arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2503.11249cs.LGcs.AI
收藏

球面树切片Wasserstein距离

Spherical Tree-Sliced Wasserstein Distance

FPT Software AI Center(FPT软件人工智能中心);Movian AI, Vietnam(越南Movian人工智能公司);The Institute of Statistical Mathematics(统计数学研究所);RIKEN AIP(理化学研究所人工智能项目)

Viet-Hoang Tran, Thanh T. Chu, Khoi N. M. Nguyen, Trang Pham, Tam Le, Tan M. Nguyen

AI总结:

本研究受树切片最优传输启发,提出球面Radon变换与球面树结构,推导得到球面测度的高效正交不变度量STSW距离,经理论分析和多类数值实验验证了其性能。

摘要:

切片最优传输(Optimal Transport,OT)通过将输入测度的支撑集投影到一维直线上,再利用单变量OT的闭式表达式来降低OT的计算负担,从而简化了高维空间中的OT问题。近期,研究人员提出了树切片方法,用更复杂的结构——即树系统——来替代这些直线。该方法在保持低计算成本的同时,增强了切片OT捕捉积分域拓扑信息的能力。受此方法启发,本文针对支撑在球面上的测度,提出了树系统在OT问题中的适配方案。作为树系统上Radon变换变体的对应方法,我们提出了一种新型球面Radon变换,其采用名为球面树的新型积分域。通过利用该变换并借助球面树结构,我们推导出了球面上OT问题的闭式表达式。由此,我们得到了一种高效的球面测度度量,命名为球面树切片Wasserstein(Spherical Tree-Sliced Wasserstein, STSW)距离。我们开展了全面的理论分析,证明了球面树的拓扑性质,以及我们提出的Radon变换变体的良定义性和单射性,这使得该度量在球面测度之间具有正交不变性。最后,我们开展了包括梯度流和自监督学习在内的大量数值实验,将所提度量与近期基准方法进行对比,以评估其性能。

英文摘要:

Sliced Optimal Transport (OT) simplifies the OT problem in high-dimensional spaces by projecting supports of input measures onto one-dimensional lines and then exploiting the closed-form expression of the univariate OT to reduce the computational burden of OT. Recently, the Tree-Sliced method has been introduced to replace these lines with more intricate structures, known as tree systems. This approach enhances the ability to capture topological information of integration domains in Sliced OT while maintaining low computational cost. Inspired by this approach, in this paper, we present an adaptation of tree systems on OT problems for measures supported on a sphere. As a counterpart to the Radon transform variant on tree systems, we propose a novel spherical Radon transform with a new integration domain called spherical trees. By leveraging this transform and exploiting the spherical tree structures, we derive closed-form expressions for OT problems on the sphere. Consequently, we obtain an efficient metric for measures on the sphere, named Spherical Tree-Sliced Wasserstein (STSW) distance. We provide an extensive theoretical analysis to demonstrate the topology of spherical trees and the well-definedness and injectivity of our Radon transform variant, which leads to an orthogonally invariant distance between spherical measures. Finally, we conduct a wide range of numerical experiments, including gradient flows and self-supervised learning, to assess the performance of our proposed metric, comparing it to recent benchmarks.

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。

↑