Safe Triplet Screening for Distance Metric Learning

被引:10
|
作者
Yoshida, Tomoki [1 ]
Takeuchi, Ichiro [2 ]
Karasuyama, Masayuki [3 ]
机构
[1] Nagoya Inst Technol, Nagoya, Aichi, Japan
[2] RIKEN Ctr Adv Intelligence Project, Natl Inst Mat Sci, Nagoya Inst Technol, Tsukuba, Ibaraki, Japan
[3] Japan Sci & Technol Agcy, Natl Inst Mat Sci, Nagoya Inst Technol, Kawaguchi, Saitama, Japan
关键词
metric learning; safe screening; convex optimization; DUAL APPROACH;
D O I
10.1145/3219819.3220037
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We study safe screening for metric learning. Distance metric learning can optimize a metric over a set of triplets, each one of which is defined by a pair of same class instances and an instance in a different class. However, the number of possible triplets is quite huge even for a small dataset. Our safe triplet screening identifies triplets which can be safely removed from the optimization problem without losing the optimality. Compared with existing safe screening studies, triplet screening is particularly significant because of (1) the huge number of possible triplets, and (2) the semi-definite constraint in the optimization. We derive several variants of screening rules, and analyze their relationships. Numerical experiments on benchmark datasets demonstrate the effectiveness of safe triplet screening.
引用
收藏
页码:2653 / 2662
页数:10
相关论文
共 50 条
  • [1] Safe Triplet Screening for Distance Metric Learning
    Yoshida, Tomoki
    Takeuchi, Ichiro
    Karasuyama, Masayuki
    NEURAL COMPUTATION, 2019, 31 (12) : 2432 - 2491
  • [2] Distance metric learning for ordinal classification based on triplet constraints
    Bac Nguyen
    Morell, Carlos
    De Baets, Bernard
    KNOWLEDGE-BASED SYSTEMS, 2018, 142 : 17 - 28
  • [3] A Theoretically Sound Upper Bound on the Triplet Loss for Improving the Efficiency of Deep Distance Metric Learning
    Do, Thanh-Toan
    Tran, Toan
    Reid, Ian
    Kumar, Vijay
    Hoang, Tuan
    Carneiro, Gustavo
    2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019), 2019, : 10396 - 10405
  • [4] Curvilinear Distance Metric Learning
    Chen, Shuo
    Luo, Lei
    Yang, Jian
    Gong, Chen
    Li, Jun
    Huang, Heng
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 32 (NIPS 2019), 2019, 32
  • [5] Sparse distance metric learning
    Choy, Tze
    Meinshausen, Nicolai
    COMPUTATIONAL STATISTICS, 2014, 29 (3-4) : 515 - 528
  • [6] Sparse distance metric learning
    Tze Choy
    Nicolai Meinshausen
    Computational Statistics, 2014, 29 : 515 - 528
  • [7] Distance metric learning with the Universum
    Bac Nguyen
    Morell, Carlos
    De Baets, Bernard
    PATTERN RECOGNITION LETTERS, 2017, 100 : 37 - 43
  • [8] Distance metric learning by minimal distance maximization
    Yu, Yaoliang
    Jiang, Jiayan
    Zhang, Liming
    PATTERN RECOGNITION, 2011, 44 (03) : 639 - 649
  • [9] Deep Metric Learning with Hierarchical Triplet Loss
    Ge, Weifeng
    Huang, Weilin
    Dong, Dengke
    Scott, Matthew R.
    COMPUTER VISION - ECCV 2018, PT VI, 2018, 11210 : 272 - 288
  • [10] Deep Metric Learning Using Triplet Network
    Hoffer, Elad
    Ailon, Nir
    SIMILARITY-BASED PATTERN RECOGNITION, SIMBAD 2015, 2015, 9370 : 84 - 92