How would we get a powerful AI to align itself with human preferences? What are human preferences anyway? And how can you code all this?
It turns out that maths give you the grounding to answer these fascinating and vital questions.
The Camassa-Holm (CH) equation is a nonlinear nonlocal dispersive equation which arises as a model for the propagation of unidirectional shallow water waves over a flat bottom. One of the most important features of the CH equation is the existence of peaked travelling waves, also called peakons. The aim of this talk is to review some asymptotic stability result for peakon solutions for CH-type equations as well as to present some new result for higher-order generalization of the CH equation.
The automorphism group of a d-regular tree is a topological group with many interesting features. A nice thing about this group is that while some of its features are highly non-trivial (e.g., the existence of infinitely many pairwise non-conjugate simple subgroups), often the ideas involved in the proofs are fairly intuitive and geometric.
I will present a proof for the fact that the outer automorphism group of (Aut(T)) is trivial. This is original joint work with Gil Goffer, but as is often the case in this area, was already proven by Bass-Lubotzky 20 years ago. I will mainly use this talk to hint at how algebra, topology and geometry all play a role when working with Aut(T).
Scrolls play a central role in the construction of varieties; they are ambient spaces for K3 surfaces and Fano 3-folds. In this talk, I will say in two ways what scrolls are and give examples of some embedded varieties in them.
We will be discussing a Fourier-analytic approach
to optimal matching between independent samples, with
an elementary proof of the Ajtai-Komlos-Tusnady theorem.
The talk is based on a joint work with Michel Ledoux.
It is known that random matrix distributions such as those that describe the largest eignevalue of the Gaussian Orthogonal and Symplectic ensembles (GOE, GSE) admit two types of representations: one in terms of a Fredholm Pfaffian and one in terms of a Fredholm determinant. The equality of the two sets of expressions has so far been established via involved computations of linear algebraic nature. We provide a structural explanation of this duality via links (old and new) between the model of last passage percolation and the irreducible characters of classical groups, in particular the general linear, symplectic and orthogonal groups, and by studying, combinatorially, how their representations decompose when restricted to certain subgroups. Based on joint work with Elia Bisi.
Tautological bundles on Hilbert schemes of points often enter into enumerative and physical computations. I will explain how to use the Donaldson-Thomas theory of toric threefolds to produce combinatorial identities that are expressed geometrically using tautological bundles on the Hilbert scheme of points on a surface. I'll also explain how these identities can be used to study Euler characteristics of tautological bundles over Hilbert schemes of points on general surfaces.
We show that the scalability challenges of Global Optimisation algorithms can be overcome for functions with low effective dimensionality, which are constant along certain linear subspaces. Such functions can often be found in applications, for example, in hyper-parameter optimization for neural networks, heuristic algorithms for combinatorial optimization problems and complex engineering simulations. We propose the use of random subspace embeddings within a(ny) global minimisation algorithm, extending the approach in Wang et al. (2016). Using tools from random matrix theory and conic integral geometry, we investigate the efficacy and convergence of our random subspace embeddings approach, in a static and/or adaptive formulation. We illustrate our algorithmic proposals and theoretical findings numerically, using state of the art global solvers. This work is joint with Coralia Cartis.
A group is sofic when every finite subset can be well approximated in a finite symmetric group. The outstanding question, due to Gromov, is whether every group is sofic.
Helfgott and Juschenko argued that a celebrated group constructed by Higman is unlikely to be sofic because its soficity would imply the existence of some seemingly pathological functions. I will describe joint work with Martin Kassabov and Vivian Kuperberg in which we construct variations on Higman's group and explore their soficity.
Modularity is a function on graphs which is used in algorithms for community detection. For a given graph G, each partition of the vertices has a modularity score, with higher values indicating that the partition better captures community structure in $G$. The (max) modularity $q^\ast(G)$ of the graph $G$ is defined to be the maximum over all vertex partitions of the modularity score, and satisfies $0 \leq q^\ast(G) \leq 1$.
We analyse when community structure of an underlying graph can be determined from an observed subset of the graph. In a natural model where we suppose edges in an underlying graph $G$ appear with some probability in our observed graph $G'$ we describe how high a sampling probability we need to infer the community structure of the underlying graph.
Joint work with Colin McDiarmid.
The QDWH algorithm can compute the polar decomposition of a matrix in a stable and efficient way. We generalize this method in order to compute generalized polar decompositions with respect to signature matrices. Here, the role of the QR decomposition is played by the hyperbolic QR decomposition. However, it doesn't show the same favorable properties concerning stability as its orthogonal counterpart. Remedies are found by exploiting connections to the LDL^T factorization and by employing well-conditioned permuted graph bases. The computed polar decomposition is used to formulate a structure-preserving spectral divide-and-conquer method for pseudosymmetric matrices. Applications of this method are found in computational quantum physics, where eigenvalues and eigenvectors describe optical properties of condensed matter or molecules. Additional properties guarantee fast convergence and a reduction to symmetric definite eigenvalue problems after just one step of spectral divide-and-conquer.
Numerous mathematical models have been proposed for modelling cancerous tumour invasion (Gatenby and Gawlinski 1996), angiogenesis (Owen et al 2008), growth kinetics (Wang et al 2009), response to irradiation (Gao et al 2013) and metastasis (Qiam and Akcay 2018). In this study, we attempt to model the qualitative behavior of growth, invasion, angiogenesis and fragmentation of tumours at the tissue level in an explicitly spatial and continuous manner in two dimensions. We simulate the effectiveness of radiation therapy on a growing tumour in comparison with immunotherapy and propose a novel framework based on vector fields for modelling the impact of interstitial flow on tumour morphology. The results of this model demonstrate the effectiveness of employing a system of partial differential equations along with vector fields for simulating tumour fragmentation and that immunotherapy, when applicable, is substantially more effective than radiation therapy.
A polynomial form is established for the off-shell CHY scattering equations proposed by Lam and Yao. Re-expressing this in terms of independent Mandelstam invariants provides a new expression for the polynomial scattering equations, immediately valid off shell, which makes it evident that they yield the off-shell amplitudes given by massless ϕ3 Feynman graphs. A CHY expression for individual Feynman graphs, valid even off shell, is established through a recurrence relation.
It is known that many real-world networks exhibit geometric properties. Brain networks, social networks, and wireless communication networks are a few examples. Storage and transmission of the information contained in the topologies and structures of these networks are important tasks, which, given their scale, is often nontrivial. Although some (but not much) work has been done to characterize and develop compression limits and algorithms for nonspatial graphs, little is known for the spatial case. In this talk, we will discuss an information theoretic formalism for studying compression limits for a fairly broad class of random geometric graphs. We will then discuss entropy bounds for these graphs and, time permitting, local (pairwise) connection rules that yield maximum entropy properties in the induced graph distribution.
Given a number field K, it is natural to ask whether it has a finite extension with ideal class number one. This question can be translated into a fundamental question in class field theory, namely the class field tower problem. In this talk, we are going to discuss this problem as well as its solution due to Golod and Shafarevich using methods of group cohomology.
The Steklov eigenvalue problem is an eigenvalue problem whose spectral parameters appear in the boundary condition. On a Riemannian surface with smooth boundary, Steklov eigenvalues have a very sharp asymptotic expansion. Also, a number of interesting sharp bounds for the $k$th Steklov eigenvalues have been known. We extend these results on orbisurfaces and discuss how the structure of orbifold singularities comes into play. This is joint work with Arias-Marco, Dryden, Gordon, Ray and Stanhope.
A broad challenge in the theory of finitely generated groups is to understand their subgroups. A group is commensurably coHopfian if its finite index subgroups are distinct from its infinite index subgroups (that is to say not abstractly isomorphic). We will focus primarily on hyperbolic groups, and give the first examples of one-ended hyperbolic groups that are not commensurably coHopfian.
This is joint work with Emily Stark.
"We consider a spatial Lambda-Fleming-Viot process, a model in mathematical biology, with a randomly chosen (rough) selection field. We study the scaling limit of this process in different regimes. This leads to the analysis of semi-discrete approximations of singular SPDEs, in particular the Parabolic Anderson Model and allows to extend previous results to weakly nonlinear cases. The subject presented is based on joint works with Aleksander Klimek and Nicolas Perkowski."
In this talk I will introduce a continuous wetting model consisting of the law of a Brownian meander tilted by its local time at a positive level h, with h small. I will prove that this measure converges, as h tends to 0, to the same weak limit as for discrete critical wetting models. I will also discuss the corresponding gradient dynamics, which is expected to converge to a Bessel SPDE admitting the law of a reflecting Brownian motion as invariant measure. This is based on joint work with Jean-Dominique Deuschel and Tal Orenshtein.
McDuff and Schlenk determined when a four-dimensional symplectic ellipsoid can be symplectically embedded into a four-dimensional ball. They found that if the ellipsoid is close to round, the answer is given by an ``infinite staircase" determined by the odd index Fibonacci numbers, while if the ellipsoid is sufficiently stretched, all obstructions vanish except for the volume obstruction. Infinite staircases have also been found when embedding ellipsoids into polydisks (Frenkel - Muller, Usher) and into the ellipsoid E(2, 3) (Cristofaro-Gardiner - Kleinman). In this talk, we will see how the sharpness of ECH capacities for embedding of ellipsoids implies the existence of infinite staircases for these and three other target spaces. We will then discuss the relationship with toric varieties, lattice point counting, and the Philadelphia subway system. This is joint work with Dan Cristofaro-Gardiner, Alessia Mandini,
and Ana Rita Pires.
In the process of studying the zeta-function for one parameter families of Calabi-Yau manifolds we have been led to a manifold, for which the quartic numerator of the zeta-function factorises into two quadrics remarkably often. Among these factorisations, we find persistent factorisations; these are determined by a parameter that satisfies an algebraic equation with coefficients in Q, so independent of any particular prime. We note that these factorisations are due a splitting of Hodge structure and that these special values of the parameter are rank two attractor points in the sense of IIB supergravity. To our knowledge, these points provide the first explicit examples of non-singular, non-rigid rank two attractor points for Calabi-Yau manifolds of full SU(3) holonomy. Modular groups and modular forms arise in relation to these attractor points in a way that, to a physicist, is unexpected. This is a report on joint work with Xenia de la Ossa, Mohamed Elmi and Duco van Straten.
The Mathematical Institute Colloquia are funded in part by the generosity of Oxford University Press.
This Colloquium is supported by a Leverhulme Trust Visiting Professorship award.
In this talk, we will provide an overview of results in the setting of granular crystals, consisting of spherical beads interacting through nonlinear elastic spring-like forces. These crystals are used in numerous engineering applications including, e.g., for the production of "sound bullets'' or the examination of bone quality. In one dimension we show that there exist three prototypical types of coherent nonlinear waveforms: shock waves, traveling solitary waves and discrete breathers. The latter are time-periodic, spatially localized structures. For each one, we will analyze the existence theory, presenting connections to prototypical models of nonlinear wave theory, such as the Burgers equation, the Korteweg-de Vries equation and the nonlinear Schrodinger (NLS) equation, respectively. We will also explore the stability of such structures, presenting some explicit stability criteria for traveling waves in lattices. Finally, for each one of these structures, we will complement the mathematical theory and numerical computations with state-of-the-art experiments, allowing their quantitative identification and visualization. Finally, time permitting, ongoing extensions of these themes will be briefly touched upon, most notably in higher dimensions, in heterogeneous or disordered chains and in the presence of damping and driving; associated open questions will also be outlined.
Molecules are dynamical systems that can adopt a variety of three dimensional conformations which, in general, differ in energy and physical properties. The identification of energetically favourable conformations is fundamental in molecular physics and computational chemistry, since it is closely related to important open problems such as the prediction of the folding of proteins and virtual screening for drug design.
In this talk I will present theoretical and data-driven approaches to the study of molecular conformational spaces and their associated energy landscapes. I will show that the topology of the internal molecular conformational space might change after taking its quotient by the group action of a discrete group of symmetries. I will also show that geometric and topological tools for data analysis such as procrustes analysis, local dimensionality reduction, persistent homology and discrete Morse theory provide with efficient methods to study the mathematical structures underlying the molecular conformational spaces and their energy landscapes.
Taught courses offer a range of distinctive learning opportunities from lectures to tutorials/supervisions through to individual study. Orchestration refers to the combining and sequencing of these opportunities for maximum effect. This raises a question about who does the orchestration. In school, there is a good case for suggesting that it is teachers who take responsibility for orchestration of students’ learning opportunities. Moving to university, do students take on more responsibility for orchestration?
In this session there will be a chance to look back on the learning opportunities you experienced last term and to reflect on how (or even if) they were orchestrated. What could be different in the term ahead if you pay more attention to how distinctive learning opportunities are orchestrated?
Globally, almost 38 million people are living with HIV. HPTN 071 (PopART) is the largest HIV prevention trial to date, taking place in 21 communities in Zambia and South Africa with a combined population of more than 1 million people. As part of the trial an individual-based mathematical model was developed to help in planning the trial, to help interpret the results of the trial, and to make projections both into the future and to areas where the trial did not take place. In this talk I will outline the individual-based mathematical model used in the trial, the inference framework, and will discuss examples of how the results from the model have been used to help inform policy decisions.