Imagen de Portada

Why Cannot Neural Networks Master Extrapolation? Insights from Physical Laws

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Dakhmouche, Ramzi, Gorji, Hossein
Formato:	Preprint
Publicado:	2025
Materias:	Machine Learning Numerical Analysis Probability
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Can Linear Probes Measure LLM Uncertainty?
por: Dakhmouche, Ramzi, et al.
Publicado: (2025)

Unified Stochastic Framework for Neural Network Quantization and Pruning
por: Zhang, Haoyu, et al.
Publicado: (2024)

Optimal Estimation of Generic Dynamics by Path-Dependent Neural Jump ODEs
por: Krach, Florian, et al.
Publicado: (2022)

Monte Carlo Neural PDE Solver for Learning PDEs via Probabilistic Representation
por: Zhang, Rui, et al.
Publicado: (2023)

Sampling from Bayesian Neural Network Posteriors with Symmetric Minibatch Splitting Langevin Dynamics
por: Paulin, Daniel, et al.
Publicado: (2024)

Numerical and statistical analysis of NeuralODE with Runge-Kutta time integration
por: Ehrhardt, Emily C., et al.
Publicado: (2025)

Scale-Adaptive Generative Flows for Multiscale Scientific Data
por: Chen, Yifan, et al.
Publicado: (2025)

Modified Equations for Stochastic Optimization
por: Perko, Stefan
Publicado: (2025)

Solving stochastic partial differential equations using neural networks in the Wiener chaos expansion
por: Neufeld, Ariel, et al.
Publicado: (2024)

New Trends in the Stability of Sinkhorn Semigroups
por: Del Moral, Pierre, et al.
Publicado: (2026)

Deep learning based numerical approximation algorithms for stochastic partial differential equations
por: Beck, Christian, et al.
Publicado: (2020)

Convergence rates for gradient descent in the training of overparameterized artificial neural networks with piecewise affine activation
por: Jentzen, Arnulf, et al.
Publicado: (2021)

Correction to "Wasserstein distance estimates for the distributions of numerical approximations to ergodic stochastic differential equations"
por: Paulin, Daniel, et al.
Publicado: (2024)

Convex-Geometric Error Bounds for Positive-Weight Kernel Quadrature
por: Hayakawa, Satoshi
Publicado: (2026)

Multilevel Picard approximations and deep neural networks with ReLU, leaky ReLU, and softplus activation overcome the curse of dimensionality when approximating semilinear parabolic partial differential equations in $L^p$-sense
por: Neufeld, Ariel, et al.
Publicado: (2024)

Riemannian Langevin Dynamics: Strong Convergence of Geometric Euler-Maruyama Scheme
por: Zhan, Zhiyuan, et al.
Publicado: (2026)

Sampling via Stochastic Interpolants by Langevin-based Velocity and Initialization Estimation in Flow ODEs
por: Duan, Chenguang, et al.
Publicado: (2026)

Extending Path-Dependent NJ-ODEs to Noisy Observations and a Dependent Observation Framework
por: Andersson, William, et al.
Publicado: (2023)

A Mean Field Ansatz for Zero-Shot Weight Transfer
por: Chen, Xingyuan, et al.
Publicado: (2024)

Diffusion Model's Generalization Can Be Characterized by Inductive Biases toward a Data-Dependent Ridge Manifold
por: He, Ye, et al.
Publicado: (2026)

Expressivity of Bi-Lipschitz Normalizing Flows: A Score-Based Diffusion Perspective
por: Iske, Meira, et al.
Publicado: (2026)

Accelerating Multilevel Markov Chain Monte Carlo Using Machine Learning Models
por: Reddy, Sohail, et al.
Publicado: (2024)

Nonparametric Filtering, Estimation and Classification using Neural Jump ODEs
por: Heiss, Jakob, et al.
Publicado: (2024)

Analysis of kinetic Langevin Monte Carlo under the stochastic exponential Euler discretization from underdamped all the way to overdamped
por: Kim, Kyurae, et al.
Publicado: (2025)

Space-time deep neural network approximations for high-dimensional partial differential equations
por: Hornung, Fabian, et al.
Publicado: (2020)

Physics-Informed Inference Time Scaling for Solving High-Dimensional PDE via Defect Correction
por: Fan, Zexi, et al.
Publicado: (2025)

Gaussian Measures Conditioned on Nonlinear Observations: Consistency, MAP Estimators, and Simulation
por: Chen, Yifan, et al.
Publicado: (2024)

Single-seed generation of Brownian paths and integrals for adaptive and high order SDE solvers
por: Jelinčič, Andraž, et al.
Publicado: (2024)

Weak Physics Informed Neural Networks for Geometry Compatible Hyperbolic Conservation Laws on Manifolds
por: Zhou, Hanfei, et al.
Publicado: (2025)

Function Extrapolation with Neural Networks and Its Application for Manifolds
por: Hay, Guy, et al.
Publicado: (2024)

Neural Operators Can Play Dynamic Stackelberg Games
por: Alvarez, Guillermo, et al.
Publicado: (2024)

Gaussian Processes and Reproducing Kernels: Connections and Equivalences
por: Kanagawa, Motonobu, et al.
Publicado: (2025)

DeepMartingale: Duality of the Optimal Stopping Problem with Expressivity and High-Dimensional Hedging
por: Ye, Junyan, et al.
Publicado: (2025)

Data-driven approximation of transfer operators for mean-field stochastic differential equations
por: Ioannou, Eirini, et al.
Publicado: (2025)

Provable Diffusion Posterior Sampling for Bayesian Inversion
por: Chang, Jinyuan, et al.
Publicado: (2025)

Samplet limits and multiwavelets
por: Giacchi, Gianluca, et al.
Publicado: (2026)

Analysis of singular subspaces under random perturbations
por: Wang, Ke
Publicado: (2024)

Full error analysis of the random deep splitting method for nonlinear parabolic PDEs and PIDEs
por: Neufeld, Ariel, et al.
Publicado: (2024)

Langevin dynamics based algorithm e-TH$\varepsilon$O POULA for stochastic optimization problems with discontinuous stochastic gradient
por: Lim, Dong-Young, et al.
Publicado: (2022)

Non-asymptotic convergence analysis of the stochastic gradient Hamiltonian Monte Carlo algorithm with discontinuous stochastic gradient with applications to training of ReLU neural networks
por: Liang, Luxu, et al.
Publicado: (2024)