跳到正文
返回论文列表
stat.ME提交于 已译

FréchetFlow:面向非欧几里得数据的流匹配

Fr\'echetFlow: Flow Matching for Non-Euclidean Data

Yidong Zhou · Muqing Cui · Hans-Georg M\"uller

中文摘要

概率分布、树和协方差矩阵等科学观测数据需要尊重其非欧几里得几何结构的生成模型。流匹配虽已通过测地线和速度扩展到流形上,但一般度量空间并不具备此类流形结构。提出 FréchetFlow,一种面向度量空间取值数据的生成建模框架,既不需要线性运算,也不需要光滑流形结构。该方法的关键特点是利用成熟的欧几里得流学习到一组地标(landmark)距离的联合分布,再根据生成的地标表示重建新对象。重建步骤借鉴 Fréchet 回归,使用参考对象的加权 Fréchet 均值。建立了 Wasserstein 误差界,将生成对象分布的精度与地标放置、流近似和重建误差联系起来,并在有限二阶矩下证明一致性。由此得到的收敛速率取决于目标对象的度量复杂度、尾部矩和拟合精度。实验表明,在欧几里得模拟中精度与已有流匹配方法相当,在多个科学数据集上 Wasserstein 误差低于专门定制的方法。在各类模拟设置中,生成误差随样本量增大而下降。

关键要点

  1. 01问题:一般度量空间缺乏流形结构,现有流匹配难以直接应用
  2. 02方法:用欧几里得流学习地标距离分布,再经加权 Fréchet 均值重建对象
  3. 03结果:给出 Wasserstein 误差界与有限二阶矩下的一致性及收敛速率
  4. 04结果:多个科学数据集上 Wasserstein 误差低于定制方法,模拟中与流匹配相当
  5. 05局限:收敛速率依赖度量复杂度、尾部矩与拟合精度

解读

尚无解读。

原始英文摘要

arXiv:2611.15220v1 Announce Type: cross Abstract: Scientific observations such as probability distributions, trees and covariance matrices call for generative models that respect their non-Euclidean geometry. While flow matching has been extended to manifolds by utilizing geodesics and velocities, such manifold structures are not available in a general metric space. We introduce Fr\'echetFlow, a generative modeling framework for metric space-valued data that requires neither linear operations nor a smooth manifold structure. A key feature of the proposed approach is that we utilize well-established Euclidean flows to learn the joint distribution of distances to a set of landmarks, whereupon new objects are reconstructed from generated landmark representations. The reconstruction step utilizes weighted Fr\'echet means of reference objects, drawing on Fr\'echet regression. We establish Wasserstein error bounds that connect accuracy of the generated object distributions to the placement of landmarks, flow approximation and reconstruction error and prove consistency under finite second moments. This leads to convergence rates that depend on the metric complexity of the target objects, tail moments and fitting accuracy. Experiments show competitive accuracy with established flow matching methods in Euclidean simulations and lower Wasserstein errors than specifically tailored methods for several scientific datasets. Across simulation settings, we find that generation error decreases as sample size increases.

同方向论文 · stat.ME

查看全部 →