Provable Preimage Under-Approximation for Neural Networks

被引:3
|
作者
Zhang, Xiyue [1 ]
Wang, Benjie [1 ]
Kwiatkowska, Marta [1 ]
机构
[1] Univ Oxford, Dept Comp Sci, Oxford, England
来源
TOOLS AND ALGORITHMS FOR THE CONSTRUCTION AND ANALYSIS OF SYSTEMS, PT III, TACAS 2024 | 2024年 / 14572卷
关键词
VERIFICATION;
D O I
10.1007/978-3-031-57256-2_1
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
Neural network verification mainly focuses on local robustness properties, which can be checked by bounding the image (set of outputs) of a given input set. However, often it is important to know whether a given property holds globally for the input domain, and if not then for what proportion of the input the property is true. To analyze such properties requires computing preimage abstractions of neural networks. In this work, we propose an efficient anytime algorithm for generating symbolic under-approximations of the preimage of any polyhedron output set for neural networks. Our algorithm combines a novel technique for cheaply computing polytope preimage under-approximations using linear relaxation, with a carefully-designed refinement procedure that iteratively partitions the input region into subregions using input and ReLU splitting in order to improve the approximation. Empirically, we validate the efficacy of our method across a range of domains, including a high-dimensional MNIST classification task beyond the reach of existing preimage computation methods. Finally, as use cases, we showcase the application to quantitative verification and robustness analysis. We present a sound and complete algorithm for the former, which exploits our disjoint union of polytopes representation to provide formal guarantees. For the latter, we find that our method can provide useful quantitative information even when standard verifiers cannot verify a robustness property.
引用
收藏
页码:3 / 23
页数:21
相关论文
共 50 条
  • [41] Approximation by neural networks is not continuous
    Kainen, PC
    Kurková, V
    Vogt, A
    NEUROCOMPUTING, 1999, 29 (1-3) : 47 - 56
  • [42] Piecewise approximation and neural networks
    Revayova, Martina
    Torok, Csaba
    KYBERNETIKA, 2007, 43 (04) : 547 - 559
  • [43] Approximation theory and neural networks
    Mhaskar, HN
    WAVELETS AND ALLIED TOPICS, 2001, : 247 - 289
  • [44] Neural networks and approximation theory
    Mhaskar, HN
    NEURAL NETWORKS, 1996, 9 (04) : 721 - 722
  • [45] Uniform approximation by neural networks
    Makovoz, Y
    JOURNAL OF APPROXIMATION THEORY, 1998, 95 (02) : 215 - 228
  • [46] Simultaneous approximation with neural networks
    Dingankar, AT
    Phatak, DS
    IJCNN 2000: PROCEEDINGS OF THE IEEE-INNS-ENNS INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOL IV, 2000, : 232 - 237
  • [47] Pointwise approximation for neural networks
    Cao, FL
    Xu, ZB
    Li, YM
    ADVANCES IN NEURAL NETWORKS - ISNN 2005, PT 1, PROCEEDINGS, 2005, 3496 : 39 - 44
  • [48] NEURAL NETWORKS FOR LOCALIZED APPROXIMATION
    CHUI, CK
    LI, X
    MHASKAR, HN
    MATHEMATICS OF COMPUTATION, 1994, 63 (208) : 607 - 623
  • [49] NEURAL NETWORKS AND THE APPROXIMATION THEORY
    Enachescu, Calin
    PROCEEDINGS OF THE EUROPEAN INTEGRATION: BETWEEN TRADITION AND MODERNITY, VOL 5, 2013, 5 : 1155 - 1164
  • [50] Function Approximation by Neural Networks
    Li, Fengjun
    ADVANCES IN NEURAL NETWORKS - ISNN 2008, PT I, PROCEEDINGS, 2008, 5263 : 384 - 390