Global quantitative robustness of regression feed-forward neural networks

被引:0
|
作者
Werner, Tino [1 ]
机构
[1] Institute for Mathematics, Carl von Ossietzky Universität Oldenburg, Carl-von-Ossietzky-Strasse 9-11, Lower Saxony, Oldenburg,26129, Germany
关键词
Feedforward neural networks;
D O I
10.1007/s00521-024-10289-w
中图分类号
学科分类号
摘要
Neural networks are an indispensable model class for many complex learning tasks. Despite the popularity and importance of neural networks and many different established techniques from literature for stabilization and robustification of the training, the classical concepts from robust statistics have rarely been considered so far in the context of neural networks. Therefore, we adapt the notion of the regression breakdown point to regression neural networks and compute the breakdown point for different feed-forward network configurations and contamination settings. In an extensive simulation study, we compare the performance, measured by the out-of-sample loss, by a proxy of the breakdown rate and by the training steps, of non-robust and robust regression feed-forward neural networks in a plethora of different configurations. The results indeed motivate to use robust loss functions for neural network training.
引用
收藏
页码:19967 / 19988
页数:21
相关论文
共 50 条
  • [1] Probabilistic robustness estimates for feed-forward neural networks
    Couellan, Nicolas
    NEURAL NETWORKS, 2021, 142 : 138 - 147
  • [2] Feed-forward neural networks
    Bebis, George
    Georgiopoulos, Michael
    IEEE Potentials, 1994, 13 (04): : 27 - 31
  • [3] A training-time analysis of robustness in feed-forward neural networks
    Alippi, C
    Sana, D
    Scotti, F
    2004 IEEE INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOLS 1-4, PROCEEDINGS, 2004, : 2853 - 2858
  • [4] Patterns of synchrony for feed-forward and auto-regulation feed-forward neural networks
    Aguiar, Manuela A. D.
    Dias, Ana Paula S.
    Ferreira, Flora
    CHAOS, 2017, 27 (01)
  • [5] Feed-forward Neural Networks with Trainable Delay
    Ji, Xunbi A.
    Molnar, Tamas G.
    Avedisov, Sergei S.
    Orosz, Gabor
    LEARNING FOR DYNAMICS AND CONTROL, VOL 120, 2020, 120 : 127 - 136
  • [6] On lateral connections in feed-forward neural networks
    Kothari, R
    Agyepong, K
    ICNN - 1996 IEEE INTERNATIONAL CONFERENCE ON NEURAL NETWORKS, VOLS. 1-4, 1996, : 13 - 18
  • [7] Optimizing dense feed-forward neural networks
    Balderas, Luis
    Lastra, Miguel
    Benitez, Jose M.
    NEURAL NETWORKS, 2024, 171 : 229 - 241
  • [8] Maximizing the margin with Feed-Forward Neural Networks
    Romero, E
    Alquézar, R
    PROCEEDING OF THE 2002 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOLS 1-3, 2002, : 743 - 748
  • [9] Internal workings of feed-forward neural networks
    Zhang, QJ
    Stanley, SJ
    Smith, DW
    JOURNAL OF ENVIRONMENTAL ENGINEERING AND SCIENCE, 2004, 3 : S1 - S12
  • [10] Hybridization of the probabilistic neural networks with feed-forward neural networks for forecasting
    Khashei, Mehdi
    Bijari, Mehdi
    ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE, 2012, 25 (06) : 1277 - 1288