Global rates of convergence for nonconvex optimization on manifolds

N Boumal, PA Absil, C Cartis - IMA Journal of Numerical …, 2019 - academic.oup.com
We consider the minimization of a cost function f on a manifold using Riemannian gradient
descent and Riemannian trust regions (RTR). We focus on satisfying necessary optimality …

Metric flow matching for smooth interpolations on the data manifold

K Kapuśniak, P Potaptchik, T Reu, L Zhang… - arXiv preprint arXiv …, 2024 - arxiv.org
Matching objectives underpin the success of modern generative models and rely on
constructing conditional paths that transform a source distribution into a target distribution …

Representation learning via manifold flattening and reconstruction

M Psenka, D Pai, V Raman, S Sastry, Y Ma - Journal of Machine Learning …, 2024 - jmlr.org
A common assumption for real-world, learnable data is its possession of some low-
dimensional structure, and one way to formalize this structure is through the manifold …

Manifold learning with arbitrary norms

J Kileel, A Moscovich, N Zelesko, A Singer - Journal of Fourier Analysis …, 2021 - Springer
Manifold learning methods play a prominent role in nonlinear dimensionality reduction and
other tasks involving high-dimensional data sets with low intrinsic dimensionality. Many of …

On Taylor-like Estimates for Polynomial Approximations

A Wodecki - arXiv preprint arXiv:2401.10169, 2024 - arxiv.org
Polynomial series approximations are a central theme in approximation theory. Two types of
series, which are featured most prominently in pure and applied mathematics, are Taylor …

Elucidating Flow Matching ODE Dynamics with respect to Data Geometries

G Mishne, Z Wan, Q Wang, Y Wang - arXiv preprint arXiv:2412.18730, 2024 - arxiv.org
Diffusion-based generative models have become the standard for image generation. ODE-
based samplers and flow matching models improve efficiency, in comparison to diffusion …

Application of Hertzian theory to torus on plane contacts

J Kelley, K Babaalihaghighi, N Bader… - Proceedings of the …, 2022 - journals.sagepub.com
Hertzian theory includes a well-known analytical solution for the calculation of the contact
area and pressure between two bodies. Hertzian theory is not unconditionally applicable, in …

Warped geometric information on the optimisation of Euclidean functions

M Hartmann, B Williams, H Yu, M Girolami… - arXiv preprint arXiv …, 2023 - arxiv.org
We consider the fundamental task of optimizing a real-valued function defined in a
potentially high-dimensional Euclidean space, such as the loss function in many machine …

The one-sided isometric extension problem

N Hungerbühler, M Wasem - Results in Mathematics, 2017 - Springer
Let Σ Σ be a codimension one submanifold of an n-dimensional Riemannian manifold M,
n\geqslant 2 n⩾ 2. We give a necessary condition for an isometric immersion of Σ Σ …

Regularity of the geodesic flow on submanifolds

C Lange - arXiv preprint arXiv:2401.03417, 2024 - arxiv.org
arXiv:2401.03417v1 [math.DG] 7 Jan 2024 Page 1 arXiv:2401.03417v1 [math.DG] 7 Jan 2024
REGULARITY OF THE GEODESIC FLOW ON SUBMANIFOLDS CHRISTIAN LANGE Abstract …