Journ ees GECKO / TERA2008 En l'honneur du 60 e anniversaire de Marc GIUSTI - 24-28 novembre 2008 Ecole polytechnique Amphith eˆatre Pierre ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Journées GECKO / TERA2008 En l’honneur du 60e anniversaire de Marc GIUSTI 24–28 novembre 2008 École polytechnique Amphithéâtre Pierre Faurre Programme 1
1 Programme LUNDI 24 NOVEMBRE 10h30 Accueil. Café 11h15 Luis Miguel PARDO Quelques réflexions incomplètes sur la résolution non-universelle des équations polynomiales 11h45 Mike STILLMAN Computing in intersection theory and intersection rings of flag bundles and Grassmannians 12h45-14h30 Déjeuner 14h30 Joris VAN DER HOEVEN On the art of guessing 15h30 Frédéric CHYZAK Products of Ordinary Differential Operators by Evaluation and Interpolation 16h–16h15 Pause café Coffee break 16h15 Laurent BUSÉ On the singularities of plane rational curves 16h45 Antonio CAFURE Bit complexity for polynomial solving over the integers 17h15 Guillermo MATERA Lower bounds for robust interpolation algorithms. 17h45 Pablo SOLERNÓ Sur l’indice et l’ordre de systèmes d’équations algebro-différentielles ordinaires 18h30 Cocktail WELCOME RECEPTION MARDI 25 NOVEMBRE 10h Joos HEINTZ Natural geometric objects and intrinsic complexity 11h Pause Café Coffee break 11h15 Teo MORA An FGLM-like algorithm for computing the radical of a 0-dim. ideal 11h45 Dave BAYER Graph colorings and toric algebra 12h45-14h30 Déjeuner 14h30 Bernd BANK Wavelet construction via algorithmic real algebraic geometry 15h30 Ian MORRISON New Gröbner Approaches to Hilbert Stability 16h–16h15 Pause café Coffee break 16h15 Bernard MOURRAIN Border bases, perturbations and walk on the Hilbert scheme 16h45 Teresa KRICK On the effective Nullstellensatz 17h15 Carlos BELTRÀN Integration in the space of singular maps : Where Geometry seems to link Real and Complex Analysis 17h45 Xavier DAHAN Lexicographic Gröbner bases and equiprojectable decompositions MERCREDI 26 NOVEMBRE 10h André GALLIGO Polynomials, Factorization and Randomness 11h Pause café Coffee break 11h15 Éric SCHOST Evaluation properties of invariant polynomials 11h45 Norbert A’CAMPO Computation of Monodromy 12h45 Déjeuner JEUDI 27 NOVEMBRE ATTENTION, le jeudi matin c’est en amphi Cauchy. 10h Grégoire LECERF Recent theoretical and practical advances in mutivariate polynomial factorization 11h Pause café Coffee break 11h15 Michel DEMAZURE Réécriture algébrique sans soustraction 11h45 Bruno SALVY Combinatorial Newton Iteration and Efficient Random Generation 2
12h15 Alin BOSTAN The full counting function of Gessel walks is algebraic 12h45-14h30 Déjeuner 14h30 Joris VAN DER HOEVEN Mathemagix I : General introduction 15h Grégoire LECERF Mathemagix II : C++ architecture, fast algebraic operations 15h30–15h45 Pause café Coffee break 15h45 Bernard MOURRAIN Mathemagix III : polynomial systems, geometry 16h Joris VAN DER HOEVEN Mathemagix IV : analysis, singularities 16h15 Daouda Niang DIATTA Mathemagix V : topology of curves and surfaces 16h45–17h Pause café Coffee break 17h GECKO BUSINESS MEETING 20h30 DÎNER Restaurant “Bel Canto”, 72, quai de l’Hôtel de Ville, 75004 Paris. VENDREDI 28 NOVEMBRE 10h Mike SHUB Homotopy Methods for solving systems of polynomial equations 11h Pause café Coffee break 11h15 Jean-Claude YAKOUBSOHN Calcul certifié du pgcd approché 11h45 Paola BOITO and Jean-Pierre DEDIEU Condition Geodesics in Matrix Spaces 12h45-14h30 Déjeuner THE END 3
2 Résumés Computation of Monodromy Norbert A’CAMPO, Université de Bâle 26/XI 11h45 The TQFT-monodromy for singularities of plane curves is a strong invariant. It is an interesting problem to compute this monodromy. We construct inside of the Milnor fiber of an isolated hypersurface singularity a stratified subset. For the case of functions of three complex variables this subset is build with strata of real dimension two. The visualization of this subset would give new understanding of monodromy. Wavelet construction via algorithmic real algebraic geometry Bernd BANK & Lutz LEHMANN, Humboldt Universität zu Berlin Joint work with Marc Giusti, Joos Heintz, Luis Miguel Pardo 25/XI 14h30 As a result of the TERA–project a new type, highly efficient probabilistic algorithm for the solution of systems of polynomial equations was developed and implemented for the complex case. The geometry of polar varieties allows to extend this algorithm to a method that finds real solutions of systems of polynomial equations. In order to test this method special emphasis was placed on the fact that example problems are of real-life and practical importance. In the talk we report on the application of the TERA–method to real polynomial equation systems solving basic for the design of fast wavelet transforms. The wavelet transforms we have in mind should reflect the practical important properties of symmetry and orthogonality. These requirements are expressible by a finite number of real parameters satisfying a finite system of polynomial equations. If these equations have a real solution at all, the solution set can be finite or a variety of positive dimension. Examples with real solution sets of positive dimension have the advantage that one can search for optimal solutions in the sense that the wavelets have additionally desired analytic properties. It turns out that the TERA–algorithm performes very well with this task and is able to solve larger systems than the best known commercial polynomial solvers. Graph colorings and toric algebra Dave BAYER, Barnard College, Columbia University 25/XI 11h45 The close link between integer programming and toric algebra relates maximal lattice-free polytopes to the semigroup structure of an associated toric ring. Monomial modules allow us to think of this toric ring as an infinite-periodic monomial ideal, whose injective hull determines the maximal lattice-free polytopes. Given a graph, there is a natural toric ring whose injective hull determines the colorability of the graph. 4
Integration in the space of singular maps : Where Geometry seems to link Real and Complex Analysis Carlos BELTRÀN, Dept. of Mathematics, University of Toronto 25/XI 17h15 Complex varieties in the affine space have very special properties. Some of them are : 1) they are minimal manifolds, 2) when intersecting the variety with a ball of increasing radius, the Hausdorff measure grows at least as in the linear case, 3) the volume growth of a tube around the variety is controlled by the Hausdorff measure of the variety. Indeed, these three facts are intimately related, as 1 ⇒ 2 ⇒ 3. In this talk, I will show that property 3 is also satisfied by a very particular class of real varieties : the set of rank-deficient matrices, and I will discuss the geometric resemblance of the real and complex situations, which naturally poses the following question : Are these real varieties minimal manifolds ? One motivation for this study is related to Numerical Analysis : The volume growth of the tube around these varieties describes the probability that a rank-deficient matrix is “easy to solve” in floating point computations, in the sense that the condition number of the problem is small. I will show very precise bounds for this numerical problem. The full counting function of Gessel walks is algebraic Alin BOSTAN, INRIA Paris-Rocquencourt 27/XI 12h15 The aim of the talk is to show how a difficult combinatorial problem has been recently solved using an experimental-mathematics approach combined with (rather involved) computer algebra techniques. More precisely, let f (n, i, j) denote the number of lattice walks in the quarter plane which start at the origin, end at the point (i, j), and consist of n unit steps going either west, south-west, east, or north-east. In the early nineties, Ira Gessel conjectured that the sequence of excursions f (n, 0, 0) is holonomic. We will present the computer-driven discovery and proof of the following generalization, X obtained in August 2008 together with Manuel Kauers : the full generating series F (t, x, y) = f (n, i, j)xi y j tn is an algebraic function. i,j,n≥0 On the singularities of plane rational curves Laurent BUSÉ, INRIA Sophia Antipolis 24/XI 16h15 Given a rational plane algebraic curve C, we will show that some informations on the singularities of C can be recovered from a (birational) parameterization of C. In particular, we will describe some explicit adjoint linear systems on C and will interpret them as equations of a certain Rees algebra. We will end by giving an extension of Abhyankar’s Taylor-resultant for an arbitrary rational plane curve. 5
Bit complexity for polynomial solving over the integers Antonio CAFURE, Universidad Nacional de General Sarmiento 24/XI 16h45 When solving polynomial systems over the integers, to avoid the so-called intermediate expression swell, the input system is reduced modulo a suitable prime, and from a resolution modulo this prime, by p-adic lifting, a resolution of the input system is computed. Applying different results obtained by the TERA group in the preceding years we give an upper bound on the height of the primes that enable us to perform such a modular reduction. We use this upper bound in order to obtain an upper bound on the bit complexity of solving polynomial systems over the integers. Products of Ordinary Differential Operators by Evaluation and Interpolation Frédéric CHYZAK, INRIA Paris - Rocquencourt 24/XI 15h30 It is known that multiplication of linear differential operators over ground fields of characteristic zero can be reduced to a constant number of matrix products. We give a new algorithm by evaluation and interpolation which is faster than the previously-known one by a constant factor, and prove that in characteristic zero, multiplication of differential operators and of matrices are computationally equivalent problems. In positive characteristic, we show that differential operators can be multiplied in nearly optimal time. Theoretical results are validated by intensive experiments. Lexicographic Gröbner bases and equiprojectable decompositions Xavier DAHAN, Universite de Kyushu, Fukuoka, Japon 25/XI 17h45 In 1985, D. Lazard gave a complete description of the structure of lexicographic Gröbner bases of zero- dimensional and radical ideals of bivariate polynomials. He somehow linked this structure to the primary decomposition of the ideal. We show that a more relevant decomposition is the so-called ”equiprojectable” one, not the primary. Then, relying on this decomposition, the generalization of this ”structure theorem”, to ideals of multivariate polynomials is not difficult. Two consequences can de deduced : size on the coefficients of lexicographic Gröbner bases, and a geometric-numerical criterion for choosing a ”lucky prime” (term borrowed to E. Arnold, JSC 2003), in the context of modular computation of Gröbner bases. Condition Geodesics in Matrix Spaces Paola BOITO and Jean-Pierre DEDIEU, Institut de Mathématiques, Université Paul Sabatier, Toulouse 28/XI 11h45 6
The condition metric for spaces of polynomial systems has been introduced and studied in a series of papers by Beltrán, Dedieu, Malajovich and Shub. The interest of this metric comes from the fact that the associated geodesics avoid ill-conditioned problems and are a useful tool to improve classical complexity bounds for Bézout’s theorem. The linear case is examined here : Using nonsmooth nonconvex analysis techniques, we study the behaviour of condition geodesics in the space of full rank, real or complex rectangular matrices. The main results include an existence theorem for the boundary problem, a differential inclusion for such geodesics based on Clarke’s generalized gradients, regularity properties and a detailed description of a few particular cases (diagonal and unitary matrices). Moreover, we study condition geodesics from a numerical viewpoint and develop an effective algorithm that allows to compute geodesics with given endpoints and helps to illustrate theoretical results and formulate new conjectures. Réécriture algébrique sans soustraction Michel DEMAZURE, professeur des universités retraité 27/XI 11h15 On interprète un article de Fiore et Leinster en termes de réécriture. Polynomials, Factorization and Randomness André GALLIGO, université de Nice 26/XI 10h00 The development of algorithms for polynomial factorization leads me to approximate computations and randomness. I will present 3 aspects : generic change of coordinates, early detection, distribution laws, and their use for multivariate polynomial factorization, Natural geometric objects and intrinsic complexity Joos HEINTZ, Universidad de Buenos Aires and CONICET, Argentina and Universidad de Cantabria, Santander, Spain 25/XI 10h00 My friendship and scientific collaboration with Marc Giusti started 1982 during the meeting Algorith- mique et Calcul Formel in Limoges. This led in the following years to a intensive interchange and cross fertilization of semantical concepts from algebraic geometry and syntactical views from theoretical computer science, mainly complexity theory, but also data structures and types and programming paradigms. Toge- ther with our coworkers from the international TERA (Turbo Evaluation and Rapid Algorithms) group we succeded finally to establish an agenda in symbolic and seminumeric elimination theory, which someone may like and someother not, but which is difficult to ignore. An outcome of this synergy was the development of the Kronecker Software package by G. Lecerf and collaborators and a couple of applications, one of them to image processing. In my talk I shall try to give an account of the ideas and questions which guided us during 26 years, the mathematical difficulties we met and the solutions which we finally found. In this context Marcs intuition that 7
in effective elimination theory computations should evolve along “natural” geometric objets which determine in their turn the size and complexity of the intermediate results, plays a crucial role. This view is reflected in the quest for the intrinsic complexity of elimination problems. On the effective Nullstellensatz Teresa KRICK, Université de Buenos Aires 25/XI 16h45 The subject of the talk will be the Nullstellensatz, a cornerstone in Algebraic Geometry, in its effective aspects. I will mention a method, developed by Marc Giusti and others in the 90’s, which produces the coefficients g1 , . . . , gs in an expression 1 = g1 f1 + · · · + gs fs (when possible) and allowed to obtain the best effective arithmetic results up to now. I also intend to introduce a new recent proof by Z. Jelonek which makes use only of elementary and natural tools of classical algebraic geometry and which will hopefully allow to improve those arithmetic results. Recent theoretical and practical advances in multivariate polynomial factorization Grégoire LECERF, université de Versailles 27/XI 10h00 This talk will survey the recent advances within the factorization of multivariate polynomials. We will show algorithms that are essentially subquadratic in time for almost all the tasks, with a focus on the methods developped during the period covered by the ANR GECKO. Lower bounds for robust interpolation algorithms Guillermo MATERA, Universidad Nacional de General Sarmiento 24/XI 17h15 In this talk we discuss lower bounds on the complexity of robust algorithms for solving families of interpolation problems. Our notion of robustness models the behavior of all known universal methods for solving families of interpolation problems avoiding unnecessary branchings and allowing the solution of certain limit problems. We first show that a robust algorithm solving a family of Lagrange interpolation problems with N nodes encoded by a Zariski open subset of the space CN of nodes has a cost which is at least linear in N , showing thus that standard interpolation methods are essentially optimal. Then we consider families of interpolation problems with singularities. In particular, we consider the family of problems which consists of interpolating a polynomial given by a straight-line program of length L from its value in a correct-test sequence. We show that any robust algorithm solving such a family of problems requires a number of arithmetic operations which is exponential in L. Joint work with Nardo Giménez, Joos Heintz and Pablo Solernó. 8
An FGLM-like algorithm for computing the radical of a 0-dim. ideal Teo MORA, Universita’ di Genova 25/XI 11h15 Mainly motivated by Auzinger–Stetter Algorithm and Cardinal’s Conjecture, in MEGA-92 Traverso at al. proposed the notion of Gröbner representation of a 0-dimensional ideal J ⊂ Q = K[Z1 , . . . , Zr ] which is the assignement of – a K-linearly independent = {q1 , . . . , qs } ⊂ Q, n set q 2 o (h) – the set M = M(q) := alj ∈ Ks , 1 ≤ h ≤ r of r square matrices (l) – s3 values γij∈K which satisfy 1. Q/J = ∼ Span (q), K P (h) 2. Zh ql ≡ j alj qj (mod J) for each l, j, h, 1 ≤ l, j ≤ s, 1 ≤ h ≤ r, P (l) 3. qi qj ≡ l γij ql (mod J) for each l, j, h, 1 ≤ i, j, l ≤ s. In connection with Rouillier’s Rational Univartiate Representation (actually with Kronecker Parame- trization), at the ICPSS conference in honour of Daniel Lazard, Arai and Muritzugu posed the following Problem. — Given a 0-dimensional ideal J ⊂ Q = K[Z1 , . . . , Zr ] via a Gröbner representation, compute √ Pr a Gröbner representation of J and a separating linear form Y = i=1 ai Zi for its roots. An algorithm (based on repeated application of Buchberger Algorithm) for computing both the radical and a separating linear form for its roots of a given zero-dimensional ideal was proposed in 1987 by Gianni. Inspired by old results by Giusti–Heintz and Alonso–Raimondo on decomposition, I reformulate here an FGLM-like linear algebra adaptation of Gianni’s Algorithm thus solving the given problem. New Gröbner Approaches to Hilbert Stability Ian MORRISON, Fordham University 25/XI 15h30 Recently, Gieseker’s GIT construction of M g as a quotient of the ν-canonical Hilbert scheme for ν ≥ 5 has found new applications in Hassett’s log minimal model program for M g . These models arise as quotients for ν < 5 and are moduli spaces whose boundary curves satisfy variants of Deligne-Mumford stability. Pinning down these variants, and constructing the models involve understanding, respectively, Hilbert instability and stability. To date, it has sufficed to polarize the Hilbert schemes by taking degree m 0—when the answer is independent of m—and the techniques have involved asymptotic estimates that apply only for +large. The next stages require making such constructions for fixed small values of m. I will review this work, focusing on new techniques for checking Hilbert instability and stability for small m using Gröbner techniques. The first, work of Hassett, Hyeon and Lee, uses Castelnuovo-Mumford regularity to read off stability with respect to a fixed 1-parameter subgroup from low degree pieces of the associated Gröbner basis and streamlines many instability calculations. The second, work in progress with Swinarski, is a first attack on the harder problem of checking low-m stability for curves with many automorphisms by calculations of state polytopes or Gröbner fans. 9
Border bases, perturbations and walk on the Hilbert scheme Bernard MOURRAIN, INRIA Méditerranée 25/XI 16h15 Solving polynomial equations with approximate coefficients is ubiquitous in many applications. It is also a challenge from an effective algebraic geometry point of view. To tackle this issue, border basis methods have been introduced recently. Compared to Grobner basis computation, they yield representations of the quotient algebra, which are more stable from a numerical point of view. In this presentation, we try to analysis more precisely this assertion. We first recall the main properties of these border bases, how they can be characterised, how they can be computed, how the syzygies are generated and give some examples of Border basis computations. Our motivation is to be able to improve the numerical quality of a quotient representation, after an approximate computation of a borderbasis at a given precision. We will describe an explicit Newton-type iteration for this purpose and give effective criteria to check flatness or the stability of a deformation. The connection with Hilbert scheme of points will be exploited. Quelques réflexions incomplètes sur la résolution non-universelle des équations polynomiales Luis-Miguel PARDO, Universidad de Cantabria 24/11 11h15 Dans cet exposé, on reviendra sur les méthodes éfficaces de résolution Non–Universelle des équations polynomiales : l’origine de la question, les résultats positifs et, surtout, les questions ouvertes. L’exposé contiendra plus de questions que de réponses. Combinatorial Newton Iteration and Efficient Random Generation Bruno SALVY, Inria 27/XI 11h45 We recall the principle of Boltzmann samplers for large combinatorial structures. The construction of these samplers relies on so-called ”oracles” that compute numerical values of generating series. We show that such an oracle is realized by a simple Newton iteration. The important point is that using the origin as a starting point always converges to the desired solution. The proof relies on lifting the Newton iteration as an iteration on classes of combinatorial structures. Homotopy Methods for solving systems of polynomial equations Mike SHUB, Mathematics Department, University of Toronto 28/XI 10h00 10
We consider the complexity of homotopy methods for solving systems of polynomial equations, the condition metric, and the structure of the solution variety. This is joint work with Carlos Beltran, Jean- Pierre Dedieu, Gregorio Malajovich, Luis Miguel Pardo and Steve Smale. Evaluation properties of invariant polynomials X. Dahan, Éric SCHOST, University of Western Ontario, J. Wu 26/XI 11h15 A polynomial invariant under the action of a finite group can be rewritten using generators of the invariant ring. We investigate the complexity aspects of this rewriting process ; we show that evaluation techniques enable one to reach a polynomial cost. Sur l’indice et l’ordre de systèmes d’équations algebro-différentielles ordinaires Pablo SOLERNÓ, Universidad de Buenos Aires 24/11 17h45 Nous étudions les notions d’indice et d’ordre dans le cas des systèmes quasi-réguliers, quelques estimations ”à la Jacobi” et des conséquences quantitatives. Computing in intersection theory and intersection rings of flag bundles and Grassmannians Mike STILLMAN, Cornell University 24/XI 11h45 The main idea of the talk is to describe some joint work with Dan Grayson, involving intersection rings of flag bundles, and a package for Macaulay2, “Schubert 2”, which is under development. Schubert 2 is roughly based on the Maple package Schubert, written by Katz and Stromme. We start by reviewing intersection theory and describing Grothendieck’s theorem for the intersection ring of flag bundles. We then show that, with respect to specific monomial orders, their defining ideals have simple to describe initial ideals, even over the integers. The resulting Groebner bases allow for fast computation in the intersection rings of flag bundles and Grassmannians. We then show some examples of computations in enumerative geometry using Schubert2 that take advantage of these methods. 11
On the art of guessing Joris VAN DER HOEVEN, département de Mathématiques, université Paris Sud 24/XI 14h30 Current computer algebra systems usually work in a shell mode : the user asks a question and the system hopefully gives an answer. To what extent would it be possible to discover additionnal mathematical structure in the user’s problem in an automatic fashion ? For instance, if we encounter 2.094395102393195492308428922 in the output of a numerical computation, the system might suggest its replacement by 2π/3. Even though guessing such addional relations was not explicitly requested by the user, it may lead to interesting insights and does not necessarily lead to a big increase of the overall computation time. In our talk, we will address several classical guessing algorithms and present a few new ones. We will start with rational number and function recovery, the LLL-algorithm and Pade-Hermite approximation. We will next turn our attention to guessing possible asymptotic expansions for sequences and possible relations at singularities of analytic functions. If time permits it, we will also discuss some guessing techniques for symbolic expressions. Calcul certifié du pgcd approché Jean-Claude YAKOUBSOHN, université Paul Sabatier, Toulouse III En collaboration avec Guillaume Chèze, André Galligo et Bernard Mourrain 28/XI 11h15 Le calcul du pgcd approché peut se ramener à la résolution d’un problème de minimisation d’une fraction rationnelle de deux variables. Cette formulation a été donnée par Karmarkar et Lakshman dans leur papier “On approximate gcds of univariate polynomials”, J. Symbolic Comput., 1998, 26, 653-666. Nous proposons un calcul certifié de ce problème de minimisation. Nous illustrons cette certification dans les cas où les polynômes sont représentés dans la base des monômes et dans la base de Bernstein. 12
3 Participants Norbert A’CAMPO, université de Bâle Bernd BANK, Humboldt Universität zu Berlin Dave BAYER, Barnard College, Columbia University Carlos BELTRÀN, Dept. of Mathematics, University of Toronto Jeremy BERTHOMIEU, LIX Alexandre BENOIT, INRIA-MSR Manuel BODIRSKY, LIX Paola BOITO, Institut de Mathématiques, Université Paul Sabatier, Toulouse Alin BOSTAN, INRIA Paris-Rocquencourt Jérôme BRACHAT, INRIA Sophia-Antipolis Laurent BUSÉ, INRIA Sophia Antipolis Antonio CAFURE, Universidad Nacional de General Sarmiento Jacques CALMET, Universität Karlsruhe (TH) Frédéric CHYZAK, INRIA Paris - Rocquencourt Jean-Pierre DEDIEU, Institut de Mathématiques, Université Paul Sabatier, Toulouse Xavier DAHAN, Universite de Kyushu, Fukuoka, Japon Daouda Niang DIATTA, Université de Limoges Michel DEMAZURE, professeur des universités retraité Clémence DURVYE, université de Versailles Michel FLIESS, LIX, Ecole polytechnique & Projet ALIEN, INRIA Anne FREDET, IUT de Saint-Denis André GALLIGO, université de Nice Marc GIUSTI, LIX Joos HEINTZ, Universidad de Buenos Aires and CONICET, Argentina and Universidad de Cantabria, Santander, Spain Pierre-Vincent KOSELEFF, UPMC-PARIS SALSA-INRIA Teresa KRICK, Université de Buenos Aires Romain LEBRETON, LIX Bernard MALGRANGE, Académie des Sciences Grégoire LECERF, université de Versailles Guillermo MATERA, Universidad Nacional de General Sarmiento Marc MEZZAROBBA, INRIA Rocquencourt (Équipe Algorithms) Jean MOULIN OLLAGNIER, LIX, École polytechnique et Université Paris XII Teo MORA, Universita’ di Genova Guillermo MORENO-SOCIAS, université de Versailles Ian MORRISON, Fordham University Guillaune MOROZ, INRIA/LIP6 Bernard MOURRAIN, INRIA Méditerranée François OLLIVIER, LIX, École polytechnique Luis Miguel PARDO, Universidad de Cantabria Michel PETITOT, LIFL, université Lille I Adrien POTEAUX, INRIA Galaad - Universite de Nice Sophia Antipolis Bruno SALVY, Inria Eric SCHOST, University of Western Ontario Alexandre SEDOGLAVIC, LIFL, USTL Mike SHUB, Mathematics Department, University of Toronto Pablo SOLERNÓ, Universidad de Buenos Aires Michèle SORIA, LIP6, université Paris 6 Mike STILLMAN, Cornell University 13
Philippe TREBUCHET, Paris 6 Annick VALIBOUZE, Université Pierre et marie Curie (UPMC) Giuseppe VALLA, Universita di Genova Joris VAN DER HOEVEN, département de Mathématiques, université Paris Sud Jacques-Arthur WEIL, XLIM, Université de Limoges Jean-Claude YAKOUBSON, université Paul Sabatier, Toulouse III 14
You can also read