On the Optimality of Misspecified Spectral Algorithms
Zhang, Haobo, Li, Yicheng, Lin, Qian
–arXiv.org Artificial Intelligence
In the misspecified spectral algorithms problem, researchers usually assume the underground true function $f_{\rho}^{*} \in [\mathcal{H}]^{s}$, a less-smooth interpolation space of a reproducing kernel Hilbert space (RKHS) $\mathcal{H}$ for some $s\in (0,1)$. The existing minimax optimal results require $\|f_{\rho}^{*}\|_{L^{\infty}}<\infty$ which implicitly requires $s > \alpha_{0}$ where $\alpha_{0}\in (0,1)$ is the embedding index, a constant depending on $\mathcal{H}$. Whether the spectral algorithms are optimal for all $s\in (0,1)$ is an outstanding problem lasting for years. In this paper, we show that spectral algorithms are minimax optimal for any $\alpha_{0}-\frac{1}{\beta} < s < 1$, where $\beta$ is the eigenvalue decay rate of $\mathcal{H}$. We also give several classes of RKHSs whose embedding index satisfies $ \alpha_0 = \frac{1}{\beta} $. Thus, the spectral algorithms are minimax optimal for all $s\in (0,1)$ on these RKHSs.
arXiv.org Artificial Intelligence
Aug-8-2023
- Country:
- North America > United States
- Massachusetts (0.04)
- New York > New York County
- New York City (0.14)
- Europe
- Italy (0.04)
- United Kingdom > England
- Cambridgeshire > Cambridge (0.04)
- Asia > China
- North America > United States
- Genre:
- Research Report (0.63)
- Technology: