The difficulty of computing stable and accurate neural networks: On the barriers of deep learning and Smale's 18th problem

MJ Colbrook, V Antun… - Proceedings of the …, 2022 - National Acad Sciences
Deep learning (DL) has had unprecedented success and is now entering scientific
computing with full force. However, current DL methods typically suffer from instability, even …

Residual dynamic mode decomposition: robust and verified Koopmanism

MJ Colbrook, LJ Ayton, M Szőke - Journal of Fluid Mechanics, 2023 - cambridge.org
Dynamic mode decomposition (DMD) describes complex dynamic processes through a
hierarchy of simpler coherent features. DMD is regularly used to understand the …

The mpEDMD algorithm for data-driven computations of measure-preserving dynamical systems

MJ Colbrook - SIAM Journal on Numerical Analysis, 2023 - SIAM
Koopman operators globally linearize nonlinear dynamical systems and their spectral
information is a powerful tool for the analysis and decomposition of nonlinear dynamical …

Rigorous data‐driven computation of spectral properties of Koopman operators for dynamical systems

MJ Colbrook, A Townsend - Communications on Pure and …, 2024 - Wiley Online Library
Koopman operators are infinite‐dimensional operators that globally linearize nonlinear
dynamical systems, making their spectral information valuable for understanding dynamics …

Beyond expectations: residual dynamic mode decomposition and variance for stochastic dynamical systems

MJ Colbrook, Q Li, RV Raut, A Townsend - Nonlinear Dynamics, 2024 - Springer
Koopman operators linearize nonlinear dynamical systems, making their spectral
information of crucial interest. Numerous algorithms have been developed to approximate …

Convergence rates for learning linear operators from noisy data

MV de Hoop, NB Kovachki, NH Nelsen… - SIAM/ASA Journal on …, 2023 - SIAM
This paper studies the learning of linear operators between infinite-dimensional Hilbert
spaces. The training data comprises pairs of random input vectors in a Hilbert space and …

The multiverse of dynamic mode decomposition algorithms

MJ Colbrook - arxiv preprint arxiv:2312.00137, 2023 - arxiv.org
Dynamic Mode Decomposition (DMD) is a popular data-driven analysis technique used to
decompose complex, nonlinear systems into a set of modes, revealing underlying patterns …

On the computation of geometric features of spectra of linear operators on Hilbert spaces

MJ Colbrook - Foundations of Computational Mathematics, 2024 - Springer
Computing spectra is a central problem in computational mathematics with an abundance of
applications throughout the sciences. However, in many applications gaining an …

Implicit regularization in AI meets generalized hardness of approximation in optimization--Sharp results for diagonal linear networks

JS Wind, V Antun, AC Hansen - arxiv preprint arxiv:2307.07410, 2023 - arxiv.org
Understanding the implicit regularization imposed by neural network architectures and
gradient based optimization methods is a key challenge in deep learning and AI. In this work …

[HTML][HTML] A contour method for time-fractional PDEs and an application to fractional viscoelastic beam equations

MJ Colbrook, LJ Ayton - Journal of Computational Physics, 2022 - Elsevier
We develop a rapid and accurate contour method for the solution of time-fractional PDEs.
The method inverts the Laplace transform via an optimised stable quadrature rule, suitable …