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

被引:3
作者
Beinert, Robert [1 ]
Heiss, Cosmas [1 ]
Steidl, Gabriele [1 ]
机构
[1] Tech Univ Berlin, Inst Math, D-10623 Berlin, Germany
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 17 条
  • [1] On a Linear Gromov-Wasserstein Distance
    Beier, Florian
    Beinert, Robert
    Steidl, Gabriele
    [J]. IEEE TRANSACTIONS ON IMAGE PROCESSING, 2022, 31 : 7292 - 7305
  • [2] Subspace Detours Meet Gromov-Wasserstein
    Bonet, Clement
    Vayer, Titouan
    Courty, Nicolas
    Septier, Francois
    Drumetz, Lucas
    [J]. ALGORITHMS, 2021, 14 (12)
  • [3] CHAPEL L., 2020, P ADV NEUR INF PROC
  • [4] Gromov-Wasserstein Averaging in a Riemannian Framework
    Chowdhury, Samir
    Needham, Tom
    [J]. 2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW 2020), 2020, : 3876 - 3884
  • [5] Chowdhury S, 2021, PR MACH LEARN RES, V130, P712
  • [6] Hur Y, 2023, Arxiv, DOI arXiv:2109.14090
  • [7] Jin HW, 2022, PR MACH LEARN RES, V180, P917
  • [8] Sampled Gromov Wasserstein
    Kerdoncuff, Tanguy
    Emonet, Remi
    Sebban, Marc
    [J]. MACHINE LEARNING, 2021, 110 (08) : 2151 - 2186
  • [9] Gromov-Wasserstein Distances and the Metric Approach to Object Matching
    Memoli, Facundo
    [J]. FOUNDATIONS OF COMPUTATIONAL MATHEMATICS, 2011, 11 (04) : 417 - 487
  • [10] Nadjahi K., 2021, ADV NEUR IN, V34