Past

Come along to hear from several PhD students and PostDocs about their research. There will also be a Q&A about doing a Master's/PhD and a chance to mingle with postgraduate students. 

Speakers include:

• Shaked Bader, DPhil Student in Geometric Group Theory, 
• Eoin Hurley, PostDoc in Combinatorics, 
• Patricia Lamirande, DPhil Student in Mathematical Biology

Several PhD students from the department will give short 5 minute talks on their research. This is also targeted at undergraduates interested in doing PhDs .

Quantum computers are becoming a reality and current generations of machines are already well beyond the 50-qubit frontier. However, hardware imperfections still overwhelm these devices and it is generally believed the fault-tolerant, error-corrected systems will not be within reach in the near term: a single logical qubit needs to be encoded into potentially thousands of physical qubits which is prohibitive.
 
Due to limited resources, in the near term, hybrid quantum-classical protocols are the most promising candidates for achieving early quantum advantage but these need to resort to quantum error mitigation techniques. I will explain the basic concepts and introduce hybrid quantum-classical protocols are the most promising candidates for achieving early quantum advantage. These have the potential to solve real-world problems---including optimisation or ground-state search---but they suffer from a large number of circuit repetitions required to extract information from the quantum state. I will detail a range of application areas of randomised quantum circuits, such as quantum algorithms, classical shadows, and quantum error mitigation introducing recent results that help lower the barrier for practical quantum advantage.

 

: With motivation from string compactifications, I will present work on the use of machine learning methods for the computation of geometric and topological properties of Calabi-Yau manifolds.

The Seventeenth Brooke Benjamin Lecture 2024

The Elusive Singularity

I will describe the open problems of singularity formation in incompressible fluids. I will discuss a list of related models, some results, and some more open problems.

Date: Monday, 11 November 2024 

Time: 5pm GMT

Location: Lecture Theatre 1, Mathematical Institute 

Speaker: Professor Peter Constantin        

More information about The Brooke Benjamin Lecture.

The Elusive Singularity

I will describe the open problems of singularity formation in incompressible fluids. I will discuss a list of related models, some results, and some more open problems.

Date: Monday, 11 November 2024 

Time: 5pm GMT

Location: Lecture Theatre 1, Mathematical Institute 

Speaker: Professor Peter Constantin            

Peter Constantin is the John von Neumann Professor of Mathematics and Applied and Computational Mathematics at Princeton University. Peter Constantin received his B.A and M.A. summa cum laude from the University of Bucharest, Faculty of Mathematics and Mechanics. He obtained his Ph.D. from The Hebrew University of Jerusalem under the direction of Shmuel Agmon.

Constantin’s work is focused on the analysis of PDE and nonlocal models arising in statistical and nonlinear physics. Constantin worked on scattering for Schr¨odinger operators, on finite dimensional aspects of the dynamics of Navier-Stokes equations, on blow up for models of Euler equations. He introduced active scalars, and, with Jean-Claude Saut, local smoothing for general dispersive PDE. Constantin worked on singularity formation in fluid interfaces, on turbulence shell models, on upper bounds for turbulent transport, on the inviscid limit, on stochastic representation of Navier-Stokes equations, on the Onsager conjecture. He worked on critical nonlocal dissipative equations, on complex fluids, and on ionic diffusion in fluids.

Constantin has advised thirteen graduate students in mathematics, and served in the committee of seven graduate students in physics. He mentored twenty-five postdoctoral associates. 

Constantin served as Chair of the Mathematics Department of the University of Chicago and as the Director of the Program in Applied and Computational Mathematics at Princeton University.

Constantin is a Fellow of the Institute of Physics, a SIAM Fellow, and Inaugural Fellow of the American Mathematical Society, a Fellow of the American Academy of Arts and Sciences and a member of the National Academy of Sciences

In general, the classification of finitely generated subgroups of a given group is intractable. Restricting to two-generator subgroups in a geometric setting is an exception. For example, a two-generator subgroup of a right-angled Artin group is either free or free abelian. Jaco and Shalen proved that a two-generator subgroup of the fundamental group of an orientable atoroidal irreducible 3-manifold is either free, free-abelian, or finite-index. In this talk I will present recent work proving a similar classification theorem for two generator mapping-torus groups of free group endomorphisms: every two generator subgroup is either free or conjugate to a sub-mapping-torus group. As an application we obtain an analog of the Jaco-Shalen result for free-by-cyclic groups with fully irreducible atoroidal monodromy. While the statement is algebraic, the proof technique uses the topology of finite graphs, a la Stallings. This is joint work with Naomi Andrew, Ilya Kapovich, and Stefano Vidussi.
 

I will present a notion of spin structure on a perfect complex in characteristic zero, generalizing the classical notion for an (algebraic) vector bundle. For a complex $E$ on $X$ with an oriented quadratic structure one obtains an associated ${\mathbb Z}/2{\mathbb Z}$-gerbe over X which obstructs the existence of a spin structure on $E$. This situation arises naturally on moduli spaces of coherent sheaves on Calabi-Yau fourfolds. Using spin structures as orientation data, we construct a categorical refinement of a K-theory class constructed by Oh-Thomas on such moduli spaces.

Self-predictive learning (aka non-contrastive learning) has become an increasingly important paradigm for representation learning. Self-predictive learning is simple yet effective: it learns without contrastive examples yet extracts useful representations through a self-predicitve objective. A common myth with self-predictive learning is that the optimization objective itself yields trivial representations as globally optimal solutions, yet practical implementations can produce meaningful solutions. 

We reconcile the theory-practice gap by studying the learning dynamics of self-predictive learning. Our analysis is based on analyzing a non-linear ODE system that sheds light on why despite a seemingly problematic optimization objective, self-predictive learning does not collapse, which echoes with important implementation "tricks" in practice. Our results also show that in a linear setup, self-predictive learning can be understood as gradient based PCA or SVD on the data matrix, hinting at meaningful representations to be captured through the learning process.

This talk is based on our ICML 2023 paper "Understanding self-predictive learning for reinforcement learning".

Celestial holography posits that the 4D S-matrix may be calculated holographically by a 2D conformal field theory. However, bulk translation invariance forces low-point massless celestial amplitudes to be distributional, which is an unusual property for a 2D CFT. In this talk, I show that translation-invariant MHV gluon amplitudes can be extracted from smooth 'leaf' amplitudes, where a bulk interaction vertex is integrated only over a hyperbolic slice of spacetime. After describing gluon leaf amplitudes' soft and collinear limits, I will show that MHV leaf amplitudes can be generated by a simple 2D system of free fermions and the semiclassical limit of Liouville theory, showing that translation-invariant, distributional amplitudes can be obtained from smooth correlation functions. An important step is showing that, in the semiclassical limit of Liouville theory, correlation functions of light operators are given by contact AdS Witten diagrams. This talk is based on a series of papers with Atul Sharma, Andrew Strominger, and Tianli Wang [2312.07820, 2402.04150,2403.18896]. 

North meets South is a tradition founded by and for early-career researchers. One speaker from the North of the Andrew Wiles Building and one speaker from the South each present an idea from their work in an accessible yet intriguing way. 

North Wing

Speaker: Paul-Hermann Balduf
Title: Statistics of Feynman integral
Abstract: In quantum field theory, one way to compute predictions for physical observables is perturbation theory, which means that the sought-after quantity is expressed as a formal power series in some coupling parameter. The coefficients of the power series are Feynman integrals, which are, in general, very complicated functions of the masses and momenta involved in the physical process. However, there is also a complementary difficulty: A higher orders, millions of distinct Feynman integrals contribute to the same series coefficient.

My talk concerns the statistical properties of Feynman integrals, specifically for phi^4 theory in 4 dimensions. I will demonstrate that the Feynman integrals under consideration follow a fairly regular distribution which is almost unchanged for higher orders in perturbation theory. The value of a given Feynman integral is correlated with many properties of the underlying Feynman graph, which can be used for efficient importance sampling of Feynman integrals. Based on 2305.13506 and 2403.16217.

South Wing

Speaker: Marc Suñé 
Title: Extreme mechanics of thin elastic objects
Abstract: Exceptionally hard --- or soft -- materials, materials that are active and response to different stimuli, elastic objects that undergo large deformations; the advances in the recent decades in robotics, 3D printing and, more broadly, in materials engineering, have created a new world of opportunities to test the (extreme) mechanics of solids. 

In this colloquium I will focus on the elastic instabilities of slender objects. In particular, I will discuss the transverse actuation of a stretched elastic sheet. This problem is a peculiar example of buckling under tension and it has a vast potential scope of applications, from understanding the mechanics of graphene and cell tissues, to the engineering of meta-materials.

In this talk, I will give an overview of recent joint work on Topological Data Analysis (TDA). The first one is an application of TDA to quantify porosity in pathological bone tissue. The second is an extension of persistent homology to directed simplicial complexes. Lastly, we present an evaluation of the persistent Laplacian in machine learning tasks. This is joint work with Ysanne Pritchard, Aikta Sharma, Claire Clarkin, Helen Ogden, and Sumeet Mahajan; David Mendez; and Tom Davies and Zhengchao Wang, respectively.
 

We will explore the connection between Celestial and Euclidean Anti-de Sitter (EAdS) holography in the massive scalar case. Specifically, exploiting the so-called hyperbolic foliation of Minkowski space-time, we will show that each contribution to massive Celestial correlators can be reformulated as a linear combination of contributions to corresponding massive Witten correlators in EAdS. This result will be demonstrated explicitly both for contact diagrams and for the four-point particle exchange diagram, and it extends to all orders in perturbation theory by leveraging the bootstrapping properties of the Celestial CFT (CCFT).  Within this framework, the Kantorovic-Lebedev transform plays a central role. This transform will allow us to make broader considerations regarding non-perturbative properties of a CCFT.

This week's Fridays@2 will be a panel discussion focusing on what it is like to pursue a research degree. The panel will share their thoughts and experiences in a question-and-answer session, discussing some of the practicalities of being a postgraduate student, and where a research degree might lead afterwards.

Introduction to flat space holography in three dimensions and Carrollian CFT2, with selected results on correlation functions, thermal entropy, entanglement entropy and an outlook to Bondi news in 3d.

Ramification theory serves the dual purpose of a diagnostic tool and treatment by helping us locate, measure, and treat the anomalous behavior of mathematical objects. In the classical setup, the degree of a finite Galois extension of "nice" fields splits up neatly into the product of two well-understood numbers (ramification index and inertia degree) that encode how the base field changes. In the general case, however, a third factor called the defect (or ramification deficiency) can pop up. The defect is a mysterious phenomenon and the main obstruction to several long-standing open problems, such as obtaining resolution of singularities. The primary reason is, roughly speaking, that the classical strategy of "objects become nicer after finitely many adjustments" fails when the defect is non-trivial. I will discuss my previous and ongoing work in ramification theory that allows us to understand and treat the defect.

We frame dynamic persuasion in a partial observation stochastic control game with an ergodic criterion. The receiver controls the dynamics of a multidimensional unobserved state process. Information is provided to the receiver through a device designed by the sender that generates the observation process. 

The commitment of the sender is enforced and an exogenous information process outside the control of the sender is allowed. We develop this approach in the case where all dynamics are linear and the preferences of the receiver are linear-quadratic.

We prove a verification theorem for the existence and uniqueness of the solution of the HJB equation satisfied by the receiver’s value function. An extension to the case of persuasion of a mean field of interacting receivers is also provided. We illustrate this approach in two applications: the provision of information to electricity consumers with a smart meter designed by an electricity producer; the information provided by carbon footprint accounting rules to companies engaged in a best-in-class emissions reduction effort. In the first application, we link the benefits of information provision to the mispricing of electricity production. In the latter, we show that when firms declare a high level of best-in-class target, the information provided by stringent accounting rules offsets the Nash equilibrium effect that leads firms to increase pollution to make their target easier to achieve.

This is a joint work with Prof. René Aïd, Prof. Giorgia Callegaro and Prof. Luciano Campi.

 One of the main entries in the AdS/CFT dictionary is a relation between the bulk on-shell partition function with specified boundary conditions and the generating function of correlation functions of primary operators in the boundary CFT. In this talk, I will show how to construct a similar relation for gravity in 4d asymptotically flat spacetimes. For simplicity, we will restrict to the leading infrared sector, where a careful treatment of soft modes and their canonical partners leads to a non-vanishing on-shell action. I will show that this action localizes to a codimension-2 surface and coincides with the generating function of 2d CFT correlators involving insertions of Kac-Moody currents. The latter were previously shown, using effective field theory methods, to reproduce the leading soft graviton theorems in 4d. I will conclude with comments on the implications of these results for the computation of soft charge fluctuations in the vacuum. 

In this seminar I will begin by giving an overview of some problems in stochastic simulation and uncertainty quantification. I will then outline the Multilevel Monte Carlo for situations in which accurate simulations are very costly, but it is possible to perform much cheaper, less accurate simulations.  Inspired by the multigrid method, it is possible to use a combination of these to achieve the desired overall accuracy at a much lower cost.

We investigate an interacting particle model to simulate a foraging colony of ants, where each ant is represented as a so-called active Brownian particle. Interactions among ants are mediated through chemotaxis, aligning their orientations with the upward gradient of the pheromone field. We show how the empirical measure of the interacting particle system converges to a solution of a mean-field limit (MFL) PDE for some subset of the model parameters. We situate the MFL PDE as a non-gradient flow continuity equation with some other recent examples. We then demonstrate that the MFL PDE for the ant model has two distinctive behaviors: the well-known Keller--Segel aggregation into spots and the formation of lanes along which the ants travel. Using linear and nonlinear analysis and numerical methods we provide the foundations for understanding these particle behaviors at the mean-field level. We conclude with long-time estimates that imply that there is no infinite time blow-up for the MFL PDE.

In Bayesian inverse problems, it is common to consider several hyperparameters that define the prior and the noise model that must be estimated from the data. In particular, we are interested in linear inverse problems with additive Gaussian noise and Gaussian priors defined using Matern covariance models. In this case, we estimate the hyperparameters using the maximum a posteriori (MAP) estimate of the marginalized posterior distribution. 

However, this is a computationally intensive task since it involves computing log determinants.  To address this challenge, we consider a stochastic average approximation (SAA) of the objective function and use the preconditioned Lanczos method to compute efficient function evaluation approximations. 

We can therefore compute the MAP estimate of the hyperparameters efficiently by building a preconditioner which can be updated cheaply for new values of the hyperparameters; and by leveraging numerical linear algebra tools to reuse information efficiently for computing approximations of the gradient evaluations.  We demonstrate the performance of our approach on inverse problems from tomography. 

The global COVID19 pandemic has highlighted the lethality and morbidity associated with infectious respiratory diseases. These diseases can lead to devastating syndrome known as the acute respiratory distress syndrome (ARDS) where bacterial/viral infections cause excessive lung inflammation, pulmonary edema, and severe hypoxemia (shortness of breath). Although ARDS patients require artificial mechanical ventilation, the complex biofluid and biomechanical forces generated by the ventilator exacerbates lung injury leading to high mortality. My group has used mathematical and computational modeling to both characterize the complex mechanics of lung injury during ventilation and to identify novel ways to prevent injury at the cellular level. We have used in-vitro and in-vivo studies to validate our mathematical predictions and have used engineering tools to understand the biological consequences of the mechanical forces generated during ventilation. In this talk I will specifically describe how our mathematical/computational approach has led to novel cytoskeletal based therapies and how coupling mathematics and molecular biology has led to the discovery of a gene regulatory mechanisms that can minimize ventilation induced lung injury. I will also describe how we are currently using nanotechnology and gene/drug delivery systems to enhance the lung’s native regulatory responses and thereby prevent lung injury during ARDS.