Non-asymptotic spectral bounds on the $\varepsilon$-entropy of kernel classes
Fuente:
arXiv
Salvato in:
| Autore principale: | |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2022
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909442521432064 |
|---|---|
| author | Takhanov, Rustem |
| author_facet | Takhanov, Rustem |
| contents | Let $K: \boldsymbolΩ\times \boldsymbolΩ$ be a continuous Mercer kernel defined on a compact subset of ${\mathbb R}^n$ and $\mathcal{H}_K$ be the reproducing kernel Hilbert space (RKHS) associated with $K$. Given a finite measure $ν$ on $\boldsymbolΩ$, we investigate upper and lower bounds on the $\varepsilon$-entropy of the unit ball of $\mathcal{H}_K$ in the space $L_p(ν)$. This topic is an important direction in the modern statistical theory of kernel-based methods.
We prove sharp upper and lower bounds for $p\in [1,+\infty]$. For $p\in [1,2]$, the upper bounds are determined solely by the eigenvalue behaviour of the corresponding integral operator $ϕ\to \int_{\boldsymbolΩ} K(\cdot,{\mathbf y})ϕ({\mathbf y})dν({\mathbf y})$. In constrast, for $p>2$, the bounds additionally depend on the convergence rate of the truncated Mercer series to the kernel $K$ in the $L_p(ν)$-norm.
We discuss a number of consequences of our bounds and show that they are substantially tighter than previous bounds for general kernels. Furthermore, for specific cases, such as zonal kernels and the Gaussian kernel on a box, our bounds are asymptotically tight as $\varepsilon\to +0$. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2204_04512 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | Non-asymptotic spectral bounds on the $\varepsilon$-entropy of kernel classes Takhanov, Rustem Machine Learning Functional Analysis 41A46, 47B06, 46E22, 68Q32 Let $K: \boldsymbolΩ\times \boldsymbolΩ$ be a continuous Mercer kernel defined on a compact subset of ${\mathbb R}^n$ and $\mathcal{H}_K$ be the reproducing kernel Hilbert space (RKHS) associated with $K$. Given a finite measure $ν$ on $\boldsymbolΩ$, we investigate upper and lower bounds on the $\varepsilon$-entropy of the unit ball of $\mathcal{H}_K$ in the space $L_p(ν)$. This topic is an important direction in the modern statistical theory of kernel-based methods. We prove sharp upper and lower bounds for $p\in [1,+\infty]$. For $p\in [1,2]$, the upper bounds are determined solely by the eigenvalue behaviour of the corresponding integral operator $ϕ\to \int_{\boldsymbolΩ} K(\cdot,{\mathbf y})ϕ({\mathbf y})dν({\mathbf y})$. In constrast, for $p>2$, the bounds additionally depend on the convergence rate of the truncated Mercer series to the kernel $K$ in the $L_p(ν)$-norm. We discuss a number of consequences of our bounds and show that they are substantially tighter than previous bounds for general kernels. Furthermore, for specific cases, such as zonal kernels and the Gaussian kernel on a box, our bounds are asymptotically tight as $\varepsilon\to +0$. |
| title | Non-asymptotic spectral bounds on the $\varepsilon$-entropy of kernel classes |
| topic | Machine Learning Functional Analysis 41A46, 47B06, 46E22, 68Q32 |
| url | https://arxiv.org/abs/2204.04512 |