发表机构
Interdisziplinäres Zentrum für wissenschaftliches Rechnen, Universität Heidelberg(海德堡大学跨学科计算科学中心)
机构由 AI 辅助整理,请以论文原文为准。AI 中文总结
本文提出贝叶斯框架学习无限维空间间的非线性算子,证明高斯先验下后验存在性,并建立后验均值的代数收敛速率及极小极大最优性,应用于Darcy流算子。
AI 中文摘要
我们发展了一个用于学习无限维空间之间非线性算子的贝叶斯框架。给定可分希尔伯特空间之间的映射 $g_0:\mathcal{X}\to\mathcal{Y}$,我们研究从 $n\in\mathbb{N}$ 个带噪声的输入-输出对 $(\boldsymbol{X},\boldsymbol{Z})=(X_i,Z_i)_{i=1}^n$ 中恢复 $g_0$,其中 $Z_i= g_0 (X_i ) + E_i$。这里 $X_i\in\mathcal{X}$ 是随机抽取的“设计”点,位于 $\mathcal X$ 的一个紧子集中,而 $E_i$ 被假定为从以 $\mathcal{Y}$ 为索引的高斯白噪声过程中独立同分布抽取。对于任何支撑在连续算子空间上的“算子值”先验 $\mathbb{P}_G$,我们证明了后验 $\mathbb P_{G|(\boldsymbol{X},\boldsymbol{Z})}$ 作为正则条件分布的存在性,并给出了其 Radon-Nikodym 导数的刻画。对于高斯先验,我们建立了后验均值向真实值的代数(关于样本量 $n$)收敛速率;这对应于岭正则化核估计器。此外,我们证明了当先验的光滑性与真实值的光滑性匹配时,后验均值在超矩形上是极小极大最优的(至多差对数因子)。为了说明我们分析的适用性,我们推导了 Darcy 流解算子的显式学习速率。
英文摘要
We develop a Bayesian framework for learning nonlinear operators between infinite-dimensional spaces. Given a map $g_0:\mathcal{X}\to\mathcal{Y}$ between separable Hilbert spaces, we study the recovery of $g_0$ from $n\in\mathbb{N}$ noisy input-output pairs $(\boldsymbol{X},\boldsymbol{Z})=(X_i,Z_i)_{i=1}^n$ with $Z_i= g_0 (X_i ) + E_i$. Here the $X_i\in\mathcal{X}$ are randomly drawn 'design' points in a compact subset of $\mathcal X$, and the $E_i$ are assumed to be i.i.d. draws from a Gaussian white noise process indexed by $\mathcal{Y}$. For any 'operator-valued' prior $\mathbb{P}_G$ supported on the space of continuous operators, we show existence of the posterior $\mathbb P_{G|(\boldsymbol{X},\boldsymbol{Z})}$ as a regular conditional distribution, and provide a characterization of its Radon-Nikodym derivative. For Gaussian priors, we establish algebraic (in the sample size $n$) convergence rates for the posterior mean towards the ground truth; this corresponds to a ridge regularized kernel estimator. Moreover, we show that the posterior mean is minimax optimal (up to logarithmic factors) over hyperrectangles when the smoothness of the prior matches that of the ground truth. To illustrate the applicability of our analysis, we derive explicit learning rates for the Darcy flow solution operator.