analiza aiqbench
This research establishes the first rigorous characterization of ReLU neural network approximation errors for infinite-dimensional functional inputs, linking accuracy directly to the joint decay of coordinate magnitudes and directional sensitivities. By proving that upper and lower bounds match under generalized exponential decay conditions, the study defines the nearly optimal stretched-exponential convergence rates achievable within a fixed network budget. These findings provide critical theoretical limits for deep learning models operating in high-dimensional or functional data regimes.
ReLU Neural Network Approximation to Smooth Functional Operator: Dimensional Decay and Error Analysis
arXiv:2609.15355v1 Announce Type: cross Abstract: We study the uniform approximation of smooth scalar-valued functionals on an infinite-dimensional separable Hilbert space by deep ReLU neural networks. Writing the functional input as $X(t)=\sum_{d\geq1}\xi_d\nu_d(t)$, we quantify the importance of coordinate $d$ through $w_ds_d$, where $s_d$ bounds the magnitude of the corresponding basis score and $w_d$ controls the directional Fr\'echet sensitivity of the target functional. Our constructive analysis combines coordinate truncation, anisotropic partitioning, local Taylor approximation, and ReLU network realization, while allowing unrestricted interactions among the retained coordinates. We establish a general nonasymptotic upper bound for the uniform approximation error and a complementary pseudo-dimension-based lower bound for the worst-case approximation error. Under generalized exponential coordinate decay $w_ds_d\asymp\exp(-cd^\rho)$, with $\rho>0$, the upper and lower bounds match at the leading order and thus yield the nearly optimal approximation rate, which is stretched-exponential in the logarithm of the network budget. This is the first work to characterize neural network approximation error for infinite-dimensional functional inputs explicitly through the joint dimensional decay of coordinate magnitudes and directional sensitivities.
Abstract:We study the uniform approximation of smooth scalar-valued functionals on an infinite-dimensional separable Hilbert space by deep ReLU neural networks. Writing the functional input as $X(t)=\sum_{d\geq1}\xi_d\nu_d(t)$, we quantify the importance of coordinate $d$ through $w_ds_d$, where $s_d$ bounds the magnitude of the corresponding basis score and $w_d$ controls the directional Fréchet sensitivity of the target functional. Our constructive analysis combines coordinate truncation, anisotropic partitioning, local Taylor approximation, and ReLU network realization, while allowing unrestricted interactions among the retained coordinates. We establish a general nonasymptotic upper bound for the uniform approximation error and a complementary pseudo-dimension-based lower bound for the worst-case approximation error. Under generalized exponential coordinate decay $w_ds_d\asymp\exp(-cd^\rho)$, with $\rho>0$, the upper and lower bounds match at the leading order and thus yield the nearly optimal approximation rate, which is stretched-exponential in the logarithm of the network budget. This is the first work to characterize neural network approximation error for infinite-dimensional functional inputs explicitly through the joint dimensional decay of coordinate magnitudes and directional sensitivities.
Machine Learning (stat.ML); Machine Learning (cs.LG)
arXiv:2609.15355 [stat.ML]
(or arXiv:2609.15355v1 [stat.ML] for this version)
https://doi.org/10.48550/arXiv.2609.15355 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 262 cuvinte
Subiecte conexe
Exploding variance of means of exponentials: least-squares to the rescue – Machine Learning Research Blog
francisbach.com · 27 sept. 2026
Approximation and learning of anisotropic and mixed smooth functions by deep ReLU neural networks
arXiv · 27 iul. 2026
Minimax rates for learning spectral Barron functions by deep ReLU neural networks
arXiv · 1 oct. 2026
SDC-GON: Singular Decomposition and Consistency-Regularized Green's Operator Networks for Solving Partial Differential Equations
arXiv · 22 sept. 2026