erdos Wikipedia page
    Danielle (daughter)'s website
    Erdos Number Project
    Budapest Semesters in Mathematics
    Combinatorialist Rogue Gallery

    [TOP]