Refine
Year of publication
Document Type
- Article (55)
- Postprint (4)
- Monograph/Edited Volume (1)
- Review (1)
Language
- English (61)
Is part of the Bibliography
- yes (61) (remove)
Keywords
- data assimilation (7)
- ensemble Kalman filter (7)
- Bayesian inference (5)
- Data assimilation (3)
- GNSS Reflectometry (3)
- gradient flow (3)
- localization (3)
- wind speed (3)
- DDM simulation (2)
- Ensemble Kalman filter (2)
- Fokker-Planck equation (2)
- continuous-time data assimilation (2)
- correlated noise (2)
- eye movements (2)
- multi-scale diffusion processes (2)
- multiplicative noise (2)
- nonlinear filtering (2)
- optimal transport (2)
- parameter estimation (2)
- rain attenuation (2)
- rain effect (2)
- sequential data assimilation (2)
- Advanced scatterometer (ASCAT) (1)
- Atmosphere (1)
- Bayesian inverse problems (1)
- COVID-19 (1)
- CRPS (1)
- Data augmentation (1)
- Data-driven modelling (1)
- Dynamical systems (1)
- Earthquake modeling (1)
- Ensemble Kalman (1)
- Ensemble Kalman Filter (1)
- Error analysis (1)
- European Centre for Medium-Range Weather Forecasts (ECMWF) (1)
- Force splitting (1)
- Fuzzy logic (1)
- GNSS forward scatterometry (1)
- GNSS reflectometry (1)
- Gaussian kernel estimators (1)
- Gaussian mixtures (1)
- Gaussian process (1)
- Generalized hybrid Monte Carlo (1)
- Hamiltonian dynamics (1)
- Hawkes process (1)
- Kalman Bucy filter (1)
- Kalman filter (1)
- Kalman-Bucy Filter (1)
- Lagrangian modeling (1)
- Lagrangian modelling (1)
- Lagrangian-averaged equations (1)
- Langevin dynamics (1)
- MCMC (1)
- MCMC modelling (1)
- McKean-Vlasov (1)
- Modified Hamiltonians (1)
- Molecular dynamics (1)
- Mollification (1)
- Monte Carlo method (1)
- Multigrid (1)
- Multiple time stepping (1)
- NWP (1)
- Nonlinear filters (1)
- Numerical weather prediction (1)
- Optimal transportation (1)
- Paleoclimate reconstruction (1)
- Poincare inequality (1)
- Proxy forward modeling (1)
- RMSE (1)
- Random feature maps (1)
- Sampling (1)
- Self-exciting point process (1)
- Sequential data assimilation (1)
- Sinkhorn approximation (1)
- Sinkhorn problem (1)
- Spatio-temporal ETAS model (1)
- Spectral analysis (1)
- Stochastic epidemic model (1)
- Stormer-Verlet method (1)
- Strike-slip fault model (1)
- TDS-1 (1)
- TechDemoSat-1 (TDS-1) (1)
- Turbulence (1)
- accuracy (1)
- adaptive (1)
- affine (1)
- affine invariance (1)
- asymptotic behavior (1)
- balanced dynamics (1)
- chemistry (1)
- climate reconstructions (1)
- co-limitation (1)
- conservative discretization (1)
- convergence assessment (1)
- diffusion maps (1)
- distribution (1)
- dynamical model (1)
- dynamical models (1)
- electromagnetic scattering (1)
- ensemble (1)
- ensembles (1)
- feedback particle filter (1)
- filter (1)
- fluid mechanics (1)
- forecasting (1)
- framework (1)
- fuzzy logic (1)
- geophysics (1)
- gradient-free (1)
- gradient-free sampling methods (1)
- graph Laplacian (1)
- high resolution paleoclimatology (1)
- highly (1)
- holonomic constraints (1)
- hybrids (1)
- hydrostatic atmosphere (1)
- idealised turbulence (1)
- interacting particle systems (1)
- interacting particles (1)
- interindividual differences (1)
- invariance (1)
- kernel methods (1)
- likelihood (1)
- likelihood function (1)
- limiting factors (1)
- linear programming (1)
- linearly implicit time stepping methods (1)
- mean-field equations (1)
- mesoscale forecasting (1)
- model comparison (1)
- model fitting (1)
- models (1)
- multilevel Monte Carlo (1)
- non-dissipative regularisations (1)
- nonlinear data assimilation (1)
- numerical analysis/modeling (1)
- numerical weather prediction (1)
- numerical weather prediction/forecasting (1)
- ocean surface (1)
- oscillatory systems (1)
- paleoclimate reconstruction (1)
- particle filter (1)
- particle filters (1)
- proposal densities (1)
- proxy forward modeling (1)
- rain detection (1)
- rain splash (1)
- reading (1)
- reanalysis (1)
- regularization (1)
- resampling (1)
- saccades (1)
- semi-Lagrangian method (1)
- shallow-water equations (1)
- short-range prediction (1)
- smoother (1)
- sparse proxy data (1)
- spread correction (1)
- stability (1)
- stiff ODE (1)
- stochastic differential equations (1)
- stochastic systems (1)
- temporal discretization (1)
- transdimensional inversion (1)
- transformations (1)
- variability (1)
- verification (1)
- weight-based formulations (1)
- well-posedness (1)
Diffusion maps is a manifold learning algorithm widely used for dimensionality reduction. Using a sample from a distribution, it approximates the eigenvalues and eigenfunctions of associated Laplace-Beltrami operators. Theoretical bounds on the approximation error are, however, generally much weaker than the rates that are seen in practice. This paper uses new approaches to improve the error bounds in the model case where the distribution is supported on a hypertorus. For the data sampling (variance) component of the error we make spatially localized compact embedding estimates on certain Hardy spaces; we study the deterministic (bias) component as a perturbation of the Laplace-Beltrami operator's associated PDE and apply relevant spectral stability results. Using these approaches, we match long-standing pointwise error bounds for both the spectral data and the norm convergence of the operator discretization. We also introduce an alternative normalization for diffusion maps based on Sinkhorn weights. This normalization approximates a Langevin diffusion on the sample and yields a symmetric operator approximation. We prove that it has better convergence compared with the standard normalization on flat domains, and we present a highly efficient rigorous algorithm to compute the Sinkhorn weights.