Search by speaker

Filter institution








Filter content






Tue, 01. Jul at 13:15
Humboldt-Universi...
A Quasi-Optimal Space-Time FEM With Local Mesh Refinements For Parabolic Problems
Abstract. We present a space-time finite element method for the heat equation that computes quasi-optimal approximations with respect to natural norms while incorporating local mesh refinements in space-time. The discretized problem is solved with a conjugate gradient method with a (nearly) optimal preconditioner. This is joint work with Lars Diening and Rob Stevenson.
Wed, 02. Jul at 10:00
Weierstrass-Insti...
Multiple Tests for Mean Functions of Functional Data
Abstract. Functional data analysis is becoming increasingly popular to study data from real-valued random functions. Nevertheless, there is a lack of multiple testing procedures for such data. These are particularly important in factorial designs for comparing different groups or inferring factor effects. We propose a new class of testing procedures for arbitrary linear hypotheses in general factorial designs with functional data. Our methods allow global as well as multiple inference of both univariate and multivariate mean functions without assuming particular error distributions or homoscedasticity. That is, we allow for different structures of the covariance functions between groups. We analyse the (joint) asymptotic behaviour of suitable test statistics and propose a resampling approach to approximate the limit distributions. The resulting global and multiple testing procedures are asymptotically valid under weak conditions and applicable in general functional MANOVA settings. We evaluate their small-sample performance in extensive simulations and finally illustrate their applicability by analysing a data set.
Wed, 02. Jul at 10:00
HVP 11 a, R.313, ...
Wed, 02. Jul at 11:30
Weierstrass Lectu...
Two edges suffice: the planar lattice two-neighbor graph percolates
Abstract. The $k$-neighbor graph is a directed percolation model on the hypercubic lattice $\mathbb{Z}^d$ in which each vertex independently picks exactly $k$ of its $2d$ nearest neighbors at random, and we open directed edges towards those. We prove that the $2$-neighbor graph percolates on $\mathbb{Z}^2$, i.e., that the origin is connected to infinity with positive probability. The proof rests on duality, an exploration algorithm, a comparison to i.i.d. bond percolation under constraints as well as enhancement arguments. As a byproduct, we show that i.i.d. bond percolation with forbidden local patterns has a strictly larger percolation threshold than $1/2$. Additionally, our main result provides further evidence that, in low dimensions, less variability is beneficial for percolation.
Wed, 02. Jul at 11:30
online
Opinion dynamics with stochastic leaders
Abstract
Wed, 02. Jul at 13:15
Room: 3.007 John ...
Polarized K3 surfaces of genus 19
Abstract. We discuss the structure of the moduli space of polarized K3 surfaces of genus 19 using a connection to the moduli space of higher rank vector bundles on a curve of genus 2.
Wed, 02. Jul at 15:00
Humboldt-Universi...
Lower energy bounds for the Landau-deGennes model
Wed, 02. Jul at 15:15
WIAS, Erhard-Schm...
From large deviations around porous media, to PDEs with irregular coefficients, to gradient flow structures
Abstract
Wed, 02. Jul at 15:30
Humboldt-Universi...
Introduction to a new class of BDF schemes
Wed, 02. Jul at 16:00
Humboldt-Universi...
Low-regularity exponential integrators for the 'good' Boussinesq equation with rough solutions
Wed, 02. Jul at 16:00
Wed, 02. Jul at 16:00
Wed, 02. Jul at 16:30
Rudower Chaussee ...
Global Kuranishi charts in symplectic Gromov-Witten theory, part 2
Abstract
Thu, 03. Jul at 09:00
1.023 (BMS Room, ...
Dynamics and disc filling methods
Abstract. First, we give a gentle introduction to contact geometry through the exploration of dynamical questions leading to the Weinstein conjecture (now Taubes' theorem). The true goal of the talk is to explain how holomorphic curve theory can help study dynamical questions, in this case the existence of closed orbits. After an introduction to basic features of holomorphic curves (closed, with boundary and with punctures) we use them to explain Hofer's proof of the Weinstein conjecture for 3-manifolds that are overtwisted or have π_2.
Thu, 03. Jul at 14:15
Humboldt-Universi...
Thu, 03. Jul at 14:15
Humboldt-Universi...
Residual-based a posteriori error analysis for symmetric mixed FEM in elasticity
Abstract. The development of mixed finite element methods for linear elasticity with strongly imposed symmetry has been a long standing problem until the beginning of this century. Surprisingly for a mixed method, nodal stress degrees of freedom are necessary in order to fulfill the strong symmetry. This interesting mixed finite element also poses some difficulties for the derivation of residual based a posteriori error estimators. In a first attempt we make use of the residual a posteriori error estimator techniques for weakly symmetric stresses introducing an auxiliary approximation of the skew-symmetric gradient via a postprocessing. The second version then makes fully use of the imposed symmetry of the stress approximations utilising integration by parts twice and a suitable decomposition into tangential-tangential and normal-normal parts, similarly to the residual a posteriori error analysis for plate problems.
Thu, 03. Jul at 15:15
Rudower Chaussee ...
Finsler geodesics and finite-strain plasticity
Abstract. The theory of energetic rate-independent systems is an elegant way to describe nonlinear systems in mechanics and other fields. One particular advantage is that it yields a natural time discretization that consists of a sequence of minimization problems. Unfortunately, in many interesting cases the objective functional is only given implicitly as the solution of a second minimization problem for a curve length in the state space. Therefore, its evaluation and obtaining derivatives can be very costly. Instead, we present a transformation based on the Finsler exponential map that turns the second minimization problem into an initial-value-problem for a second-order ODE. Solutions of this can be found much cheaper numerically, or may even be available in closed form. We show examples of this construction, and how to use it to obtain fast and robust Proximal Newton solvers for finite-strain elastoplasticity.
Thu, 03. Jul at 16:15
TU Berlin, Instit...
How to Learn from Data in Stochastic Control Problems – An Approach Based on Statistics
Abstract. While theoretical solutions to many stochastic control problems are well understood, their practicality often suffers from the assumption of known dynamics of the underlying stochastic process, which raises the statistical challenge of developing purely data-driven controls. In this talk, we discuss how stochastic control and statistics can be brought together, which we study for various classical control problems with underlying one- and multi-dimensional diffusions and jump processes. The dilemma between exploration and exploitation plays an essential role in the considerations. We find exact sublinear-order convergence rates for the regret and compare the results numerically with those of deep Q-learning algorithms. The talk is based on: - Nonparametric learning for impulse control problems (with C. Strauch), Annals of Applied Probability 33 (2), 1369 - 1387, 2023 - Learning to reflect: A unifying approach for data-driven stochastic control strategies (with C. Strauch, Lukas Trottner), Bernoulli 30 (3) 2074 - 2101, August 2024 - Data-driven rules for multidimensional reflection problems (with Asbjørn Holk Thomsen, Lukas Trottner), SIAM/ASA Journal on Uncertainty Quantification, Volume 12, Issue 4, 2024 - Data-driven optimal stopping: A pure exploration analysis (with Niklas Dexheimer, Claudia Strauch), 2024 (preprint on arXiv)
Thu, 03. Jul at 17:15
TU Berlin, Instit...
An Efficient On-Policy Deep Learning Framework for Stochastic Optimal Control
Abstract. We present a novel on-policy algorithm for solving stochastic optimal control (SOC) problems. By leveraging the Girsanov theorem, our method directly computes on-policy gradients of the SOC objective without expensive back propagation through stochastic differential equations or adjoint problem solutions. This approach significantly accelerates the optimization of neural network control policies while scaling efficiently to high-dimensional problems and long time horizons. We evaluate our method on classical SOC benchmarks as well as applications to sampling from unnormalized distributions via Schrödinger-Föllmer processes and fine-tuning pre-trained diffusion models. Experimental results demonstrate substantial improvements in both computational speed and memory efficiency compared to existing approaches. Joint work with Mengjian Hua and Eric Vanden-Eijnden
Fri, 04. Jul at 14:15
TU (C 130)
Kovalevskaya Lecture
Sun, 06. Jul at 15:30
Rudower Chaussee ...
Positive sectional curvature and Ricci flow
Abstract
Tue, 08. Jul at 13:15
Humboldt-Universi...
Tue, 08. Jul at 14:00
1.023 (BMS Room, ...
Wed, 09. Jul at 10:00
HVP 11 a, R.313, ...
Wed, 09. Jul at 11:30
Weierstrass Lectu...
Wed, 09. Jul at 13:15
Room: 3.007 John ...
p-adic zeta function and Hodge theory
Abstract. In 1988, Igusa proposed a mysterious conjecture, in the spirit of Weil's conjecture, which predicts certain arithmetic quantities of a given polynomial (poles of the p-adic zeta function) are in fact topological/geometric, i.e. they must induce roots of the Bernstein-Sato polynomial and monodromy eigenvalues on the cohomology of Milnor fibers. This conjecture is widely open, but for hyperplane arrangements, Budur-Mustațǎ-Teitler proposed the n/d-conjecture in 2009 and showed that it will imply Igusa's conjecture in this case. In this talk, I will report a recent work of mine with Dougal Davis, where we prove the n/d-conjecture, finishing the previous work of Saito, Walther, Budur, Yuzvinsky, Veys, Bath, Shi-Zuo and many others. The proof uses Hodge theory in an essential way, especially Schmid's nilpotent orbit theorem and its consequences.
Wed, 09. Jul at 14:00
WIAS, Erhard-Schm...
The three limits of the hydrostatic approximation
Abstract
Wed, 09. Jul at 15:30
WIAS, Erhard-Schm...
A mediocre two-component variant of the famous result on equilibration in scalar Fokker-Planck equations
Abstract
Wed, 09. Jul at 16:30
EN 058
Is the squared inverse of the distance between kissing polytopes always an integer?
Abstract. A lattice $(d,k)$-polytope is the convex hull of a set of points in dimension $d$ whose coordinates are integers ranging from $0$ to $k$. We investigate the smallest possible distance between two disjoint lattice $(d,k)$-polytopes. A pair of such polytopes are called kissing polytopes. This question arises in various contexts where the minimal distance between such polytopes appears in complexity bounds for optimization algorithms. We provide nearly matching lower and upper bounds for this distance and propose an algebraic model. Our formulation yields explicit formulas in dimensions $2$ and $3$, and allows for the computation of previously intractable values. We also discuss related results, such as the Alon–Vu bounds for flat simplices — that is, the minimum distance between a vertex of a lattice $(d,1)$-simplex and the affine space spanned by the remaining vertices. Finally, we observe that all the known squared distances between kissing polytopes are inverses of integers, and we ask whether this observation holds in general. Based on joint-work with Shmuel Onn (Technion), Sebastian Pokutta (Zuse Institute Berlin and TU Berlin), and Lionel Pournin (Université Paris 13).
Thu, 10. Jul at 10:00
WIAS, Erhard-Schm...
Thu, 10. Jul at 13:00
Fri, 11. Jul at 14:15
FU (T9)
On optimality conditions for nonsmooth functions
Mon, 14. Jul at 13:00
Rudower Chaussee ...
Tue, 15. Jul at 11:15
1.023 (BMS Room, ...
Wed, 16. Jul at 10:00
Weierstrass-Insti...
Wed, 16. Jul at 11:30
online
Data-Adaptive Discretization of Inverse Problems
Abstract
Wed, 16. Jul at 16:00
Thu, 17. Jul at 10:00
WIAS, Erhard-Schm...
Thu, 17. Jul at 15:00
Lorentzian polynomials and the incidence geometry of tropical linear spaces
Abstract. Tropical linear spaces are complicated. Even the most elementary questions about their incidence geometry can be hard. I will give some answers to three types of such questions that are especially important to other recent advancements. The central object is the moduli space of all codimension-1 tropical linear subspaces of a given tropical linear space. The structure of this moduli space is closely related to the structure of spaces of Lorentzian polynomials. I will show how convexity results about tropical linear spaces can be derived and can be used to derive convexity results about Lorentzian polynomials.