Past

There is great interest in the molecular characterisation of intestinal metaplasia, such as Barrett’s esophagus (BE), to understand the basic biology of metaplastic development from a tissue of origin. BE is asymptomatic, so it is not generally known how long a patient has lived with this precursor of esophageal adenocarcinoma (EAC) when initially diagnosed in the clinic. We previously constructed a BE clock model using patient-specific methylation data to estimate BE onset times using Bayesian inference techniques, and thus obtain the biological age of BE tissue (Curtius et al. 2016). We find such epigenetic drift to be widely evident in BE tissue (Luebeck et al. 2017) and the corresponding tissue ages show large inter-individual heterogeneity in two patient populations.               

From a basic biological mechanism standpoint, it is not fully understood how the Barrett’s tissue first forms in the human esophagus because this process is never observed in vivo, yet such information is critical to inform biomarkers of risk based on temporal features (e.g., growth rates, tissue age) reflecting the evolution toward cancer. We analysed multi-region samples from 17 BE patients to

1) measure the spatial heterogeneity in biological tissue ages, and 2) use these ages to calibrate mathematical models (agent-based and continuum) of the mechanisms for formation of the segment itself. Most importantly, we found that tissue must be regenerated nearer to the stomach, perhaps driven by wound healing caused by exposure to reflux, implying a gastric tissue of origin for the lesions observed in BE. Combining bioinformatics and mechanistic modelling allowed us to infer evolutionary processes that cannot be clinically observed and we believe there is great translational promise to develop such hybrid methods to better understand multiscale cancer data.

References:

Curtius K, Wong C, Hazelton WD, Kaz AM, Chak A, et al. (2016) A Molecular Clock Infers Heterogeneous Tissue Age Among Patients with Barrett's Esophagus. PLoS Comput Biol 12(5): e1004919

Luebeck EG, Curtius K, Hazelton WD, Made S, Yu M, et al. (2017) Identification of a key role of epigenetic drift in Barrett’s esophagus and esophageal adenocarcinoma. J Clin Epigenet 9:113

Data science has become a topic of great interest lately and has triggered new widescale research activities around efficientl first order methods for optimisation and Bayesian sampling. The National Physical Laboratory is addressing some of these challenges with particular focus on  robustness and confidence in the solution.  In this talk, I will present some problems and recent results concerning i. robust learning in the presence of outliers based on the Median of Means (MoM) principle and ii. stability of the solution in super-resolution (joint work with A. Thompson and B. Toader).

 The talk is about the normal modal logics of elementary classes defined by first-order formulas of the form
 'for all x_0 there exist x_1, ..., x_n phi(x_0, x_1, ... x_n)' with phi being a conjunction of binary atoms.
 I'll show that many properties of these logics, such as finite axiomatisability,
 elementarity,  axiomatisability by a set of canonical formulas or by a single generalised Sahlqvist formula,
 together with modal definability of the initial formula, either simultaneously hold or simultaneously do not hold.
 

The Asai (or twisted tensor) L-function attached to a Bianchi modular form is the 'restriction to the rationals' of the standard L-function. Introduced by Asai in 1977, subsequent study has linked its special values to the arithmetic of the corresponding form. In this talk, I will discuss joint work with David Loeffler in which we construct a p-adic Asai L-function -- that is, a measure on Z_p* that interpolates the critical values L^As(f,chi,1) -- for ordinary weight 2 Bianchi modular forms. We use a new method for constructing p-adic L-functions, using Kato's system of Siegel units to build a 'Betti analogue' of an Euler system, building on algebraicity results of Ghate. I will start by giving a brief introduction to p-adic L-functions and Bianchi modular forms, and if time permits, I will briefly mention another case where the method should apply, that of non-self-dual automorphic representations for GL(3).

A two-layer shear flow in the presence of surfactants is considered. The flow configuration comprises two superposed layers of viscous and immiscible fluids confined in a long horizontal channel, and characterised by different densities, viscosities and thicknesses. The surfactants can be insoluble, i.e. located at the interface between the two fluids only, or soluble in the lower fluid in the form of monomers (single molecules) or micelles (multi-molecule aggregates). A mathematical model is formulated, consisting of governing equations for the hydrodynamics and appropriate transport equations for the surfactant concentration at the interface, the concentration of monomers in the bulk fluid and the micelle concentration. A primary objective of this study is to investigate the effect of surfactants on the stability of the interface, and in particular surfactants in high concentrations and above the critical micelle concentration (CMC). Interfacial instabilities are induced due to the acting forces of gravity and inertia, as well as the action of Marangoni forces generated as a result of the dependence of surface tension on the interfacial surfactant concentration. The underlying physical mechanism responsible for the formation of interfacial waves will be discussed, together with the complex flow dynamics (typical nonlinear phenomena associated with interfacial flows include travelling waves, solitary pulses, quasi-periodic and chaotic dynamics).

I will describe a family of 2d TQFTs, due to Moore-Tachikawa, which take values in a category whose objects are Lie groups and whose morphisms are holomorphic symplectic varieties. They link many interesting aspects of geometry, such as moduli spaces of solutions to Nahm equations, hyperkähler reduction, and geometric invariant theory.

 Multiscale methods based on coupled partial differential equations defined on bulk and embedded manifolds are still poorly explored from the theoretical standpoint, although they are successfully used in applications, such as microcirculation and flow in perforated subsurface reservoirs. This work aims at shedding light on some theoretical aspects of a multiscale method consisting of coupled partial differential equations defined on one-dimensional domains embedded into three-dimensional ones. Mathematical issues arise because the dimensionality gap between the bulk and the inclusions is larger than one, named as the high dimensionality gap case. First, we show that such model derives from a system of full three-dimensional equations, by the application of a topological model reduction approach. Secondly, we rigorously analyze the problem, showing that the averaging operators applied for the model reduction introduce a regularization effect that resolves the issues due to the singularity of solutions and to the ill-posedness of restriction operators. Then, we discretize the problem by means of the finite element method and we analyze the approximation error. Finally, we exploit the structure of the model reduction technique to analyze the modeling error. This study confirms that for infinitesimally small inclusions, the modeling error vanishes.

This is a joint work with Federica Laurino, Department of Mathematics, Politecnico di Milano.

In this talk, we study the long time behaviour of a cloud of weakly interacting Brownian particles, conditionally on the observation of their initial and final configuration. In particular, we connect this problem, which may be regarded as a nonlinear version of the Schrödinger problem, to the study of the long time behaviour of Mean Field Games. Combining tools from optimal transport and stochastic control we prove convergence towards the equilibrium configuration and establish convergence rates. A key ingredient to derive these results is a new functional inequality, which generalises Talagrand’s inequality to the entropic transportation cost.

A stacking is a lift of an immersion of graphs $A\to B$ to an embedding of $A$ into the product of $B$ with the real line; their existence relates to orderability properties of groups. I will describe how Louder and Wilton used them to prove Wise's "$w$-cycles" conjecture: given a primitive word $w$ in a free group $F$, and a subgroup $H < F$, the number of conjugates of $H$ which intersect $<w>$ nontrivially is at most rank($H$). I will also discuss applications of the result to questions of coherence, and possible extensions of it.

The Grothendieck ring of varieties over a field is a simple idea that formalizes various cut-and-paste arguments in algebraic geometry. We will explain how this intuitive construction leads to nontrivial results, such as computing Euler characteristics, counting points of varieties over finite fields, and determining Hodge numbers. As an example, we will investigate cubic hypersurfaces, especially the varieties parametrizing lines on them. If time permits, we will discuss some of the stranger properties of the Grothendieck ring.

We study random (simple) graphs with given vertex degrees, in the sparse case where the average degree is bounded. Assume also that the second moment of the vertex degree is bounded. The standard method then is to use the configuration model to construct a random multigraph and condition it on
being simple.

This works well for results of the type that something holds with high probability, or that something converges in probability, but it does not immediately apply to convergence in distribution, for example asymptotic normality. (Although this has been done by special arguments in a couple of cases, by Janson and Luczak and by Riordan.) A typical example is the recent result by Barbour and Röllin on asymptotic normality of the size of the giant component of the multigraph (in the supercritical case); it is an obvious conjecture that the same results hold for the random simple graph.

We discuss two new approaches to this, both based on old methods. Both apply to the size of the giant component, using rather minor special arguments.

One approach uses the method of moments to obtain joint convergence of the variable of interest together with the numbers of loops and multiple edges
in the  multigraph.

The other approach uses switchings to modify the multigraph and construct a simple graph. This simple random graph will not have a uniform distribution,
but almost, and this is good enough.

The idea of adjusting prices in order to sell goods at the highest acceptable price, such as haggling in a market, is as old as money itself. We consider the problem of pricing multiple products on a network of resources, such as that faced by an airline selling tickets on its flight network. In this talk I will consider various optimization relaxations to the deterministic dynamic pricing problem on a network. This is joint work with Raphael Hauser.

In the 1990s Moy and Prasad revolutionized the representation theory of p-adic groups by showing how to use Bruhat-Tits theory to assign invariants to representations of p-adic groups. The tools they introduced resulted in rapid advancements in both representation theory and harmonic analysis -- areas of central importance in the Langlands program. A crucial ingredient for many results is an explicit construction of (types for) representations of p-adic groups. In this talk I will indicate why, survey what constructions are known (no knowledge about p-adic groups assumed) and present recent developments based on a refinement of Moy and Prasad's invariants.​

In recent years, there has been an increased interest in exploiting rank structure of matrices arising from the discretization of partial differential equations to develop fast direct solvers. In this talk, I will outline the fundamental ideas of this topic in the context of solving the integral equation formulation of the Helmholtz equation, known as the Lippmann-Schwinger equation, and will discuss some plans for future work to develop new, higher-order solvers. This is joint work with Gunnar Martinsson.

We propose a novel class of network models for temporal dyadic interaction data. Our objective is to capture important features often observed in social interactions: sparsity, degree heterogeneity, community structure and reciprocity. We use mutually-exciting Hawkes processes to model the interactions between each (directed) pair of individuals. The intensity of each process allows interactions to arise as responses to opposite interactions (reciprocity), or due to shared interests between individuals (community structure). For sparsity and degree heterogeneity, we build the non time dependent part of the intensity function on compound random measures following (Todeschini et al., 2016). We conduct experiments on real- world temporal interaction data and show that the proposed model outperforms competing approaches for link prediction, and leads to interpretable parameters.

Link to paper: https://papers.nips.cc/paper/7502-modelling-sparsity-heterogeneity-reci…

 I will give a short survey on classical inequalities for Laplace eigenvalues, tell about related history and questions. I will then discuss the so-called Korevaar method, and new results generalising to higher eigenvalues a number of classical inequalities known for the first Laplace eigenvalue only. 

In his famous book on partial differential relations Gromov formulates an exercise concerning local deformations of solutions to open partial differential relations. We will explain the content of this fundamental assertion and sketch a proof. 

In the sequel we will apply this to extend local deformations of closed $G_2$ structures, and to construct 
$C^{1,1}$-Riemannian metrics which are positively curved "almost everywhere" on arbitrary manifolds. 

This is joint work with Christian Bär (Potsdam).

We consider Brownian motion with Kato class measure-valued drift.   A small time large deviation principle and a Varadhan type asymptotics for the Brownian motion with singular drift are established. We also study the existence and uniqueness of the associated Dirichlet boundary value problems.

We first give a short introduction to analysis and stochastic processes on fractal state spaces and the typical difficulties involved.

We then discuss gradient operators and semilinear PDE. They are used to formulate the main result which establishes the hydrodynamic limit of the weakly asymmetric exclusion process on the Sierpinski gasket in the form of a law of large numbers for the particle density. We will explain some details and, if time permits, also sketch a corresponding large deviations principle for the symmetric case.

Roughly speaking, a commutative-by-finite Hopf algebra is a Hopf
algebra which is an extension of a commutative Hopf algebra by a
finite dimensional Hopf algebra.
There are many big and significant classes of such algebras
(beyond of course the commutative ones and the finite dimensional ones!).
I'll make the definition precise, discuss examples
and review results, some old and some new.
No previous knowledge of Hopf algebras is necessary.
 

For a polynomial $h(x)$ in $F[x]$, where $F$ is any field, let $A$ be the
$F$-algebra given by generators $x$ and $y$ and relation $[y, x]=h$.
This family of algebras include the Weyl algebra, enveloping algebras of
$2$-dimensional Lie algebras, the Jordan plane and several other
interesting subalgebras of the Weyl algebra.

In a joint work in progress with Samuel Lopes, we computed the Hochschild
cohomology $HH^*(A)$ of $A$ and determined explicitly the Gerstenhaber
structure of $HH^*(A)$, as a Lie module over the Lie algebra $HH^1(A)$.
In case $F$ has characteristic $0$, this study has revealed that $HH^*(A)$
has finite length as a Lie module over $HH^1(A)$ with pairwise
non-isomorphic composition factors and the latter can be naturally
extended into irreducible representations of the Virasoro algebra.
Moreover, the whole action can be understood in terms of the partition
formed by the multiplicities of the irreducible factors of the polynomial
$h$.
 

Ice streams are fast flowing regions of ice that generally slide over a layer of unconsolidated, water-saturated subglacial sediment known as till.  A striking feature that has been observed geophysically is that subglacial till has been found to accumulate distinctively into sedimentary wedges at the grounding zones (regions where ice sheets begin to detach from the bedrock to form freely floating ice shelves) of both past and present-day ice sheets. These grounding-zone wedges have important implications for ice-sheet stability against grounding zone retreat in response to rising sea levels, and their origins have remained a long-standing question. Using a combination of mathematical modelling, a series of laboratory experiments, field data and numerical simulations, we develop a fluid-mechanical model that explains the mechanism of the formation of these sedimentary wedges in terms of the loading and unloading of deformable till in three dynamical regimes. We also undertake a series of analogue laboratory experiments, which reveal that a similar wedge of underlying fluid accumulates spontaneously in experimental grounding zones, we formulate local conditions relating wedge slopes in each of the scenarios and compare them to available geophysical radargram data at the well lubricated, fast-flowing Whillans Ice Stream.

Mathematical biology has grown enormously over the past 40 years and has changed considerably. At first, biology inspired mathematicians to come up with models that could, at an abstract level, "explain" biological phenomena - one of the most famous being Alan Turing's model for biological pattern formation. However, with the enormous recent advances in biotechnology and computation, the field is now truly inter- and multi-disciplinary. We shall discuss the changing role mathematics is playing in applications to biology and medicine.

Genome replication is a stochastic process whereby each cell exhibits different patterns of origin activation and replication fork movement.  Despite this heterogeneity, replication is a remarkably stable process that works quickly and correctly over hundreds of thousands of iterations. Existing methods for measuring replication dynamics largely focus on how a population of cells behave on average, which precludes the detection of low probability errors that may have occurred in individual cells.  These errors can have a severe impact on genome integrity, yet existing single-molecule methods, such as DNA combing, are too costly, low-throughput, and low-resolution to effectively detect them.  We have created a method that uses Oxford Nanopore sequencing to create high-throughput genome-wide maps of DNA replication dynamics in single molecules.  I will discuss the informatics approach that our software uses, our use of mathematical modelling to explain the patterns that we observe, and questions in DNA replication and genome stability that our method is uniquely positioned to answer.

Given a stability condition defined over a category, every object in this category
is filtered by some distinguished objects called semistables. This
filtration, that is unique up-to-isomorphism, is know as the
 Harder-Narasimhan filtration.
One less studied property of stability conditions, when defined over an
 abelian category, is the fact that each of them induce a chain of torsion
classes that is naturally indexed.
 In this talk we will study arbitrary indexed chain of torsion classes. Our
first result states that every indexed chain of torsion classes induce a
 Harder-Narasimhan filtration. Following ideas from Bridgeland we
 show that the set of all indexed chains of torsion classes satisfying a mild 
 technical condition forms a topological space. If time we
 will characterise the neighbourhood or some distinguished points. 

Landmark-based human action recognition in videos is a challenging task in computer vision. One crucial step is to design discriminative features for spatial structure and temporal dynamics. To this end, we use and refine the path signature as an expressive, robust, nonlinear, and interpretable representation for landmark-based streamed data. Instead of extracting signature features from raw sequences, we propose path disintegrations and transformations as preprocessing to improve the efficiency and effectiveness of signature features. The path disintegrations spatially localize a pose into a collection of m-node paths from which the signatures encode non-local and non-linear geometrical dependencies, while temporally transform the evolutions of spatial features into hierarchical spatio-temporal paths from which the signatures encode long short-term dynamical dependencies. The path transformations allow the signatures to further explore correlations among different informative clues. Finally, all features are concatenated to constitute the input vector of a linear fully-connected network for action recognition. Experimental results on four benchmark datasets demonstrated that the proposed feature sets with only linear network achieves comparable state-of-the-art result to the cutting-edge deep learning methods. 

Starting with a countable transitive model of V=L, we show that by 
adding a single Cohen real, c, most intermediate models do no satisfy choice. In 
fact, most intermediate models to L[c] are not even definable.

The key part of the proof is the Bristol model, which is intermediate to L[c], 
but is not constructible from a set. We will give a broad explanation of the 
construction of the Bristol model within the constraints of time.

Gallagher's theorem is a strengthening of the Littlewood conjecture that holds for almost all pairs of real numbers. I'll discuss some recent refinements of Gallagher's theorem, one of which is joint work with Niclas Technau. A key new ingredient is the correspondence between Bohr sets and generalised arithmetic progressions. It is hoped that these are the first steps towards a metric theory of multiplicative diophantine approximation on manifolds. 

I will give an overview of the Nielsen-Thurston theory of the mapping class group and its connection to hyperbolic geometry and dynamics. Time permitting, I will discuss the surface entropy conjecture and a theorem of Hamenstadt on entropies of `generic' elements of the mapping class group. No prior knowledge of the concepts involved is required.

Injection of a gas into a gas/liquid foam is known to give rise to instability phenomena on a variety of time and length scales. Macroscopically, one observes a propagating gas-filled structure that can display properties of liquid finger propagation as well as of fracture in solids. Using a discrete model, which incorporates the underlying film instability as well as viscous resistance from the moving liquid structures, we describe brittle cleavage phenomena in line with experimental observations. We find that  the dimensions of the foam sample significantly affect the speed of the  cracks as well as the pressure necessary to sustain them: cracks in wider samples travel faster at a given driving stress, but are able to avoid arrest and maintain propagation at a lower pressure (the  velocity gap becomes smaller). The system thus becomes a study case for stress concentration and the transition between discrete and continuum systems in dynamical fracture; taking into account the finite dimensions of the system improves agreement with experiment.

Driven by its numerous applications in computational science, the approximation of smooth, high-dimensional functions via sparse polynomial expansions has received significant attention in the last five to ten years.  In the first part of this talk, I will give a brief survey of recent progress in this area.  In particular, I will demonstrate how the proper use of compressed sensing tools leads to new techniques for high-dimensional approximation which can mitigate the curse of dimensionality to a substantial extent.  The rest of the talk is devoted to approximating functions defined on irregular domains.  The vast majority of works on high-dimensional approximation assume the function in question is defined over a tensor-product domain.  Yet this assumption is often unrealistic.  I will introduce a method, known as polynomial frame approximation, suitable for broad classes of irregular domains and present theoretical guarantees for its approximation error, stability, and sample complexity.  These results show the suitability of this approach for high-dimensional approximation through the independence (or weak dependence) of the various guarantees on the ambient dimension d.  Time permitting, I will also discuss several extensions.

Stein ($1981$) proved the borderline Sobolev embedding result which states that for $n \geq 2,$ $u \in L^{1}(\mathbb{R}^{n})$ and $\nabla u \in L^{(n,1)}(\mathbb{R}^{n}; \mathbb{R}^{n})$ implies $u$ is continuous. Coupled with standard Calderon-Zygmund estimates for Lorentz spaces, this implies $u \in C^{1}(\mathbb{R}^{n})$ if $\Delta u \in L^{(n,1)}(\mathbb{R}^{n}).$ The search for a nonlinear generalization of this result culminated in the work of Kuusi-Mingione ($2014$), which proves the same result for $p$-Laplacian type systems. \paragraph{} In this talk, we shall discuss how these results can be extended to differential forms. In particular, we can prove that if $u$ is an $\mathbb{R}^{N}$-valued $W^{1,p}_{loc}$ $k$-differential form with $\delta \left( a(x) \lvert du \rvert^{p-2} du \right) \in L^{(n,1)}_{loc}$ in a domain of $\mathbb{R}^{n}$ for $N \geq 1,$ $n \geq 2,$ $0 \leq k \leq n-1, $ $1 < p < \infty, $ with uniformly positive, bounded, Dini continuous scalar function $a$, then $du$ is continuous.

I will give a self-contained introduction to the theory of cross ratios on boundaries of Gromov hyperbolic and CAT(-1) spaces, focussing on the connections to the following two questions. When are two spaces with the 'same' Gromov boundary isometric/quasi-isometric? Are closed Riemannian manifolds completely determined (up to isometry) by the lengths of their closed geodesics?

Young diagrams frequently appear in the study of partitions and representations of the symmetric group. By filling these diagrams with numbers, we obtain Young tableau, combinatorial objects onto which we can define the structure of a monoid via insertion algorithms. We will explore this structure and it's connection to a basis of the ring of symmetric polynomials. If we have time, we will mention alternative monoid structures and their corresponding bases.

Polar classes are very classical objects in Algebraic Geometry. A brief introduction to the subject will be presented and ideas and preliminarily results towards generalisations will be explained. These ideas can be applied towards variety sampling and relevant applications. 
 

This talk will feature a brief introduction to persistent homology, the vanguard technique in topological data analysis. Nothing will be required of the audience beyond a willingness to row-reduce enormous matrices (by hand if we can, by machine if we must).

In combinatorics, the 'nicest' way to prove that two sets have the same size is to find a bijection between them, giving more structure to the seeming numerical coincidences. In representation theory, many of the outstanding conjectures seem to imply that the characteristic p of the ground field can be allowed to vary, and we can relate different groups and different primes, to say that they have 'the same' representation theory. In this talk I will try to make precise what we could mean by this

For decades, researchers have been studying efficient numerical methods to solve differential equations, most of them optimised for one-core processors. However, we are about to reach the limit in the amount of processing power we can squeeze into a single processor. This explains the trend in today's computing industry to design high-performance processors looking at parallel architectures. As a result, there is a need to develop low-complexity parallel algorithms capable of running efficiently in terms of computational time and electric power.

Parallelisation across time appears to be a promising way to provide more parallelism. In this talk, we will introduce the main algorithms, following (Gander, 2015), with a particular focus on the parareal algorithm.

Unintended low energy thermal or mechanical stimuli can lead to the accidental ignition of explosive materials. During such events, described as ‘insults’ in the literature, ignition of the explosive is caused by localised regions of high temperature known as ‘hot spots’. We develop a model which helps us to understand how highly localised shear deformation, so-called shear banding, acts as a mechanism for hot spot generation. Through a boundary layer analysis, we give a deeper insight into how the additional self heating caused by chemical reactions affects the initiation and development of shear bands,  and highlight the key physical properties which control this process.

I will talk about the implications of UV completion of quantum gravity on the low energy spectrums. I will introduce the constraints on low-energy effective theory due to unitarity and analyticity of scattering amplitudes, in particular an infinite set of new unitarity constraints on the forward-limit limit of four-point scattering amplitudes due to the work of Arkani-Hamed et al. In three dimensions, we find the constraints imply that light states with charge-to-mass ratio z greater than 1 can only be consistent if there exists other light states, preferably neutral. Applied to the 3D Standard Model like spectrum, where the low energy couplings are dominated by the electron with z \sim 10^22, this provides a novel understanding of the need for light neutrinos.