.png){kind=icon}
Return
Provable wavelet-based neural approximation
.png){kind=icon}
OA
AI
DOI：10.1016/j.amc.2025.129821.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• We investigate neural network approximations with a wavelet-based framework via the wavelet frame theory on spaces of homogeneous type. • Our analysis provides sufficient conditions on activation functions that ensure the approximations with explicit error estimates, including the case of oscillatory activations. • We establish a generalized approximation result for non-smooth activations, where the error is controlled by their distance from smooth activations.
Keywords:
Neural networks
Function approximation
Wavelet frames
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
3.4
Papers:
2.3W
Citations:
3.3W
Organization
Cited Papers
No cited papers available