Date/Time : 06 Mar 2026 14:30
Venue : IMS Executive Seminar Room (S17 Level 3)
Speaker : Max Welling, CuspAI & University of Amsterdam

Prof. Dr. Max Welling is a full professor and research chair in machine learning at the University of Amsterdam and a Merkin distinguished visiting professor at Caltech. He is co-founder and CTO of the startup CuspAI – a frontier AI company harnessing AI to unlock materials breakthroughs. He is also a member of the Royal Dutch Academy of Sciences, a fellow at the Canadian Institute for Advanced Research (CIFAR) and the European Lab for Learning and Intelligent Systems (ELLIS) where he served on the founding board. His previous appointments include Partner and VP at Microsoft Research, VP at Qualcomm Technologies, professor at UC Irvine.

Prof. Max Welling finished his PhD in theoretical high energy physics under supervision of Nobel laureate prof. Gerard ‘t Hooft. He then switched fields to focus on machine learning, first as a postdoc at Caltech under supervision of prof. Pietro Perona and then as postdoc under supervision of Nobel laureate prof. Geoffrey Hinton at UCL & U. Toronto.

Prof. Max Welling is a preeminent computer scientist and physicist with over 200,000 Google Scholar citations. His landmark contributions include inventing the Variational Autoencoder (VAE), a foundational architecture in generative AI, and Graph Convolutional Networks (GCNs), which revolutionized the processing of structured data such as molecules and social networks. Furthermore, he pioneered Stochastic Gradient Langevin Dynamics (SGLD) for Bayesian learning and has led the advancement of Equivariant Neural Networks, which integrate fundamental geometric symmetries into deep learning models.

Add to calendar :

Date/Time : 04 Mar 2026 14:00
Venue : S17 #05-11 (Seminar Room 5)
Speaker : Ralf Spatzier, University of Michigan

I will introduce the main ingredients for classifying hyperbolic  higher rank habelian actions.

Add to calendar :

Date/Time : 04 Mar 2026 15:00
Venue : S17 #04-06 (Seminar Room 1)
Speaker : Sven Leyffer, Argonne National Laboratory

Many design, planning and decision problems arising in engineering, sciences, finance, and statistics can be modeled as mixed-integer nonlinear optimization problems. A challenging class of mixed-integer problems are topology design problem, such as additive manufacturing and the design of cloaking devices. In topology optimization, the physical response of the design is modeled as partial-differential equations (PDEs) and the design is modeled with binary variables defined on each element of the discretization of the PDE. This approach results in mixed integer PDE-constrained optimization (MIPDECO) problem that combine the computational challenges of PDEs with the combinatorial challenges of a massive number of discrete variables. We show that despite their seemingly hopeless complexity, certain MIPDECOs can be solved efficiently (at a cost comparable to a few continuous PDE-constrained optimization solves). We discuss two classes of methods: rounding techniques that are asymptotically optimal, and randomized search techniques. We illustrate these solution techniques with examples from topology optimization.

Add to calendar :

Date/Time : 02 Mar 2026 15:00
Venue : S17 #05-11 (Seminar Room 5)
Speaker : Chenji Fu, University of Bonn

Eteve computed the Genestier-Lafforgue L-parameters for depth-zero representation for a equal-characteristic local field. I will explain how to adapt his method to the setting of Fargues-Scholze. We use certain ”smoothness” and ”properness” of the stack of shtuka to reduce the question to the local Hecke stack, where we can do an explicit computation using hyperbolic localizations following the work of ALWY on central sheaves in mixed characteristic.

Add to calendar :

Date/Time : 26 Feb 2026 15:00
Venue : S17 #04-05 (Seminar Room 2)
Speaker : Lingrui Ge, Peking University

We develop the dual based approach and introduce several new concepts and tools for the study of one-frequency quasiperiodic operators with analytic potentials. As applications we solve some long-standing arithmetic conjectures: the ten martini problem, sharp phase transition in frequency and absolute continuity of the integrated density of states, for the entire class of type I operators.

Add to calendar :

Date/Time : 25 Feb 2026 15:00
Venue : S17 #05-11 (Seminar Room 5)
Speaker : Chengyang Shao, Institut des Hautes Études Scientifiques

This is a mini-course with 3 lectures (1.5 hours each).   The time and venue are:
20 Feb (Friday) 3pm-4:30pm S16-03-06
23 Feb (Monday) 3pm-4:30pm S17-05-11
25 Feb (Wednesday) 3pm-4:30pm S17-05-11

This mini-course equips the audience with essential mathematical tools for the construction of stable and unstable invariant manifolds in hydrodynamical free boundary problems. In particular, it will be shown that the nonlinear instability for a moving capillary water jet follows precisely the Rayleigh-Plateau instability law. Being fully nonlinear and nonlocal in nature, such problems exhibit severe loss of derivatives and cannot be treated by standard methods. Therefore, it has remained largely open. Paradifferential calculus, however, provides a refined framework for analyzing nonlinearities, allowing them to be handled in a manner analogous to linear operators. It yields a unified framework addressing the construction of stable and unstable invariant manifolds for quasilinear PDEs or hydrodynamical free boundary problems.

The course consists of three lectures. The first lecture offers a concise introduction to classical partially hyperbolic dynamical systems and the Lyapunov-Perron method, followed by a review of its extensions to semilinear dispersive PDEs and Eulerian systems with fixed boundaries. The specific difficulties arising in quasilinear and free boundary problems are then discussed. The second lecture provides a brief introduction to the calculus of paradifferential operators, emphasizing its algebraic similarities with classical (pseudo)differential operators. In particular, it will be shown how this calculus applies to the analysis of Dirichlet-Neumann operator in free boundary problems. Finally, the third lecture demonstrates how these tools can be applied to construct stable and unstable invariant manifolds for hydrodynamical free boundary problems, thereby extending the classical Lyapunov-Perron method to this setting.

Some familiarity with the standard theory of hyperbolic PDEs, such as paraproduct decomposition and energy estimates, will be helpful. No prior background in dynamical systems is required.

Add to calendar :

Date/Time : 25 Feb 2026 15:00
Venue : S17 #04-06 (Seminar Room 1)
Speaker : Ralf Spatzier, University of Michigan

Symmetry and extremal properties have played a major role in mathematics for a long time. While I will start to discuss this in the context of some Riemannian geometry, I will emphasize more recent work on dynamical systems. Here symmetry expresses itself in terms of having non-trivially commuting maps or flows, or an action of some group with complicated relations. Case in point are actions of semisimple Lie groups, especially ones of higher rank, e.g. SL(n,R) with n at least 3. This is the so-called Zimmer program. I will discuss some recent highlights.

Add to calendar :

Date/Time : 24 Feb 2026 15:00
Venue : S17 #04-04 (Seminar Room 3)
Speaker : Jing An, Shanghai Jiaotong University

A classical problem in the study of the long-time behavior of advection–reaction–diffusion equations is to characterize the precise asymptotics of front locations. The strength of the advection term induces pushed-to-pulled front transitions, which can be captured by the so-called weighted Hopf–Cole transform. In this talk, I will revisit existing results on convergence to traveling waves in both the pulled and pushed regimes, and demonstrate how this transformation can be applied to the chemotaxis–FKPP system, an important nonlocal advection–reaction–diffusion model, to obtain refined front asymptotics for a wider range of parameters.

Add to calendar :

Date/Time : 23 Feb 2026 16:30
Venue : S17 #04-04 (Seminar Room 3)
Speaker : Nils Thuerey, Technical University of Munich

In this talk, I will present recent advances from the area of AI and deep learning for physics simulations. A key focus is the utilization of numerical solvers capable of providing gradient information, i.e. “differentiable simulators”. These solvers seamlessly integrate with deep learning algorithms, presenting numerous advantages in arising from AI-based components in solvers, particularly in the context of flow simulations. However, the regular, first-order quantities like gradients are not necessarily optimal for optimizations and learning tasks. I will outline alternative updates, specifically those arising from modified back propagation paths, and discuss their advantages and disadvantages. In addition, I will give an outlook towards generative modeling for PDEs.

Add to calendar :