Past

A topological quantum field theory (TQFT) is a functor from a category of bordisms to a category of vector spaces. Classifying low-dimensional TQFTs often involves considering presentations of bordism categories in terms of generators and relations. In this talk, we will introduce these concepts and outline a program for obtaining such presentations using Morse–Cerf theory.

Schwarzian Theory is a quantum field theory which has attracted a lot of attention in the physics literature in the context of two-dimensional quantum gravity, black holes and AdS/CFT correspondence. It is predicted to be universal and arise in many systems with emerging conformal symmetry, most notably in Sachdev--Ye--Kitaev random matrix model and Jackie--Teitelboim gravity.

In this talk we will discuss our recent progress on developing rigorous mathematical foundations of the Schwarzian Field Theory, including rigorous construction of the corresponding measure, calculation of both the partition function and a natural class of correlation functions, and a large deviation principle.

The stable uniqueness theorem for KK-theory asserts that a Cuntz-pair of *-homomorphisms between separable C*-algebras gives the zero element in KK if and only if the *-homomorphisms are stably homotopic through a unitary path, in a specific sense. This result, along with its group equivariant analogue, has been crucial in the classification theory of C*-algebras and C*-dynamics. In this talk, I will present a unitary tensor category analogue of the stable uniqueness theorem and explore its application to a duality in tensor category equivariant KK-theory. To make the talk approachable even for those unfamiliar with actions of unitary tensor categories or KK-theory, I will introduce the relevant definitions and concepts, drawing comparisons with the case of group actions. This is joint work with Kan Kitamura and Robert Neagu.

Simplicial volume is a homotopy invariant of manifolds introduced by Gromov to study their metric and rigidity properties. One of the strongest vanishing results for simplicial volume of closed manifolds is in presence of amenable covers with controlled multiplicity. I will discuss some conditions under which this result can be extended to manifolds with boundary. To this end, I will follow Gromov's original approach via the theory of multicomplexes, whose foundations have been recently laid down by Frigerio and Moraschini.

All atmospheric phenomena, from daily weather patterns to the global climate system, are invariably influenced by atmospheric flow. Despite its importance, its complex behaviour makes extracting informative features from its dynamics challenging. In this talk, I will present a network-based approach to explore relationships between different flow structures. Using three phenomenon- and model-independent methods, we will investigate coherence patterns, vortical interactions, and Lagrangian coherent structures in an idealised model of the Northern Hemisphere stratospheric polar vortex. I will argue that networks built from fluid data retain essential information about the system's dynamics, allowing us to reveal the underlying interaction patterns straightforwardly and offering a fresh perspective on atmospheric behaviour.

In a graph $H$ whose edges are coloured (not necessarily properly) a rainbow copy of a graph $G$ is a (not necessarily induced) subgraph of $H$ that is isomorphic to $G$ and whose edges are all coloured differently. In this talk I will explain why the problem of finding such rainbow copies is interesting, survey what we know, concentrating mainly on the case where $G$ is a Hamilton cycle, and then tell you a bit about a new result about finding rainbow Hamilton cycles resiliently in random graphs (which is joint work with Peter Allen and Liana Yepremyan).

Smooth generic representations of $GL_n$ over a $p$-adic field $F$, i.e. representations admitting a nondegenerate Whittaker model, are an important class of representations, for example in the setting of Rankin-Selberg integrals. However, in recent years there has been an increased interest in non-generic representations and their degenerate Whittaker models. By the theory of Bernstein-Zelevinsky derivatives we can associate to each smooth irreducible representation of $GL_n(F)$ an integer partition of $n$, which encodes the "degeneracy" of the representation. By using these "highest derivative partitions" we can define a stratification of the category of smooth complex representations and prove the surprising fact that all of the strata categories are equivalent to module categories over commutative rings. This is joint work with David Helm.

In the talk, I will start by recalling some basics of optimal transport and how it can be used to define Ricci curvature lower bounds for singular spaces, in a synthetic sense. Then, I will present some joint work with De Luca-De Ponti and Tomasiello,  where we show that some singular spaces,  naturally showing up in gravity compactifications (namely, Dp-branes),  enter the aforementioned setting of non-smooth spaces satisfying Ricci curvature lower bounds in a synthetic sense.  Time permitting, I will discuss some applications to the Kaluza-Klein spectrum.

On arbitrary Carnot groups, the only hypoelliptic Hodge-Laplacians on forms that have been introduced are 0-order pseudodifferential operators constructed using the Rumin complex.  However, to address questions where one needs sharp estimates, this 0-order operator is not suitable. Indeed, this is a rather difficult problem to tackle in full generality, the main issue being that the Rumin exterior differential is not homogeneous on arbitrary Carnot groups. In this talk, I will focus on the specific example of the free Carnot group of step 3 with 2 generators, where it is possible to introduce different hypoelliptic Hodge-Laplacians on forms. Such Laplacians can be used to obtain sharp div-curl type inequalities akin to those considered by Bourgain & Brezis and Lanzani & Stein for the de Rham complex, or their subelliptic counterparts obtained by Baldi, Franchi & Pansu for the Rumin complex on Heisenberg groups

A successful strategy to handle problems involving primes is to approximate them by a more 'simple' function. Two aspects need to be balanced. On the one hand, the approximant should be simple enough so that the considered problem can be solved for it. On the other hand, it needs to be close enough to the primes in order to make it an admissible to replacement. In this talk I will present how one can construct general approximants in the context of the Circle Method and will use this to give a different perspective on Goldbach type applications.

We show that linear reflection groups in the sense of Vinberg are often Zariski dense in PGL(n). Among the applications are examples of low-dimensional closed hyperbolic manifolds whose fundamental groups virtually embed as Zariski-dense subgroups of SL(n,Z), as well as some one-ended Zariski-dense subgroups of SL(n,Z) that are finitely generated but infinitely presented, for all sufficiently large n. This is joint work with Jacques Audibert, Gye-Seon Lee, and Ludovic Marquis.

We consider the problem of parametric and non-parametric statistical inference for systems of weakly interacting diffusions and of their mean field limit. We present several parametric inference methodologies, based on stochastic gradient descent in continuous time, spectral methods and the method of moments. We also show how one can perform fully nonparametric Bayesian inference for the mean field McKean-Vlasov PDE. The effect of non-uniqueness of stationary states of the mean field dynamics on the inference problem is elucidated.

During the 14th to the 16th centuries CE, a succession of Indian scholars, collectively referred to as the Kerala school, made remarkable contributions in the fields of mathematics and astronomy. Mādhava of Saṅgamagrāma, a gifted mathematician and astronomer, is considered the founder of this school, and is perhaps best known for discovering an infinite series for pi, among other achievements. Subsequently, Mādhava's lineage of disciples, consisting of illustrious names such as Parameśvara, Dāmodara, Nīlakaṇṭha, Jyeṣṭhadeva, Śaṅkara Vāriyar, Citrabhānu, Acyuta Piṣaraṭi etc., made numerous important contributions of their own in the fields of mathematics and astronomy. Later scholars of the Kerala school flourished up to the 19th century. This talk will provide a historical overview of the Kerala school and highlight its important contributions.

Electron cryomicroscopy (cryo-EM) is an imaging technique widely used in structural biology to determine the three-dimensional structure of biological molecules from noisy two-dimensional projections with unknown orientations. As the typical pipeline involves processing large amounts of data, efficient algorithms are crucial for fast and reliable results. The stochastic gradient descent (SGD) algorithm has been used to improve the speed of ab initio reconstruction, which results in a first, low-resolution estimation of the volume representing the molecule of interest, but has yet to be applied successfully in the high-resolution regime, where expectation-maximization algorithms achieve state-of-the-art results, at a high computational cost. 
In this work, we investigate the conditioning of the optimisation problem and show that the large condition number prevents the successful application of gradient descent-based methods at high resolution. 
Our results include a theoretical analysis of the condition number of the optimisation problem in a simplified setting where the individual projection directions are known, an algorithm based on computing a diagonal preconditioner using Hutchinson's diagonal estimator, and numerical experiments showing the improvement in the convergence speed when using the estimated preconditioner with SGD. The preconditioned SGD approach can potentially enable a simple and unified approach to ab initio reconstruction and high-resolution refinement with faster convergence speed and higher flexibility, and our results are a promising step in this direction.

Abstract: I will introduce and explain a new symmetry structure for type IIA string theory, called string^h. Using string^h I will explain  how some objects of stable homotopy theory relating to elliptic cohomology enter into type IIA string theory.

Persistent homology is infeasible to compute when a dataset is very large. Inspired by the bootstrapping method, Chazal et al. (2014) proposed a multiple subsampling approach to approximate the persistence landscape of a massive dataset. In this talk, I will present an extension of the multiple subsampling method to a broader class of vectorizations of persistence diagrams and to persistence diagrams directly. First, I will review the statistical foundation of the multiple subsampling approach as applied to persistence landscapes in Chazal et al. (2014). Next, I will talk about how this analysis extends to a class of vectorized persistence diagrams called Hölder continuous vectorizations. Finally, I will address the challenges in applying this method to raw persistence diagrams for two measures of centrality: the mean persistence measure and the Fréchet mean of persistence diagrams. I will demonstrate these methods through simulation results and applications in estimating data shapes. 

This week's Fridays@2 will feature a panel discussion on how to manage your time during your degree. The panel will share their thoughts and experiences in a Q&A session, discussing some of the practicalities of juggling lectures, the many ways to study independently and non-maths activities. 

The Bruhat-Tits building is a crucial combinatorial tool in the study of reductive p-adic groups and their representation theory. Given a p-adic group, its Bruhat-Tits building is a simplicial complex upon which it acts with remarkable properties. In this talk I will give an introduction to the Bruhat-Tits building by sketching its definition and going over some of its basic properties. I will then show the usefulness of the Bruhat-Tits by determining the maximal compact subgroups of a p-adic group up to conjugacy by using the Bruhat-Tits building.

At the heart of both cross-section calculations at the Large Hadron Collider and gravitational wave physics lie the evaluation of Feynman integrals. These integrals are meromorphic functions (or distributions) of the parameters on which they depend and understanding their analytic structure has been an ongoing quest for over 60 years. In this talk, I will demonstrate how these integrals fits within the framework of generalized hypergeometry by Gelfand, Kapranov, and Zelevinsky (GKZ). In this framework the singularities are simply calculated by the principal A-determinant and I will show that some Feynman integrals can be used to generate Cohen-Macaulay rings which greatly simplify their analysis. However, not every integral fits within the GKZ framework and I will show how the singularities of every Feynman integral can be calculated using Whitney stratifications.

Cells must reliably coordinate responses to noisy external stimuli for proper functionality whether deciding where to move or initiate a response to threats. In this talk I will present a perspective on such cellular decision making problems with extreme statistics. The central premise is that when a single stochastic process exhibits large variability (unreliable), the extrema of multiple processes has a remarkably tight distribution (reliable). In this talk I will present some background on extreme statistics followed by two applications. The first regards antigen discrimination - the recognition by the T cell receptor of foreign antigen. The second concerns directional sensing - the process in which cells acquire a direction to move towards a target. In both cases, we find that extreme statistics explains how cells can make accurate and rapid decisions, and importantly, before any steady state is reached.

This study introduces a novel suite of historical large language models (LLMs) pre-trained specifically for accounting and finance, utilising a diverse set of major textual resources. The models are unique in that they are year-specific, spanning from 2007 to 2023, effectively eliminating look-ahead bias, a limitation present in other LLMs. Empirical analysis reveals that, in trading, these specialised models outperform much larger models, including the state-of-the-art LLaMA 1, 2, and 3, which are approximately 50 times their size. The findings are further validated through a range of robustness checks, confirming the superior performance of these LLMs.