arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2501.09097math.OCmath.FAmath.PR
收藏

概率测度空间上的最小二乘问题

Least-Squares Problem Over Probability Measure Space

Qin Li, Li Wang, Yunan Yang

AI总结:

本研究针对概率测度空间上的最小二乘变分问题展开,探究不同概率测度差异度量方式对极小化元的影响,发现φ-散度对应恢复条件分布、Wasserstein距离对应恢复边缘分布,完善了该类问题与线性空间最小二乘问题的类比体系。

摘要:

在本研究中,我们探究变分问题$$ρ_x^\u2009= \text{argmin}_{ρ_x} D(G\\#ρ_x, ρ_y)\\,, $$其中$D$用于量化两个概率测度之间的差异,$G$是将变量$x$映射为$y=G(x)$的前向算子。该问题可被视为其在线性空间(如欧几里得空间)中对应问题$\text{argmin}_x \\|G(x) - y\\|^2$的类比。与范数$\\|\cdot\\|$的选择会影响$\u211d^d$或其他线性空间中的最优解类似,该概率变分问题中的极小化元也依赖于$D$的选择。我们的研究结果表明:将$ϕ$-散度用作$D$可恢复$ρ_y$的条件分布,而采用Wasserstein距离则会恢复边缘分布。

英文摘要:

In this work, we investigate the variational problem $$ρ_x^\ast = \text{argmin}_{ρ_x} D(G\#ρ_x, ρ_y)\,, $$ where $D$ quantifies the difference between two probability measures, and ${G}$ is a forward operator that maps a variable $x$ to $y=G(x)$. This problem can be regarded as an analogue of its counterpart in linear spaces (e.g., Euclidean spaces), $\text{argmin}_x \|G(x) - y\|^2$. Similar to how the choice of norm $\|\cdot\|$ influences the optimizer in $\mathbb R^d$ or other linear spaces, the minimizer in the probabilistic variational problem also depends on the choice of $D$. Our findings reveal that using a $ϕ$-divergence for $D$ leads to the recovery of a conditional distribution of $ρ_y$, while employing the Wasserstein distance results in the recovery of a marginal distribution.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑