Adversarial Risk via Optimal Transport and Optimal Couplings

被引:14
|
作者
Pydi, Muni Sreenivas [1 ]
Jog, Varun [2 ]
机构
[1] Univ Wisconsin, Dept Elect & Comp Engn, 1415 Johnson Dr, Madison, WI 53706 USA
[2] Univ Cambridge, Dept Pure Math & Math Stat, Cambridge CB3 0WB, England
关键词
Couplings; Standards; Measurement; Kernel; Perturbation methods; Loss measurement; Q measurement; Machine learning; statistical learning; robustness; couplings; information theory; DEEP NEURAL-NETWORKS; ROBUST; ALGORITHMS; GO;
D O I
10.1109/TIT.2021.3100107
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Modern machine learning algorithms perform poorly on adversarially manipulated data. Adversarial risk quantifies the error of classifiers in adversarial settings; adversarial classifiers minimize adversarial risk. In this paper, we analyze adversarial risk and adversarial classifiers from an optimal transport perspective. We show that the optimal adversarial risk for binary classification with 0-1 loss is determined by an optimal transport cost between the probability distributions of the two classes. We develop optimal transport plans (probabilistic couplings) for univariate distributions such as the normal, the uniform, and the triangular distribution. We also derive optimal adversarial classifiers in these settings. Our analysis leads to algorithm-independent fundamental limits on adversarial risk, which we calculate for several real-world datasets. We extend our results to general loss functions under convexity and smoothness assumptions.
引用
收藏
页码:6031 / 6052
页数:22
相关论文
共 50 条
  • [11] Optimal Markovian Couplings
    陈木法
    ProgressinNaturalScience, 1994, (03) : 111 - 113
  • [12] On optimal multivariate couplings
    Ruschendorf, L
    Uckelmann, L
    DISTRIBUTIONS WITH GIVEN MARGINALS AND MOMENT PROBLEMS, 1997, : 261 - 273
  • [13] Adaptive mesh methods on compact manifolds via Optimal Transport and Optimal Information Transport
    Turnquist A.G.R.
    Journal of Computational Physics, 2024, 500
  • [14] Metric Reconstruction Via Optimal Transport
    Adamaszek, Michal
    Adams, Henry
    Frick, Florian
    SIAM JOURNAL ON APPLIED ALGEBRA AND GEOMETRY, 2018, 2 (04): : 597 - 619
  • [15] Revisiting colocalization via optimal transport
    Wang, Shulei
    Yuan, Ming
    NATURE COMPUTATIONAL SCIENCE, 2021, 1 (03): : 177 - 178
  • [16] Revisiting colocalization via optimal transport
    Shulei Wang
    Ming Yuan
    Nature Computational Science, 2021, 1 : 177 - 178
  • [17] Model Fusion via Optimal Transport
    Singh, Sidak Pal
    Jaggi, Martin
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 33, NEURIPS 2020, 2020, 33
  • [18] Natural gradient via optimal transport
    Li W.
    Montúfar G.
    Information Geometry, 2018, 1 (2) : 181 - 214
  • [19] EQUILIBRIUM SELECTION VIA OPTIMAL TRANSPORT
    Chow, Shui-Nee
    Li, Wuchen
    Lu, Jun
    Zhou, Haomin
    SIAM JOURNAL ON APPLIED MATHEMATICS, 2020, 80 (01) : 142 - 159
  • [20] Margin-aware Adversarial Domain Adaptation with Optimal Transport
    Dhouib, Sofien
    Redko, Ievgen
    Lartizien, Carole
    25TH AMERICAS CONFERENCE ON INFORMATION SYSTEMS (AMCIS 2019), 2019,