Erion Morina is going to talk about: 'Learning PDE-based models from data: An analysis-driven perspective on identifiability, consistency, and interpretability'

Learning governing equations from data is a central problem in scientific machine learning. Given noisy and incomplete observations of a physical process, the goal is to recover the underlying law. This is often approached by fitting a neural network or a dictionary of candidate terms to the observed dynamics. A good fit, however, does not determine identifiability of the law, its approximability by what is computed, physical consistency, or interpretability as a formula. These depend largely on how the learning problem is posed.

This talk formulates the learning task as a regularized inverse problem in function space and studies how these properties can be established. Over a parameterized class of candidate laws, one minimizes a model residual, a data misfit, and a regularizer, with the class and the regularizer as the design choices. With a sufficiently expressive class and suitable regularization, one obtains a regularization-based notion of identifiability, under which the regularization-minimizing law consistent with the data is unique. As the approximation scale grows and the regularization parameters are chosen accordingly, minimizers of the parameterized problem converge to that law. For classes carrying the structural constraints of the underlying physical model, the learned models are physically consistent and well-posed by construction at every finite approximation scale. For symbolic networks built from rational building blocks, the recovered law is a readable formula. The emphasis of this talk is analytical, and first numerical experiments illustrate the recovery in practice. 

Overall, the results show that identifiability, consistency, and interpretability are not competing objectives, but can be unified in one analysis-driven framework.

Aperiodic (quasicrystalline) tilings, such as Penrose's tiling, can be built up from (for example) kites and darts, squares and triangles, rhombi or shield-shaped tiles and can have a variety of different symmetries. However, almost all quasicrystals occurring in soft matter are of the dodecagonal (12-fold rotation symmetry) type, and many can be described in terms of square and equilateral triangular tiles. Here, we explore what contributes to the thermodynamic stability of soft-matter quasicrystals, both in two dimensions and in three, and how the details of how soft-matter particles interact leads to different kinds of aperiodic tilings. Although dodecagonal quasicrystals are the most common, this work points to how more general (beyond dodecagonal) quasicrystals can be designed in soft matter.

Lorenzo Pareschi is going to talk about; 'Bridging high-order numerics and machine learning for kinetic plasma simulation'

Reliable uncertainty quantification is a central challenge in kinetic plasma simulation, where high dimensionality, multiple physical scales, and sensitivity to uncertain inputs make repeated high-fidelity computations prohibitively expensive. This is particularly relevant in fusion-oriented applications, for which accurate predictions require sophisticated numerical solvers but direct sampling is often out of reach.

In this talk, I will present a multifidelity framework for the Vlasov–Poisson–Landau system designed to combine, rather than replace, high-order numerical simulation with machine learning. At the high-fidelity level, asymptotic-preserving and structure-aware solvers provide accurate kinetic descriptions across different regimes. These are coupled with reduced plasma models and tensor neural surrogates constructed through a micro–macro decomposition, so that the dominant physical structure is treated analytically and numerically, while learning is used only for the lower-complexity kinetic correction.  The resulting hierarchy produces inexpensive low-fidelity samples that remain strongly correlated with the high-fidelity kinetic solution.  When used as control variates, these models yield substantial variance reduction and computational savings while retaining the high-order solver as the reference description.

Beyond the specific plasma application, the main message is that classical numerical analysis and machine learning need not be competing approaches. High-order solvers can provide structure, reliability, and asymptotic consistency, while learned models provide efficient approximations that can be exploited within rigorous multifidelity estimators. This interaction offers a general route toward trustworthy machine learning for computational science.
 

Calibrated fibrations are expected to play an important role in exceptional holonomy, much as special Lagrangian fibrations do in the SYZ picture for Calabi--Yau manifolds. Singular fibres are expected to be essential, and a natural question is whether they are forced by the geometry. In his PhD thesis, Baraglia showed that coassociative fibrations of compact full-holonomy G_2-manifolds must have singular fibres. The analogous problem for Cayley fibrations remained open for many years and turns out to be substantially more difficult, with its resolution relying on deep results from 4-manifold topology.

The proof takes some unexpected twists and turns, involving a Diophantine equation arising from the Spin(7)-structure, families Seiberg--Witten theory and parametrized homotopy theory. After a short crash course on Spin(7) geometry, I will explain how these pieces fit together. This is joint work with Jianfeng Lin and Viktor Majewski.

to follow

Conlon and Gowers in 2016 described a general approach to proving sparse random analogues of extremal results in combinatorics, such as bounding the minimum and maximum number of triangles in any subgraph of G(n,p) with a given number of edges. The general part of this approach is a functional-analytic statement which, given a sparse setting, constructs a 'dense model'. However there is a condition which must be shown to hold with high probability to apply the dense model theorem. In Conlon and Gowers' work, there is a technical difficulty with the probabilistic part which leads to a rather involved proof, which applies only in a restricted setting (for example, they can handle triangles but not triangles with a pendant edge), and with quite poor bounds on 'high probability'.

Around the same time Schacht, with a very different method, was able to prove a related result, which is applicable in a much more general setting and which has optimal bounds on 'high probability': but Schacht's result does not provide sharp lower bounds on the number of triangles, and does not provide any upper bounds. What it does do is identify the number of edges in a subgraph of G(n,p) which guarantee that triangles will appear; subsequently the 'hypergraph container method' provides another approach to proving this kind of result, but again does not provide sharp lower bounds or any upper bounds.

We revisit Conlon and Gowers' approach, and show how to avoid their technical problem, giving a simpler proof of their counting result which applies in the general setting and with optimal probability bounds. As a corollary, we prove the 'Counting KLR' theorem of Conlon, Gowers, Samotij and Schacht, but for general hypergraphs and with optimal probability bounds. This is joint work with Julia Boettcher, Joanna Lada and Domenico Mergoni.

to follow

TBA

Internal gravity waves (IGWs) are oscillatory modes of motion that exist in the interior of a stably stratified fluid with gravity as the restoring force. In Earth's oceans, these waves play a fundamental role in the transport and mixing of energy, momentum, heat, carbon and nutrients. Much like ocean surface waves, IGWs become unstable and break when their amplitude is large and density interfaces become steeply inclined. This talk will explore the evolution and fate of such waves in two different scenarios.

I will first focus on IGWs that propagate into horizontally sheared currents, explaining how the fate of these waves depends both on their geometry and the strength of the shear flow, and why the total energy dissipated by ensuing turbulence is unexpectedly large. Second, I will discuss the dynamics of internal bores – long, large-amplitude solitary waves that propagate along sharp density interfaces. I will outline the mathematical links between internal bores and gravity currents and, by studying observations off the coast of California, demonstrate their relevance to understanding transport and dissipation on the continental shelf.

Ruth Misener will talk about: 'Optimizing over graphs: Challenges, Formulations, and Applications'

Applications involving optimization over graphs include molecular design, graph neural network verification, neural architecture search, etc. This talk discusses formulating graph spaces using mixed-integer optimization and incorporating application-specific constraints. We discuss computational challenges with these mixed-integer optimization formulations and zoom in on the practical implications for these applications. We mention what has been done (by both ourselves and others) and what other research still needs to be done.

Co-authors: Shiqiang Zhang, Yilin Xie, Christopher Hojny, Juan Campos, Jixiang Qing, Christian Feldmann, David Walz, Frederik Sandfort, Miriam Mathea, Calvin Tsay

This talk is hosted by Rutherford Appleton Laboratory, Harwell Campus

TBA

TBA

Hypertoric varieties are conical symplectic singularities equipped with a hamiltonian action of a torus of maximal possible dimension. They behave analogously to toric varieties in many ways, with the hamiltonian torus replacing the dense torus. Examples include: (crepant partial resolutions of) rational surface singularities of type A, the cotangent bundle of projective space, and Nakajima quiver varieties with the `all 1s' dimension vector.

Arbo and Proudfoot conjectured a combinatorial classification of hypertoric varieties by zonotopal tilings. In this talk I will discuss a proof of the conjecture.

TBA

to follow

Any countable group G comes equipped with a canonical dynamical system, namely its conjugation action on its space of subgroups Sub(G). This 0-dimensional compact space is a central object in measured and geometric group theory, especially because it supports invariant and stationary random subgroups.

In general, describing this space is hard. In 2024, Carderi, Gaboriau, Le Maître, and Stalder initiated the study of the space of subgroups of non-amenable Baumslag-Solitar groups BS(m,n). They provided an explicit description of the perfect kernel of Sub(BS(m,n)). This is the largest closed subspace without isolated points, i.e. the space that remains after performing successive derivations that remove the isolated points.

In this talk, I will provide a complete classification of the spaces Sub(BS(m,n)) up to homeomorphism. I will prove that there exist exactly four homeomorphism types among the non-amenable ones. This relies on a detailed study of the Cantor-Bendixson erasing process, which depends on the arithmetical properties of the parameters m,n. This is based on a joint work with Damien Gaboriau, François Le Maître, and Yves Stalder.

TBA

TBA

TBA 

Recently, data-driven methods based on path signatures have gained prominence in mathematical finance. They rely on universal approximation theorems stating that continuous functionals on path space can be approximated uniformly on compact sets by linear functionals of the signature. In financial applications, this has led to the use of Stratonovich-signatures, although Itô integration is often the natural modeling framework.

In this talk, we establish a universality result for signatures of non-geometric rough paths. By augmenting the path with its rough path bracket, we obtain a quasi-shuffle structure that provides the algebraic basis for universality. For continuous semimartingales, this yields a universal approximation property for Itô signatures.

This talk is based on joint work with A. P. Kwossek and D. J. Prömel.

TBA

Associate Professor Zhanzing Zhu is going to talk about; 'Understanding Transformers: Token Distinguishability, Learnability, and Safety Alignment'

The underlying mechanism of Large Language Models’ (LLMs) reasoning capabilities remains a big puzzle. This talk will introduce several theoretical works from our group, which attempt to answer some fundamental questions about the Transformer in LLMs:

1. Does softmax attention still work for a very long context (e.g. millions) to distinguish each token for learning? How can we scale the logit to produce a sharp attention distribution?
2. Why can Transformers acquire reasoning ability through Chain-of-Thought (CoT) with fine-tuning? What is the difference between supervised fine-tuning (SFT) and reinforcement learning (RL) for fine-tuning LLMs?
3. Safety alignment in LLMs is fragile in part because it is often shallow: fine-tuning mainly reshapes the model’s behavior near the first few output tokens. Why does the fine-tuning lead to such fragile safety alignment? How can we robustify it?

Z/2 harmonic 1-forms are generalizations of harmonic 1-forms that allow topological twisting around a subspace of codimension 2. These objects were introduced by Taubes to compactify the moduli spaces of solutions to generalized Seiberg-Witten equations, and they show up in many other gauge theoretical problems.

Donaldson showed there is a deformation theory for so-called non-degenerate Z/2-harmonic 1-forms. In this presentation we study the deformations of the remaining degenerate solutions. For a natural class of degenerate examples, we prove that after a suitable perturbation of the ambient Riemannian metric, the form can be deformed to a nearby non-degenerate Z/2-harmonic 1-form.

TBA

to follow

to follow

TBA

TBA