Learning dynamical systems from data: a simple cross-validation perspective. I: Parametric kernel flows

DOI10.1016/j.physd.2020.132817OpenAlexW3128400532MaRDI QIDQ2077645

Houman Owhadi, Boumediene Hamzi

Publication date: 21 February 2022

Published in: Physica D (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2007.05074

zbMATH Keywords

kernel methodsdatakernel flowslearning dynamical systems

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Time series analysis of dynamical systems (37M10)

Related Items (20)

Feedforward Neural Networks and Compositional Functions with Applications to Dynamical Systems ⋮ Computational graph completion ⋮ Learning dynamical systems from data: a simple cross-validation perspective. III: Irregularly-sampled time series ⋮ A framework for machine learning of model error in dynamical systems ⋮ Learning dynamical systems from data: a simple cross-validation perspective. IV: Case with partial observations ⋮ Probabilistic forecast of nonlinear dynamical systems with uncertainty quantification ⋮ Gaussian process hydrodynamics ⋮ Data-Driven Kernel Designs for Optimized Greedy Schemes: A Machine Learning Perspective ⋮ Approximation of compositional functions with ReLU neural networks ⋮ Learning to Forecast Dynamical Systems from Streaming Data ⋮ Ensemble forecasts in reproducing kernel Hilbert space family ⋮ Kernel methods are competitive for operator learning ⋮ Comparison of prediction accuracy between interpolation and artificial intelligence application of CFD data for 3D cavity flow ⋮ A note on microlocal kernel design for some slow-fast stochastic differential equations with critical transitions and application to EEG signals ⋮ Consistency of empirical Bayes and kernel flow for hierarchical parameter estimation ⋮ Error bounds of the invariant statistics in machine learning of ergodic Itô diffusions ⋮ Recovering sparse networks: basis adaptation and stability under extensions ⋮ Learning ``best kernels from data in Gaussian process regression. With application to aerodynamics ⋮ One-shot learning of stochastic differential equations with data adapted kernels ⋮ Do ideas have shape? Idea registration as the continuous limit of artificial neural networks

Uses Software

• astsa

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Discovering governing equations from data by sparse identification of nonlinear dynamical systems
• Bases for kernel-based spaces
• Brittleness of Bayesian inference and new Selberg formulas
• Nonlinear prediction of chaotic time series
• Kernel methods for the approximation of some key quantities of nonlinear systems
• Dimensionality reduction of complex metastable systems via kernel embeddings of transition manifolds
• Data-driven approximation of the Koopman generator: model reduction, system identification, and control
• Operator-theoretic framework for forecasting nonlinear time series with kernel analog techniques
• Kernel flows: from learning kernels from data into the abyss
• Brittleness of Bayesian inference under finite information in a continuous world
• On the mathematical foundations of learning
• On the Brittleness of Bayesian Inference
• Local error estimates for radial basis function interpolation of scattered data
• Using machine learning to replicate chaotic attractors and calculate Lyapunov exponents from data
• Qualitative Robustness in Bayesian Inference
• Operator-Adapted Wavelets, Fast Solvers, and Numerical Homogenization
• Bayesian Numerical Homogenization
• Kernel Methods for the Approximation of Nonlinear Systems
• Theory of Reproducing Kernels

This page was built for publication: Learning dynamical systems from data: a simple cross-validation perspective. I: Parametric kernel flows