On Assignment Problems Related to Gromov-Wasserstein Distances on the Real Line

被引:2
|
作者
Beinert, Robert [1 ]
Heiss, Cosmas [1 ]
Steidl, Gabriele [1 ]
机构
[1] Tech Univ Berlin, Inst Math, D-10623 Berlin, Germany
来源
SIAM JOURNAL ON IMAGING SCIENCES | 2023年 / 16卷 / 02期
关键词
point assignment problem; Gromov-Wasserstein distance; Gromov--Monge formulation; sliced Gromv-Wasserstein;
D O I
10.1137/22M1497808
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Let x1 < \cdot \cdot \cdot < xn and y1 < \cdot \cdot \cdot < yn, n \in N, be real numbers. We show by an example that the assignment problem max \sigma\in Sn \sumn 1 F\sigma(x, y) := 2 i,k =1 lxi -xkl\alphaly\sigma(i) -y\sigma(k)l\alpha, \alpha > 0, is in general neither solved by the identical permutation (id) nor the anti-identical permutation (a-id) if n > 2 + 2\alpha. Indeed the above maximum can be, depending on the number of points, arbitrarily far away from Fid(x, y) and FFid(x, y). The motivation to deal with such assignment problems came from their relation to Gromov-Wasserstein distances, which have recently received a lot of attention in imaging and shape analysis.
引用
收藏
页码:1028 / 1032
页数:5
相关论文
共 50 条
  • [21] Gromov-Wasserstein Averaging of Kernel and Distance Matrices
    Peyre, Gabriel
    Cuturi, Marco
    Solomon, Justin
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 48, 2016, 48
  • [22] Entropic Gromov-Wasserstein between Gaussian Distributions
    Khang Le
    Dung Le
    Huy Nguyen
    Dat Do
    Tung Pham
    Nhat Ho
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 162, 2022,
  • [23] On the Existence of Monge Maps for the Gromov-Wasserstein Problem
    Dumont, Theo
    Lacombe, Theo
    Vialard, Francois-Xavier
    FOUNDATIONS OF COMPUTATIONAL MATHEMATICS, 2024, 25 (2) : 463 - 510
  • [24] Generalized Spectral Clustering via Gromov-Wasserstein Learning
    Chowdhury, Samir
    Needham, Tom
    Proceedings of Machine Learning Research, 2021, 130 : 712 - 720
  • [25] Learning Graphons via Structured Gromov-Wasserstein Barycenters
    Xu, Hongteng
    Luo, Dixin
    Carin, Lawrence
    Zha, Hongyuan
    THIRTY-FIFTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, THIRTY-THIRD CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE AND THE ELEVENTH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2021, 35 : 10505 - 10513
  • [26] Gromov-Wasserstein Multi-modal Alignment and Clustering
    Gong, Fengjiao
    Nie, Yuzhou
    Xu, Hongteng
    PROCEEDINGS OF THE 31ST ACM INTERNATIONAL CONFERENCE ON INFORMATION AND KNOWLEDGE MANAGEMENT, CIKM 2022, 2022, : 603 - 613
  • [27] Scalable Gromov-Wasserstein Learning for Graph Partitioning and Matching
    Xu, Hongteng
    Luo, Dixin
    Carin, Lawrence
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 32 (NIPS 2019), 2019, 32
  • [28] Generalized Spectral Clustering via Gromov-Wasserstein Learning
    Chowdhury, Samir
    Needham, Tom
    24TH INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS (AISTATS), 2021, 130 : 712 - +
  • [29] Orthogonal Gromov-Wasserstein Discrepancy with Efficient Lower Bound
    Jin, Hongwei
    Yu, Zishun
    Zhang, Xinhua
    UNCERTAINTY IN ARTIFICIAL INTELLIGENCE, VOL 180, 2022, 180 : 917 - 927
  • [30] Scalable Gromov-Wasserstein Based Comparison of Biological Time Series
    Kravtsova, Natalia
    McGee II, Reginald L. L.
    Dawes, Adriana T.
    BULLETIN OF MATHEMATICAL BIOLOGY, 2023, 85 (08)