Efficient Robust Optimal Transport with Application to Multi-Label Classification

被引:2
作者
Jawanpuria, Pratik [1 ]
Satyadev, N. T., V [2 ]
Mishra, Bamdev [1 ]
机构
[1] Microsoft, Hyderabad, India
[2] Vayve Technol, Mumbai, Maharashtra, India
来源
2021 60TH IEEE CONFERENCE ON DECISION AND CONTROL (CDC) | 2021年
关键词
D O I
10.1109/CDC45484.2021.9683259
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Optimal transport (OT) is a powerful geometric tool for comparing two distributions and has been employed in various machine learning applications. In this work, we propose a novel OT formulation that takes feature correlations into account while learning the transport plan between two distributions. We model the feature-feature relationship via a symmetric positive semi-definite Mahalanobis metric in the OT cost function. For a certain class of regularizers on the metric, we show that the optimization strategy can be considerably simplified by exploiting the problem structure. For high-dimensional data, we additionally propose suitable low-dimensional modeling of the Mahalanobis metric. Overall, we view the resulting optimization problem as a non-linear OT problem, which we solve using the Frank-Wolfe algorithm. Empirical results on the discriminative learning setting, such as tag prediction and multi-class classification, illustrate the good performance of our approach.
引用
收藏
页码:1490 / 1495
页数:6
相关论文
共 26 条
[11]  
Jawanpuria P, 2020, 58TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2020), P3052
[12]  
Kantorovitch L, 1942, CR ACAD SCI URSS, V37, P199
[13]  
Kolouri Soheil, 2019, NeurIPS
[14]  
Lampert CH, 2009, PROC CVPR IEEE, P951, DOI 10.1109/CVPRW.2009.5206594
[15]  
Nath J. S., 2020, NEURIPS
[16]  
Paty F.-P., 2020, ICML
[17]  
Paty F.-P., 2019, ICML
[18]   Computational Optimal Transport [J].
Peyre, Gabriel ;
Cuturi, Marco .
FOUNDATIONS AND TRENDS IN MACHINE LEARNING, 2019, 11 (5-6) :355-607
[19]  
Qi G.-J., 2009, P 26 ANN INT C MACHI, P841
[20]  
Rosales R., 2006, KDD