Neural networks have become increasingly effective for approximating solutions to partial differential equations (PDEs). This talk presents three advances that improve both accuracy and computational efficiency. First, I introduce an augmented Lagrangian formulation of the physics-informed loss that strengthens constraint enforcement and improves accuracy near domain boundaries. Second, I develop efficient architectures based on hypernetworks and graph neural networks that learn PDE solution operators with markedly small model sizes. Finally, I describe Neural-Galerkin schemes with low rank approximations for operator learning, which achieves a favorable accuracy-efficiency trade-off.

TBA

A (positive definite and integral) quadratic form $f$ is called irrecoverable (from its subforms) if there is a quadratic form $F$ that represents all proper subforms except for $f$ itself, and such a quadratic form $F$ is called an isolation of $f$. In this talk, we present recent advances on irrecoverable quadratic forms and discuss their possible generalizations.

(The is a PhD student reading seminar to be given by Mr. Jaehong Kim.)

We present a dynamic data structure that maintains a tree decomposition of width at most 9k+8 of a dynamic graph with treewidth at most k, which is updated by edge insertions and deletions. The amortized update time of our data structure is $2^{O(k)} \log n$, where n is the number of vertices. The data structure also supports maintaining any “dynamic programming scheme” on the tree decomposition, providing, for example, a dynamic version of Courcelle’s theorem with $O_k(\log n)$ amortized update time; the $O_k(⋅)$ notation hides factors that depend on k. This improves upon a result of Korhonen, Majewski, Nadara, Pilipczuk, and Sokołowski [FOCS 2023], who gave a similar data structure but with amortized update time $2^{k^{O(1)}} n^{o(1)}$. Furthermore, our data structure is arguably simpler. Our main novel idea is to maintain a tree decomposition that is “downwards well-linked”, which allows us to implement local rotations and analysis similar to those for splay trees. This talk is based on arXiv:2504.02790.

TBD

We present a new cryptomorphic definition of orthogonal matroids with coefficients using Grassmann-Plücker functions. The equivalence is motivated by Cayley’s identities expressing principal and almost-principal minors of a skew-symmetric matrix in terms of its pfaffians. As a corollary of the new cryptomorphism, we deduce that each component of the orthogonal Grassmannian is parameterized by certain part of the Plücker coordinates. This is joint work with Changxin Ding.

This is a reading seminar of a graduate student, following the Fields medal work of Daniel Quillen on the foundation of the higher algebraic K-theory.

In this presentation, we discuss recent existence results for nonlinear diffusion equations with a divergence-type drift term, which are broadly applicable to various reaction-diffusion equations, including Keller-Segel models. We focus on identifying appropriate functional spaces for the drift, guided by the nonlinear diffusion and initial data. Using techniques from the theory of Wasserstein spaces, we construct weak solutions and establish their regularity properties.

Associated to a group action on a bifoliated plane, satisfying some reasonable conditions, one can associate a combinatorial object known as a veering triangulation. Since their introduction by Agol (in a very different setting), these triangulations have recently played an interesting role in studying pseudo-Anosov flows, the structure of fibered 3-manifolds, algorithmic properties of mapping class groups, and fixed points of surface homeomorphisms, to name just a few (from my own biased perspective). This talk will be an overview of these applications, starting with the most basic properties from the initial bifoliated plane.

In this talk, we will discuss Leray-Hopf solutions to the incompressible Navier-Stokes equations with vanishing viscosity. We explore important features of turbulence, focusing around the anomalous energy dissipation phenomenon. As a related result, I will present a recent result proving that for two-dimensional fluids, assuming that the initial vorticity is merely a Radon measure with nonnegative singular part, there is no anomalous energy dissipation. Our proof draws on several key observations from the work of J. Delort (1991) on constructing global weak solutions to the Euler equation. We will also discuss possible extensions to the viscous SQG equation in the context of Hamiltonian conservation and existence of weak solutions for a rough initial data.

Given a $k$-uniform hypergraph $F$, its Turán density $\pi(F)$ is the infimum over all $d\in [0,1]$ such that any $n$-vertex $k$-uniform hypergraph $H$ with at least $d\binom{n}{k}+o(n^k)$ edges contains a copy of $F$. While Turán densities are generally well understood for graphs ($k=2$), the problem becomes notoriously difficult for $k\geq 3$, even for small hypergraphs. We study two well-known variants of this Turán problem for hypergraphs: first, under minimum codegree conditions and, second, with a quasirandom edge distribution. Each variant defines a distinct extremal parameter, generalising the classical Turán density. Here we present recent results in both settings, with a particular emphasis on the case of hypergraphs where every link is itself quasirandom. Our results include exact solutions for key hypergraphs and general results about the behaviour of the Turán density functions.

For many variant kinetic equations, we choose an appropriate approx- imatation equations. Also, this approximation equations are solvable more easier than the original equations and it retains the expected a priori bounds. Then, we use the variant compactness theorem to pass to the limit in the sense of distributions in the approximation equations. In the kinetic theory, this compactness theorem is called the averaing lemma, that is, the averaging in velocity improves regularity in the space and time variables. For this PDE seminar, we study the basic averaging lemma. In other words, we investigate the basic properties of the free transport operator ∂t + v · ∇x.

In this talk, I will try to explain how the essence of the Weierstrass representation formula and the Bjorling representation formula for minimal surfaces in $E^3$ can be suitably applied to zero/constant mean curvature surfaces in the three-dimensional spaceforms in the Lorentz-Minkowski four-space.

TBA

In this talk, we introduce the concept of t-core partitions. We discuss the generating function and modularity, along with some results and applications of t-core partitions. Recent results on simultaneous core partitions will also be presented. Toward the end of the talk, we introduce numerical semigroups and explore connections between numerical semigroups (or numerical sets) and partitions. Additionally, we present some open problems related to these topics.

We study the partial dimensional semi-classical Weyl’s laws, describing the quantum subband structures for two-dimensional electron gases (2DEGs). As a simple application, we derive lowest free energy states for the subband models describing non-interacting 2DEGs.

We study the partial dimensional semi-classical Weyl’s laws, describing the quantum subband structures for two-dimensional electron gases (2DEGs). As a simple application, we derive lowest free energy states for the subband models describing non-interacting 2DEGs.

We consider a continuous model of graphs, introduced by Dearing and Francis in 1974, where each edge of G to be a unit interval, giving rise to an infinite metric space that contains not only the vertices of G but all points on all edges of G. Several standard graph problems can be defined and studied on continuous graphs, yielding many surprising algorithmic results and combinatorial connections. The motivation can be exemplified by the well-known Independent Set problem on graphs. Given a graph G, we want to place k facilities that are pairwise at least a distance 2 edge lengths apart. In some applications, such as when the underlying graph represents a street network, it is reasonable to allow placing a facility not only at a crossroad but also somewhere within a street, that is, not only at a vertex but also at any point on an edge between two vertices. This motivates the study of the Independent Set problem on continuous graphs. In such a setting, for example, the problem corresponds to requiring pairwise distance r=2 for the placed facilities. However, we may also study the problem where we fix r to any positive integer, rational, or even irrational number. Other problems studied in the continuous model include Dominating Set, TSP, and Coloring. In the first part of this talk, we will give a general overview of research on continuous graphs and computational problems in this setting. In the second part, we explore, as part of recent work, a coloring problem on continuous graphs akin to the well-known Hadwiger-Nelson Problem. Based on joint work with Fabian Frei, Florian Hörsch, Tom Janßen, Stefan Lendl, Dániel Marx, Prahlad Narasimhan, and Gerhard Woeginger.

The Lyapunov-Schmidt reduction is a powerful tool to solve PDEs. This method reduces the equations, which are essentially infinite-dimensional, to finite-dimensional ones. In this talk, we illustrate the reduction by showing the existence of a positive solution to the singularly perturbed problem in for positive smooth and appropriate . To show the existence, we first construct an -dimensional surface of approximate solutions. Then, we reduce the problem onto that surface by the Lyapunov-Schmidt reduction. The key to the reduction is proving the invertibility of a certain operator, which in turn, is proved by a certain uniqueness result. After the reduction, we end the proof by solving the equation on the -dimensional surface.