arXiv research

This article sketches various ideas in contact geometry that have become useful in low-dimensional topology. Specifically we (1) outline the proof of Eliashberg and Thurston's results concerning perturbations of foliatoins into contact structures, (2) discuss Eliashberg and Weinstein's symplectic handle attachments, an…

Course notes on Lie groups and Riemannian geometry, focusing on applications and low-dimensional examples.

Abstracts discuss a common framework for constructing homology theories.

Cryo-em images are found to be low-dimensional.

Explains how knots relate to 4D shapes.

Scattering networks maximize separation on low-dimensional data.

Survey on categorifying Jones polynomial.

The purpose of this thesis is to study classical combinatorial objects, such as polytopes, polytopal complexes, and subspace arrangements, using tools that have been developed in combinatorial topology, especially those tools developed in connection with (discrete) differential geometry, geometric group theory and low-…

We create real-time geodesic rendering for non-isotropic geometries.

Survey on GKM theory in low dimensions, highlighting combinatorics-geometry interplay.

We show that a car, viewed as a nonholonomic system, provides an example of a flat parabolic geometry of type $({\bf SO}(2,3),P_{12})$, where $P_{12}$ is a Borel parabolic subgroup in ${\bf SO}(2,3)$. We discuss the relations of this geometry of a car with the geometry of circles in the plane (a low dimensional Lie sph…

The study of infinite groups through their finite quotients in geometry.

A new method for generative modeling of discrete data using geometric latent subspaces.

Adversarial examples are a pervasive phenomenon of machine learning models where seemingly imperceptible perturbations to the input lead to misclassifications for otherwise statistically accurate models. We propose a geometric framework, drawing on tools from the manifold reconstruction literature, to analyze the high-…

In this paper, we build an organization of high-dimensional datasets that cannot be cleanly embedded into a low-dimensional representation due to missing entries and a subset of the features being irrelevant to modeling functions of interest. Our algorithm begins by defining coarse neighborhoods of the points and defin…

Great computational effort is invested in generating equilibrium states for molecular systems using, for example, Markov chain Monte Carlo. We present a probabilistic model that generates statistically independent samples for molecules from their graph representations. Our model learns a low-dimensional manifold that p…

Study on reliability of latent reuse in diffusion models under distribution shift.

A neural network learns efficient parametrizations of product shape spaces.

The paper proposes a method to learn 3D object pose manifolds using GANs and elasticae.

In 1995 the author, Jones, and Segal introduced the notion of "Floer homotopy theory". The proposal was to attach a (stable) homotopy type to the geometric data given in a version of Floer homology. More to the point, the question was asked, "When is the Floer homology isomorphic to the (singular) homology of a natural…

RNNs classify text by accumulating evidence on a low-dimensional manifold.

Survey on strong closing lemmas in Hamiltonian dynamics.

Let $\hat{G}$ be a group and $G$ its normal subgroup. In this paper, we study $\hat{G}$-invariant quasimorphisms on $G$ which appear in symplectic geometry and low dimensional topology. As its application, we prove the non-existence of a section of the flux homomorphism on closed surfaces of higher genus. We also prove…

Develops optimal low-dimensional approximations to high-dimensional SDEs.

The primary objects of study in the ``knot theory of complex plane curves'' are C-links: links (or knots) cut out of a 3-sphere in the complex plane by complex plane transverse and totally tangential. Transverse C-links are naturally oriented. There are many natural classes of examples: links of singularities; links at…

Researchers develop neural optimal transport with Lagrangian costs for efficient computation.

Chart autoencoders learn latent features preserving manifold topology and geometry, with robust denoising capabilities.

Paper presents a new method for learning hyperbolic representations using tree structures.

We present a novel kernel-based machine learning algorithm for identifying the low-dimensional geometry of the effective dynamics of high-dimensional multiscale stochastic systems. Recently, the authors developed a mathematical framework for the computation of optimal reaction coordinates of such systems that is based …

A simple formula is derived for the Ricci scalar curvature of any smooth level set ${ψ(x_0,x_1,...,x_n)=C}$ embedded in the Euclidean space $\mathbb R^{n+1}$, in terms of the gradient $\nablaψ$ and the Laplacian $Δψ$. Some applications are given to the geometry of low-dimensional $p$-harmonic functions and high-dime…

GH-PID uses guided harmonic paths for efficient SOT with interpretable diagnostics.

Estimates curvature of network manifolds to understand community structure.

This paper tackles high-dimensional Bayesian optimization by projecting a manifold into a lower space.

Nonlinear analysis has played a prominent role in the recent developments in geometry and topology. The study of the Yang-Mills equation and its cousins gave rise to the Donaldson invariants and more recently, the Seiberg-Witten invariants. Those invariants have enabled us to prove a number of striking results for low …

Riemannian metric matching learns the geometry of high-dimensional datasets using neural networks.

We introduce the concept of a Clifford-Weyl structure on a conformal manifold, which consists of an even Clifford structure parallel with respect to the tensor product of a metric connection on the Clifford bundle and a Weyl structure on the manifold. We show that the Weyl structure is necessarily closed except for som…

Sub-Riemannian Geometry is proved to play an important role in many applications, e.g., Mathematical Physics and Control Theory. The simplest example of sub-Riemannian structure is provided by the 3-D Heisenberg group. Sub-Riemannian Geometry enjoys major differences from the Riemannian being a generalisation of the la…

Sparse connectivity improves generalization in neural networks below the Edge of Stability.

Flag manifolds are generalizations of projective spaces and other Grassmannians: they parametrize flags, which are nested sequences of subspaces in a given vector space. These are important objects in algebraic and differential geometry, but are also increasingly being used in data science, where many types of data are…

The local geometry of high dimensional neural network loss landscapes can both challenge our cherished theoretical intuitions as well as dramatically impact the practical success of neural network training. Indeed recent works have observed 4 striking local properties of neural loss landscapes on classification tasks: …

Cartan's method of moving frames is briefly recalled in the context of immersed curves in the homogeneous space of a Lie group $G$. The contact geometry of curves in low dimensional equi-affine geometry is then made explicit. This delivers the complete set of invariant data which solves the $G$-equivalence problem via …

This paper proposes a new method for learning covers of geometric datasets to improve topological inference and visualization.

Subdivision rules create sequences of nested cell structures on CW-complexes, and they frequently arise from groups. In this paper, we develop several tools for classifying subdivision rules. We give a criterion for a subdivision rule to represent a Gromov hyperbolic space, and show that a subdivision rule for a hyperb…

A basic problem in machine learning is to find a mapping $f$ from a low dimensional latent space $\mathcal{Y}$ to a high dimensional observation space $\mathcal{X}$. Modern tools such as deep neural networks are capable to represent general non-linear mappings. A learner can easily find a mapping which perfectly fits a…

The study explains transformer scaling laws using statistical and approximation theories.

Non-linear manifold learning enables high-dimensional data analysis, but requires out-of-sample-extension methods to process new data points. In this paper, we propose a manifold learning algorithm based on deep learning to create an encoder, which maps a high-dimensional dataset and its low-dimensional embedding, and …

Physics: Similar long-distance properties can mask vastly different short-distance metrics.

Manifold learning techniques for dynamical systems and time series have shown their utility for a broad spectrum of applications in recent years. While these methods are effective at learning a low-dimensional representation, they are often insufficient for visualizing the global and local structure of the data. In thi…