Diffusion maps, spectral clustering and reaction coordinates of dynamical systems
From MaRDI portal
Publication:2497982
DOI10.1016/J.ACHA.2005.07.004zbMATH Open1103.60069arXivmath/0503445OpenAlexW2117684310MaRDI QIDQ2497982FDOQ2497982
Authors: Boaz Nadler, Stéphane Lafon, I. G. Kevrekidis, Ronald R. Coifman 
Publication date: 4 August 2006
Published in: Applied and Computational Harmonic Analysis (Search for Journal in Brave)
Abstract: A central problem in data analysis is the low dimensional representation of high dimensional data, and the concise description of its underlying geometry and density. In the analysis of large scale simulations of complex dynamical systems, where the notion of time evolution comes into play, important problems are the identification of slow variables and dynamically meaningful reaction coordinates that capture the long time evolution of the system. In this paper we provide a unifying view of these apparently different tasks, by considering a family of {em diffusion maps}, defined as the embedding of complex (high dimensional) data onto a low dimensional Euclidian space, via the eigenvectors of suitably defined random walks defined on the given datasets. Assuming that the data is randomly sampled from an underlying general probability distribution , we show that as the number of samples goes to infinity, the eigenvectors of each diffusion map converge to the eigenfunctions of a corresponding differential operator defined on the support of the probability distribution. Different normalizations of the Markov chain on the graph lead to different limiting differential operators. One normalization gives the Fokker-Planck operators with the same potential U(x), best suited for the study of stochastic differential equations as well as for clustering. Another normalization gives the Laplace-Beltrami (heat) operator on the manifold in which the data resides, best suited for the analysis of the geometry of the dataset, regardless of its possibly non-uniform density.
Full work available at URL: https://arxiv.org/abs/math/0503445
Recommendations
random walksbackward Fokker-Planck operatorhigh-dimensional stochastic systemslimiting differential operators
Classification and discrimination; cluster analysis (statistical aspects) (62H30) Applications of Brownian motions and diffusion theory (population genetics, absorption problems, etc.) (60J70) Diffusion processes (60J60)
Cites Work
- The elements of statistical learning. Data mining, inference, and prediction
- Diffusion maps
- Title not available (Why is that?)
- Equation-free, coarse-grained multiscale computation: enabling microscopic simulators to perform system-level analysis
- Laplacian Eigenmaps for Dimensionality Reduction and Data Representation
- The Fokker-Planck equation. Methods of solution and applications.
- Diffusion Maps, Reduction Coordinates, and Low Dimensional Representation of Stochastic Systems
- Geometric diffusions as a tool for harmonic analysis and structure definition of data: diffusion maps
- Machine Learning: ECML 2004
- Learning Theory
- Extracting macroscopic dynamics: model problems and algorithms
- Extracting macroscopic stochastic dynamics: Model problems
- Handbook of stochastic methods for physics, chemistry and the natural sciences.
- Learning Theory
- Title not available (Why is that?)
Cited In (97)
- Doubly stochastic normalization of the Gaussian kernel is robust to heteroskedastic noise
- Equation-free model reduction in agent-based computations: coarse-grained bifurcation and variable-free rare event analysis
- Data-driven efficient solvers for Langevin dynamics on manifold in high dimensions
- Local kernels and the geometric structure of data
- Equation free projective integration: a multiscale method applied to a plasma ion acoustic wave
- Computational coarse graining of a randomly forced one-dimensional Burgers equation
- Nonlinear Laplacian spectral analysis for time series with intermittency and low-frequency variability
- Diffusion maps
- Diffusion-based kernel methods on Euclidean metric measure spaces
- Variable bandwidth diffusion kernels
- Parsimonious representation of nonlinear dynamical systems through manifold learning: a chemotaxis case study
- Transition manifolds of complex metastable systems. Theory and data-driven computation of effective dynamics
- Non-linear independent component analysis with diffusion maps
- Patch-to-tensor embedding
- ATLAS: a geometric approach to learning high-dimensional stochastic systems near manifolds
- Orientability and diffusion maps
- Understanding the geometry of transport: diffusion maps for Lagrangian trajectory data unravel coherent sets
- Dynamics-adapted cone kernels
- An equivalence between the limit smoothness and the rate of convergence for a general contraction operator family
- Data-driven model reduction and transfer operator approximation
- Go with the flow, on Jupiter and snow. Coherence from model-free video data without trajectories
- Measure-based diffusion grid construction and high-dimensional data discretization
- Continuous-time random walks for the numerical solution of stochastic differential equations
- Towards hybrid system modeling of uncertain complex dynamical systems
- Diffusion representation for asymmetric kernels
- An efficient tree-based computation of a metric comparable to a natural diffusion distance
- Coarse analysis of collective motion with different communication mechanisms
- An experimental investigation of kernels on graphs for collaborative recommendation and semisupervised classification
- Robust Inference of Manifold Density and Geometry by Doubly Stochastic Scaling
- Time-Inhomogeneous Diffusion Geometry and Topology
- Learning Binary Hash Codes for Large-Scale Image Search
- Kernel Analog Forecasting: Multiscale Test Problems
- A bag-of-paths framework for network data analysis
- Towards mesoscopic ergodic theory
- Hearing the clusters of a graph: A distributed algorithm
- A spectral notion of Gromov-Wasserstein distance and related methods
- Landmark diffusion maps (L-dMaps): accelerated manifold learning out-of-sample extension
- A data-driven approximation of the koopman operator: extending dynamic mode decomposition
- Learning the geometry of common latent variables using alternating-diffusion
- Some extensions of E. Stein's work on Littlewood-Paley theory applied to symmetric diffusion semigroups
- Extracting Sparse High-Dimensional Dynamics from Limited Data
- An equation-free approach to analyzing heterogeneous cell population dynamics
- An equation-free approach to coarse-graining the dynamics of networks
- Fuzzy spectral clustering by PCCA+: application to Markov state models and data classification
- Thomas Bayes' walk on manifolds
- Poincaré maps for multiscale physics discovery and nonlinear Floquet theory
- Diffusion Maps, Reduction Coordinates, and Low Dimensional Representation of Stochastic Systems
- Mathematics of smoothed particle hydrodynamics: a study via nonlocal Stokes equations
- Some remarks on diffusion distances
- Metric-based upscaling
- Latent common manifold learning with alternating diffusion: analysis and applications
- Coarse analysis of collective behaviors: bifurcation analysis of the optimal velocity model for traffic jam formation
- Diffusion state distances: multitemporal analysis, fast algorithms, and applications to biological networks
- An equation-free computational approach for extracting population-level behavior from individual-based models of biological dispersal
- Data-driven control of agent-based models: an equation/variable-free machine learning approach
- A multiscale environment for learning by diffusion
- Dimensionality reduction of complex metastable systems via kernel embeddings of transition manifolds
- Physics-constrained, data-driven discovery of coarse-grained dynamics
- Coarse-grained computation for particle coagulation and sintering processes by linking quadrature method of moments with Monte-Carlo
- Manifold learning for the emulation of spatial fields from computational models
- Dimensionality reduction: an interpretation from manifold regularization perspective
- A stochastic multiscale model for electricity generation capacity expansion
- Nonlinear Laplacian spectral analysis: capturing intermittent and low‐frequency spatiotemporal patterns in high‐dimensional data
- Data-Driven Discovery of Governing Equations for Coarse-Grained Heterogeneous Network Dynamics
- Nonparametric uncertainty quantification for stochastic gradient flows
- Point cloud discretization of Fokker-Planck operators for committor functions
- Modern Koopman theory for dynamical systems
- Fuzzy diffusion distance learning for cartoon similarity estimation
- Physics-agnostic and physics-infused machine learning for thin films flows: modelling, and predictions from small data
- Time-series forecasting using manifold learning, radial basis function interpolation, and geometric harmonics
- Local and global perspectives on diffusion maps in the analysis of molecular systems
- Tipping points of evolving epidemiological networks: machine learning-assisted, data-driven effective modeling
- Towards effective dynamics in complex systems by Markov kernel approximation
- Recovering hidden components in multimodal data with composite diffusion operators
- A geometrical method for low-dimensional representations of simulations
- Diffusion maps tailored to arbitrary non-degenerate Itô processes
- Numerical bifurcation analysis of PDEs from lattice Boltzmann model simulations: a parsimonious machine learning approach
- Introducing User-Prescribed Constraints in Markov Chains for Nonlinear Dimensionality Reduction
- Functional diffusion maps
- Grassmannian diffusion maps based surrogate modeling via geometric harmonics
- Imaging geometry through dynamics: the observable representation
- Scalable extended dynamic mode decomposition using random kernel approximation
- Probing multipartite entanglement, coherence and quantum information preservation under classical Ornstein-Uhlenbeck noise
- Understanding Graph Neural Networks with Generalized Geometric Scattering Transforms
- Gaussian Process Landmarking for Three-Dimensional Geometric Morphometrics
- Data-driven prediction of multistable systems from sparse measurements
- Transient anisotropic kernel for probabilistic learning on manifolds
- A framework for self-evolving computational material models inspired by deep learning
- Parameter rating by diffusion gradient
- Clustering Dynamics on Graphs: From Spectral Clustering to Mean Shift Through Fokker–Planck Interpolation
- Reduction methods in climate dynamics -- a brief review
- Geometric fluid approximation for general continuous-time Markov chains
- A tailored convolutional neural network for nonlinear manifold learning of computational physics data using unstructured spatial discretizations
- Manifold learning for organizing unstructured sets of process observations
- Geometric scattering on measure spaces
- Maximally predictive states: from partial observations to long timescales
- Data clustering based on Langevin annealing with a self-consistent potential
This page was built for publication: Diffusion maps, spectral clustering and reaction coordinates of dynamical systems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2497982)
