Deep Neural Tangent Kernel and Laplace Kernel Have the Same RKHS
Lists
Only you see this
Abstract
We prove that the reproducing kernel Hilbert spaces (RKHS) of a deep neural tangent kernel and the Laplace kernel include the same set of functions, when both kernels are restricted to the sphere $\mathbb{S}^{d-1}$. Additionally, we prove that the exponential power kernel with a smaller power (making the kernel less smooth) leads to a larger RKHS, when it is restricted to the sphere $\mathbb{S}^{d-1}$ and when it is defined on the entire $\mathbb{R}^d$.
BibTeX
@inproceedings{chen2021deep,
author = {Lin Chen and Sheng Xu},
title = {{Deep Neural Tangent Kernel and Laplace Kernel Have the Same RKHS}},
booktitle = {International Conference on Learning Representations},
year = {2021},
url = {https://openreview.net/forum?id=vK9WrZ0QYQ},
}Loading related content...