Taming Binarized Neural Networks and Mixed-Integer Programs

被引:0
|
作者
Aspman, Johannes [1 ]
Korpas, Georgios [1 ,2 ]
Marecek, Jakub [1 ]
机构
[1] Czech Tech Univ, Dept Comp Sci, Prague, Czech Republic
[2] HSBC Holdings, HSBC Lab, Innovat & Ventures, London, England
关键词
FIELD;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
There has been a great deal of recent interest in binarized neural networks, especially because of their explainability. At the same time, automatic differentiation algorithms such as back-propagation fail for binarized neural networks, which limits their applicability. We show that binarized neural networks admit a tame representation by reformulating the problem of training binarized neural networks as a subadditive dual of a mixed-integer program, which we show to have nice properties. This makes it possible to use the framework of Bolte et al. for implicit differentiation, which offers the possibility for practical implementation of backpropagation in the context of binarized neural networks. This approach could also be used for a broader class of mixed-integer programs, beyond the training of binarized neural networks, as encountered in symbolic approaches to AI and beyond.
引用
收藏
页码:10935 / 10943
页数:9
相关论文
共 50 条
  • [21] A Comparison of Verification Methods for Neural-Network Controllers Using Mixed-Integer Programs
    Dubach, Marcel
    Ducard, Guillaume J. J.
    2022 7TH INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION ENGINEERING, ICRAE, 2022, : 43 - 48
  • [22] A Framework for Globally Optimizing Mixed-Integer Signomial Programs
    Ruth Misener
    Christodoulos A. Floudas
    Journal of Optimization Theory and Applications, 2014, 161 : 905 - 932
  • [23] Coefficient strengthening: a tool for reformulating mixed-integer programs
    Andersen, Kent
    Pochet, Yves
    MATHEMATICAL PROGRAMMING, 2010, 122 (01) : 121 - 154
  • [24] Global solution of nonlinear mixed-integer bilevel programs
    Mitsos, Alexander
    JOURNAL OF GLOBAL OPTIMIZATION, 2010, 47 (04) : 557 - 582
  • [25] Irreducible Infeasible Sets in Convex Mixed-Integer Programs
    Wiesława T. Obuchowska
    Journal of Optimization Theory and Applications, 2015, 166 : 747 - 766
  • [26] Foundation-penalty cuts for mixed-integer programs
    Glover, F
    Sherali, HD
    OPERATIONS RESEARCH LETTERS, 2003, 31 (04) : 245 - 253
  • [27] Neural benders decomposition for mixed-integer programming
    Monemi, Rahimeh Neamatian
    Gelareh, Shahin
    Maculan, Nelson
    Dai, Yu-Hong
    TOP, 2024,
  • [28] A DC Programming Approach for Mixed-Integer Linear Programs
    Niu, Yi-Shuai
    Dinh, Tao Pham
    MODELLING, COMPUTATION AND OPTIMIZATION IN INFORMATION SYSTEMS AND MANAGEMENT SCIENCES, PROCEEDINGS, 2008, 14 : 244 - 253
  • [29] Irreducible Infeasible Sets in Convex Mixed-Integer Programs
    Obuchowska, Wiesawa T.
    JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS, 2015, 166 (03) : 747 - 766
  • [30] Extending the QCR method to general mixed-integer programs
    Billionnet, Alain
    Elloumi, Sourour
    Lambert, Amelie
    MATHEMATICAL PROGRAMMING, 2012, 131 (1-2) : 381 - 401