Neural ODE Control for Classification, Approximation, and Transport

DOI10.1137/21M1411433zbMath1521.93080arXiv2104.05278OpenAlexW3155438876MaRDI QIDQ6115450

Domènec Ruiz-Balet, Enrique Zuazua

Publication date: 10 August 2023

Published in: SIAM Review (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2104.05278

zbMATH Keywords

transport equationsuniversal approximationWasserstein distancedata classificationdeep learningsimultaneous controlneural ODEs

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Control problems involving ordinary differential equations (34H05) Control/observation systems governed by ordinary differential equations (93C15)

Related Items (2)

Connections between Operator-Splitting Methods and Deep Neural Networks with Applications in Image Segmentation ⋮ Dynamical Systems–Based Neural Networks

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Computational Optimal Transport: With Applications to Data Science
• Optimal observability of the multi-dimensional wave and Schrödinger equations in quantum ergodic domains
• From averaged to simultaneous controllability
• How violent are fast controls?
• Image recovery via total variation minimization and related problems
• Complete controllability of continuous-time recurrent neural networks
• An algorithm for total variation minimization and applications
• Multilayer feedforward networks are universal approximators
• A computational fluid mechanics solution to the Monge-Kantorovich mass transfer problem
• Polytopal approximation bounding the number of \(k\)-faces
• Minimal time for the continuity equation controlled by a localized perturbation of the velocity vector field
• Nonlinear approximation and (deep) ReLU networks
• Interpolation and approximation via momentum ResNets and neural ODEs
• Optimal shape and location of sensors for parabolic equations with random initial data
• A proposal on machine learning via dynamical systems
• Learning theory: stability is sufficient for generalization and necessary and sufficient for consistency of empirical risk minimization
• Control on the manifolds of mappings with a view to the deep learning
• Fractals in Probability and Analysis
• Extensions of Lipschitz mappings into a Hilbert space
• Control to Flocking of the Kinetic Cucker--Smale Model
• Exact Controllability, Stabilization and Perturbations for Distributed Systems
• A Bang-Bang Theorem with Bounds on the Number of Switchings
• On Hausdorff Dimension of Recurrent Net Fractals
• Weak Existence for the Semigeostrophic Equations Formulated as a Coupled Monge--Ampère/Transport Problem
• Simultaneous Exact Controllability and Some Applications
• Stable architectures for deep neural networks
• Approximate and Exact Controllability of the Continuity Equation with a Localized Vector Field
• Matrix Optimal Mass Transport: A Quantum Mechanical Approach
• Dimensions and Measures of Quasi Self-Similar Sets
• 10.1162/153244302760200704
• Turnpike in Lipschitz—nonlinear optimal control
• Deep Neural Networks, Generic Universal Interpolation, and Controlled ODEs
• Error bounds for approximations with deep ReLU neural networks in Ws,p norms
• Wavelet methods for pointwise regularity and local oscillations of functions
• Turnpike in optimal control of PDEs, ResNets, and beyond
• Optimal Transport
• Approximation by superpositions of a sigmoidal function
• Error bounds for approximation with neural networks
• The geometry of adversarial training in binary classification
• Sparsity in long-time control of neural ODEs
• Deep learning via dynamical systems: an approximation perspective

This page was built for publication: Neural ODE Control for Classification, Approximation, and Transport