Abstract:
Spatio-temporal modelling central to many modern applications, but the dimensionality of the domain of such data poses both conceptual and computational statistical challenges. These are exacerbated in a nonparametric setting, where questions related to asymptotic framework and invariance assumptions arise. The areas of functional data analysis and random field analysis have been attacking these problems separately, but are now increasingly coming into contact, especially when replication may be feasible. The workshop will bring together experts from both areas to explore recent trends and novel challenges.
08:00 • EPF Lausanne, Room GA 3 21
Peter K. Friz (Technische Universität Berlin)
Rough Analysis and Applications (ONLINE) abstract
Abstract:
We plan to discuss: the algebra of iterated integrals, signatures and expected signatures, rough paths and models; Kolmogorov type criteria; Schauder estimates; p-variation, abstract Riemann-Stieltjes integration (Sewing); rough integration and reconstruction; a paracontrolled view; rough differential equations and flows; rough transport equations; second order rough partial differential equations; rough volatility; a regularity structure view; large deviations and precise asymptotics for rough volatility.A good part of the lecture will follow F-Hairer (2014,2020).
10:15 • ETH Zentrum, Rämistrasse 101, Zürich, Building HG, Room
Artem Sapozhnikov (Universität Leipzig)
Title T.B.A.
11:00 • Universität Basel, Fachbereich Mathematik, Spiegelgasse 1, Seminar Room 05.002
Efthymios Sofos (University of Glasgow)
Title T.B.A.
15:15 • EPF Lausanne, Salle tba
Gérard Besson (Institut Fourier, Université de Grenoble)
CANCELLED: Title T.B.A.
15:45 • ETH Zentrum, Rämistrasse 101, Zürich, Building HG, Room G 43
Dr. Anton Ayzenberg (Higher School of Economics, Russia)
Multipolytopes, volume polynomials, duality algebras. abstract
Abstract:
(based on joint work w. Mikiya Masuda)We are interested in combinatorics of simplicial complexes. The famous g-theorem in this field gives necessary and sufficient conditions for the integral vector (f0, f1, f2,..., fn) to be the face vector of a convex simplicial sphere. The necessary condition in this theorem was proved by Richard Stanley by associating a certain algebra to a convex polytope (namely, the cohomology algebra of the corresponding projective toric variety) and exploiting its Lefschetz property. There are several elementary models of this polytope algebra, one of the simplest being the model constructed from the volume polynomial of a polytope.We study a wider class of combinatorial and geometrical objects: multipolytopes and multifans introduced by Hattori and Masuda. The notions of volume polynomial and its corresponding duality algebra can be extended to multipolytopes. This approach gives elementary explanations of some results on the combinatorics of triangulated manifolds.
15:45 • Location T.B.A.
Prof. em. Dr. Sara van de Geer (ETH Zürich, Switzerland)
Total variation regularization abstract
Abstract:
Let Y be a n-dimensional vector of independent observations with unknown meanf^0 := E Y. We consider the estimator f_D that solves the ``analysis problem" min_f {|| Y - f ||_2^2/ n + 2 lambda || D f ||_1 } ,where D is a given , m times n matrix and lambda>0 is a tuning parameter. An example for the matrix D is the (first order) difference operator(Df )_i= f_{i}- f_{i-1} , \\ i \\in [2:n]in which case || Df ||_1 = TV(f) is the (first order) total variation of the vector f. Other examples include higher order discrete derivatives, total variation on graphs and total variation in higher dimensions.Our aim is to show that the estimator f_D is adaptive. For example, when f^0 is apiecewise linear function, we show that the analysis estimator f_D, with D thesecond order differences operator, adapts to the number of kinks of f^0. As is the case with the Lasso, the theory for the analysis estimator f_D requires a form of "restricted eigenvalue" condition. We will show that this can beestablished using interpolating vectors. We will illustrate this (with drawings) for the various examples.
Onlline Seminar
Dr. Luca De Feo (IBM Research Zürich)
eSeminar: Faster Evaluation of Isogenies of Large Prime Degree abstract
Abstract:
(**This eSeminar will take place on Zoom, using same meeting details as previous seminars. If you do not have meeting details, please contact karan.khathuria@math.uzh.ch. This talk will also be live streamed at https://defeo.lu/tube/**) <BR> An isogeny is a non-zero morphism of elliptic curves. The isogeny evaluation problem asks, given a description of an isogeny ?:E->E\' and of a point P?E, to compute ?(P). It is a fundamental algorithmic problem in computational number theory, and has gained a fair amount interest thanks to the recent developments in isogeny-based cryptography. <BR> The "atomic" case for isogeny evaluation is that of isogenies of prime degree, on top of which algorithms for isogenies of any degree are easily constructed. For the prime degree case, the classic solution is based on Vélu\'s formulas, or any of their optimized variants. Vélu\'s formulas take as input the kernel of the isogeny, e.g., as a list of points, and output the isogeny as a pair of rational functions, which are then used to evaluate the isogeny at the point. This algorithm can be implemented in time linear in the isogeny degree, which is asymptotically optimal in general; however in the fundamental case where the kernel can be described by a single generator over the base field, one could hope to find a more efficient algorithm which sidesteps the computation of the rational functions. <BR> This is exactly what I will present in this talk: starting from a very simple idea, already used by Pollard, Strassen and Chudnovsky², among others, I will present a baby-step/giant-step algorithm that solves the isogeny evaluation problem in time and space proportional to the square root of the degree. I will explain why this is important for isogeny-based cryptography, and highlight several applications where the new algorithm produces practical speedups ranging from the unimpressive to the spectacular. <BR> This is joint work with D.J. Bernstein, A. Leroux and B. Smith, the preprint can be found at https://ia.cr/2020/341. The talk will be live streamed at
16:00 • Onlline Seminar
Tilmann Gneiting (HITS Heidelberg & Universität Karlsruhe)
CANCELLED: Strictly and non-strictly positive definite functions on spheres - revisited abstract
Abstract:
In classical papers, Isaac Schoenberg (1938, 1942) laid the ground for the study of isotropic positive definite functions on Euclidean spaces and spheres, respectively. Subsequently, positive definite functions on Euclidean spaces have been studied extensively in spatial statistics and related fields. In contrast, positive definite functions on spheres have only recently moved into focus, motivated by critical applied problems in the environmental and geophysical sciences that require correlation models on the surface of Planet Earth. I will review recent progress in the area, with an eye on a listing of 18 open problems posed in 2013.
17:15 • EPF Lausanne, Salle GA 3 21