Past

Gastrulation is a critical event in vertebrate morphogenesis, characterized by coordinated large-scale multi-cellular movements. One grand challenge in modern biology is understanding how spatio-temporal morphological structures emerge from cellular processes in a developing organism and vary across vertebrates. We derive a theoretical framework that couples tissue flows, stress-dependent myosin activity, and actomyosin cable orientation. Our model, consisting of a set of nonlinear coupled PDEs, predicts the onset and development of observed experimental patterns of wild-type and perturbations of chick gastrulation as a spontaneous instability of a uniform state. We use analysis and numerics to show how our model recapitulates the phase space of gastrulation morphologies seen across vertebrates, consistent with experiments. Altogether, this suggests that early embryonic self-organization follows from a minimal predictive theory of active mechano-sensitive flows. 

 https://www.biorxiv.org/content/10.1101/2021.10.03.462928v2 

We use energy estimates to derive new bounds on the eigenvalues of a generic form of double saddle-point matrices, with and without regularization terms. Results related to inertia and algebraic multiplicity of eigenvalues are also presented. The analysis includes eigenvalue bounds for preconditioned matrices based on block-diagonal Schur complement-based preconditioners, and it is shown that in this case the eigenvalues are clustered within a few intervals bounded away from zero. The analytical observations are linked to a few multiphysics problems of interest. This is joint work with Susanne Bradley.

---

A link for this talk will be sent to our mailing list a day or two in advance.  If you are not on the list and wish to be sent a link, please contact @email.

Symmetrically Colored Gaussian Graphical Models with Toric Vanishing Ideals

Jane Coons

Gaussian graphical models are multivariate Gaussian statistical models in which a graph encodes conditional independence relations among the random variables. Adding colors to this graph allows us to describe situations where some entries in the concentration matrices in the model are assumed to be equal. In this talk, we focus on RCOP models, in which this coloring is obtained from the orbits of a subgroup of the automorphism group of the underlying graph. We show that when the underlying block graph is a one-clique-sum of complete graphs, the Zariski closure of the set of concentration matrices of an RCOP model on this graph is a toric variety. We also give a Markov basis for the vanishing ideal of this variety in these cases.

Topological persistence for multi-scale terrain profiling and feature detection in drylands hydrology

Gillian Grindstaff

With the growing availability of remote sensing products and computational resources, an increasing amount of landscape data is available, and with it, increasing demand for automated feature detection and useful morphological summaries. Topological data analysis, and in particular, persistent homology, has been applied successfully to detect landslides and characterize soil pores, but its application to hydrology is currently still limited. We demonstrate how persistent homology of a real-valued function on a two-dimensional domain can be used to summarize critical points and shape in a landscape simultaneously across all scales, and how that data can be used to automatically detect features of hydrological interest, such as: experimental conditions in a rainfall simulator, boundary conditions of landscape evolution models, and earthen berms and stock ponds, placed historically to alter natural runoff patterns in the American southwest.

We say that a graph G is Local-to-Global rigid if there exists R>0 such that every other graph whose balls of radius R are isometric to the balls of radius R in G is covered by G. Examples include the Euclidean building of PSLn(Qp). We show that the rigidity of the building goes further by proving that a reconstruction is possible from only a partial local information, called “print”. We use this to prove the rigidity of graphs quasi-isometric to the building among which are the torsion-free lattices of PSLn(Qp).

TBA

I will present a study of integrable structures and quantum chaos in a class of infinite-dimensional though computationally tractable models, called quantum resonant systems. These models, together with their classical counterparts, emerge in various areas of physics, such as nonlinear dynamics in anti-de Sitter spacetime, but also in Bose-Einstein condensate physics. The class of classical models displays a wide range of integrable properties, such as the existence of Lax pairs, partial solvability or generic chaotic dynamics. This opens a window to investigate these properties from the perspective of the corresponding quantum theory by effectively diagonalising finite-sized matrices and exploring level spacing statistics. We will furthermore analyse the implications of the symmetries for the spectrum of resonant models with partial solvability and discuss how the rich integrable structures can be exploited to constructed novel quantum coherent states that effectively capture sophisticated nonlinear solutions in the classical theory.

Many applications involve non-Euclidean data, where exploiting Riemannian geometry can deliver algorithms that are computationally superior to standard nonlinear programming approaches. This observation has resulted in an increasing interest in Riemannian methods in the optimization and machine learning community. In this talk, we consider the problem of optimizing a function on a Riemannian manifold subject to convex constraints. We introduce Riemannian Frank-Wolfe (RFW) methods, a class of projection-free algorithms for constrained geodesically convex optimization. To understand the algorithm’s efficiency, we discuss (1) its iteration complexity, (2) the complexity of computing the Riemannian gradient and (3) the complexity of the Riemannian “linear” oracle (RLO), a crucial subroutine at the heart of the algorithm. We complement our theoretical results with an empirical comparison of RFW against state-of-the-art Riemannian optimization methods. Joint work with Suvrit Sra (MIT).

There's an idea going back at least to Kirchberger in 1902 that since the only operations we can ultimately compute are +, -, *, and /, all of numerical computation must reduce to rational functions.  I've been looking into this idea and it has led in some interesting directions.

Identifying novel drug-target interactions (DTI) is a critical and rate limiting step in drug discovery. While deep learning models have been proposed to accelerate the identification process, we show that state-of-the-art models fail to generalize to novel (i.e., never-before-seen) structures. We first unveil the mechanisms responsible for this shortcoming, demonstrating how models rely on shortcuts that leverage the topology of the protein-ligand bipartite network, rather than learning the node features. Then, we introduce AI-Bind, a pipeline that combines network-based sampling strategies with unsupervised pre-training, allowing us to limit the annotation imbalance and improve binding predictions for novel proteins and ligands. We illustrate the value of AI-Bind by predicting drugs and natural compounds with binding affinity to SARS-CoV-2 viral proteins and the associated human proteins. We also validate these predictions via auto-docking simulations and comparison with recent experimental evidence. Overall, AI-Bind offers a powerful high-throughput approach to identify drug-target combinations, with the potential of becoming a powerful tool in drug discovery.

arXiv link: https://arxiv.org/abs/2112.13168

The Iwasawa algebra of a compact $p$-adic Lie group is fundamental to the study of the representations of the group. Understanding this representation theory is crucial in progress towards a (mod p) local Langlands correspondence. However, much remains unknown about Iwasawa algebras and their modules.

In this talk we'll aim to measure the size of the Iwasawa algebra and its representations. I'll explain the algebraic tools we use to do this - Krull dimension and canonical dimension - and survey previously known examples. Our main result is a new bound on these dimensions for the group $SL_2(O_F)$, where $F$ is a finite extension of the p-adic numbers. When $F$ is a quadratic extension, we find the Krull dimension is exactly 5, as predicted by a conjecture of Ardakov and Brown.

Symmetry protected topological (SPT) phases have recently attracted a lot of
attention from physicists and mathematicians as a topological classification
scheme for gapped ground states. In this talk I will briefly introduce the
operator algebraic approach to SPT phases in the infinite-volume limit. In
particular, I will focus on the quasifree (free-fermionic) setting, where we

can adapt tools from algebraic quantum field theory to describe phases of
gapped ground states using K-homology and the coarse index.

TBC

We'll sketch how the $K$-rational solutions of a system $X$ of multivariate polynomials can be viewed as the solutions of a "classical mechanics" problem on an associated affine space.

When $X$ has a suitable topology, e.g. if its $\mathbb{C}$-solutions form a Riemann surface of genus $>1$, we'll observe some of the advantages of this new point of view such as a relatively computable algorithm for effective finiteness (with some stipulations). This is joint work with Minhyong Kim.
 

Lattice homology is a combinatorial invariant of plumbed 3-manifolds due to Nemethi. The definition is a formalization of Ozsvath and Szabo's computation of the Heegaard Floer homology of plumbed 3-manifolds. Nemethi conjectured that lattice homology is isomorphic to Heegaard Floer homology. For a restricted class of plumbings, this isomorphism is known to hold, due to work of Ozsvath-Szabo, Nemethi, and Ozsvath-Stipsicz-Szabo. By using the Manolescu-Ozsvath link surgery formula for Heegaard Floer homology, we prove the conjectured isomorphism in general. In this talk, we will talk about aspects of the proof, and some related topics and extensions of the result.

The homogeneous coordinate ring of a projective variety may be constructed by geometrically quantizing the multiples of a symplectic form, using the complex structure as a polarization. In this talk, I will explain how a holomorphic Poisson structure allows us to deform the complex polarization into a generalized complex structure, leading to a non-commutative deformation of the homogeneous coordinate ring. The main tool is a conjectural construction of a category of generalized complex branes, which makes use of the A-model of an associated symplectic groupoid. I will explain this in the example of toric Poisson varieties. This is joint work with Marco Gualtieri (arXiv:2108.01658).

The global symmetries of a d-dimensional quantum field theory (QFT), and their ’t Hooft anomalies, are conveniently captured by a topological field theory (TFT) in (d+1) dimensions, which we may refer to as the Symmetry TFT of the given d-dimensional QFT. This point of view has a vast range of applicability: it encompasses both ordinary symmetries, as well as generalized symmetries. In this talk, I will discuss systematic methods to compute the Symmetry TFT for QFTs realized by M-theory on a singular, non-compact space X. The desired Symmetry TFT is extracted from the topological couplings of 11d supergravity, via reduction on the space L, the boundary of X. The formalism of differential cohomology allows us to include discrete symmetries originating from torsion in the cohomology of L. I will illustrate this framework in two classes of examples: M-theory on an ALE space (engineering 7d SYM theory); M-theory on Calabi-Yau cones (engineering 5d superconformal field theories).

In many machine learning tasks, it is crucial to extract low-dimensional and descriptive features from a data set. In this talk, I present a method to extract features from multi-dimensional space-time signals which is motivated, on the one hand, by the success of path signatures in machine learning, and on the other hand, by the success of models from the theory of regularity structures in the analysis of PDEs. I will present a flexible definition of a model feature vector along with numerical experiments in which we combine these features with basic supervised linear regression to predict solutions to parabolic and dispersive PDEs with a given forcing and boundary conditions. Interestingly, in the dispersive case, the prediction power relies heavily on whether the boundary conditions are appropriately included in the model. The talk is based on the following joint work with Andris Gerasimovics and Hendrik Weber: https://arxiv.org/abs/2108.05879

Deep learning continues to dominate machine learning and has been successful in computer vision, natural language processing, etc. Its impact has now expanded to many research areas in science and engineering. In this talk, I will mainly focus on some recent impacts of deep learning on computational mathematics. I will present our recent work on bridging deep neural networks with numerical differential equations, and how it may guide us in designing new models and algorithms for some scientific computing tasks. On the one hand, I will present some of our works on the design of interpretable data-driven models for system identification and model reduction. On the other hand, I will present our recent attempts at combining wisdom from numerical PDEs and machine learning to design data-driven solvers for PDEs and their applications in electromagnetic simulation.

The path signature is a characterization of paths that originated in Chen's iterated integral cochain model for path spaces and loop spaces. More recently, it has been used to form the foundations of rough paths in stochastic analysis, and provides an effective feature map for sequential data in machine learning. In this talk, we return to the topological foundations in Chen's construction to develop generalizations of the signature.

Adaptive agents through active inference: The main fields of research that are used to model and realise adaptive agents are optimal control, reinforcement learning and active inference. Active inference is a probabilistic description of adaptive agents that is relatively less known to mathematicians, as it originated from neuroscience in the last decade. This talk presents the mathematical underpinnings of active inference, starting from fundamental considerations about agents that maintain their structural integrity in the face of environmental perturbations. Through this, we derive a probability distribution over actions, that describes decision-making under uncertainty in adaptive agents . Interestingly, this distribution has an interesting information geometric structure, combining, for instance, drives for exploration and exploitation, which may yield a principled answer to the exploration-exploitation trade-off. Preserving this geometric structure enables to realise adaptive agents in practice. We illustrate their behaviour with simulation examples and empirical comparisons with reinforcement learning.

Scaling Limits of Random Graphs: The scaling limit of directed random graphs remains relatively unexplored compared to their undirected counterparts. In contrast, many real-world networks, such as links on the world wide web, financial transactions and “follows” on Twitter, are inherently directed. Previous work by Goldschmidt and Stephenson established the scaling limit for the strongly connected components (SCCs) of the Erdős -- Rényi model in the critical window when appropriately rescaled. In this talk, we present a result showing the SCCs of another class of critical random directed graphs will converge when rescaled to the same limit. Central to the proof is an exploration of the directed graph and subsequent encodings of the exploration as real valued random processes. We aim to present this exploration algorithm and other key components of the proof.

From Mathematics to Data Science and Back: We give an overview of the interaction between rough path theory and data science at the current time.
 

We consider spectral methods that uncover hidden structures in directed networks. We establish and exploit connections between node reordering via (a) minimizing an objective function and (b) maximizing the likelihood of a random graph model. We focus on two existing spectral approaches that build and analyse Laplacian-style matrices via the minimization of frustration and trophic incoherence. These algorithms aim to reveal directed periodic and linear hierarchies, respectively. We show that reordering nodes using the two algorithms, or mapping them onto a specified lattice, is associated with new classes of directed random graph models. Using this random graph setting, we are able to compare the two algorithms on a given network and quantify which structure is more likely to be present. We illustrate the approach on synthetic and real networks, and discuss practical implementation issues. This talk is based on a joint work with Desmond Higham and Konstantinos Zygalakis. 

Article link: https://royalsocietypublishing.org/doi/10.1098/rsos.211144