Reinforcement Learning Based Sparse Black-box Adversarial Attack on Video Recognition Models

被引:0
|
作者
Wang, Zeyuan [1 ]
Sha, Chaofeng [1 ]
Yang, Su [1 ]
机构
[1] Fudan Univ, Shanghai Key Lab Intelligent Informat Proc, Sch Comp Sci, Shanghai, Peoples R China
关键词
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
We explore the black-box adversarial attack on video recognition models. Attacks are only performed on selected key regions and key frames to reduce the high computation cost of searching adversarial perturbations on a video due to its high dimensionality. To select key frames, one way is to use heuristic algorithms to evaluate the importance of each frame and choose the essential ones. However, it is time inefficient on sorting and searching. In order to speed up the attack process, we propose a reinforcement learning based frame selection strategy. Specifically, the agent explores the difference between the original class and the target class of videos to make selection decisions. It receives rewards from threat models which indicate the quality of the decisions. Besides, we also use saliency detection to select key regions and only estimate the sign of gradient instead of the gradient itself in zeroth order optimization to further boost the attack process. We can use the trained model directly in the untargeted attack or with little fine-tune in the targeted attack, which saves computation time. A range of empirical results on real datasets demonstrate the effectiveness and efficiency of the proposed method.
引用
收藏
页码:3162 / 3168
页数:7
相关论文
共 50 条
  • [31] A low-query black-box adversarial attack based on transferability
    Ding, Kangyi
    Liu, Xiaolei
    Niu, Weina
    Hu, Teng
    Wang, Yanping
    Zhang, Xiaosong
    [J]. KNOWLEDGE-BASED SYSTEMS, 2021, 226
  • [32] An adversarial attack on DNN-based black-box object detectors
    Wang, Yajie
    Tan, Yu-an
    Zhang, Wenjiao
    Zhao, Yuhang
    Kuang, Xiaohui
    [J]. JOURNAL OF NETWORK AND COMPUTER APPLICATIONS, 2020, 161
  • [33] Boosting Black-box Adversarial Attack with a Better Convergence
    Yin, Heng
    Wang, Jindong
    Mi, Yan
    Zhang, Xiaoning
    [J]. 2020 5TH INTERNATIONAL CONFERENCE ON MECHANICAL, CONTROL AND COMPUTER ENGINEERING (ICMCCE 2020), 2020, : 1234 - 1238
  • [34] A black-box adversarial attack on demand side management
    Cramer, Eike
    Gao, Ji
    [J]. COMPUTERS & CHEMICAL ENGINEERING, 2024, 186
  • [35] Black-box Bayesian adversarial attack with transferable priors
    Zhang, Shudong
    Gao, Haichang
    Shu, Chao
    Cao, Xiwen
    Zhou, Yunyi
    He, Jianping
    [J]. MACHINE LEARNING, 2024, 113 (04) : 1511 - 1528
  • [36] Adaptive hyperparameter optimization for black-box adversarial attack
    Zhenyu Guan
    Lixin Zhang
    Bohan Huang
    Bihe Zhao
    Song Bian
    [J]. International Journal of Information Security, 2023, 22 : 1765 - 1779
  • [37] An Effective Way to Boost Black-Box Adversarial Attack
    Feng, Xinjie
    Yao, Hongxun
    Che, Wenbin
    Zhang, Shengping
    [J]. MULTIMEDIA MODELING (MMM 2020), PT I, 2020, 11961 : 393 - 404
  • [38] SCHMIDT: IMAGE AUGMENTATION FOR BLACK-BOX ADVERSARIAL ATTACK
    Shi, Yucheng
    Han, Yahong
    [J]. 2018 IEEE INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO (ICME), 2018,
  • [39] Black-Box Decision based Adversarial Attack with Symmetric α-stable Distribution
    Srinivasan, Vignesh
    Kuruoglu, Ercan E.
    Mueller, Klaus-Robert
    Samek, Wojciech
    Nakajima, Shinichi
    [J]. 2019 27TH EUROPEAN SIGNAL PROCESSING CONFERENCE (EUSIPCO), 2019,
  • [40] Black-Box Adversarial Attack via Overlapped Shapes
    Williams, Phoenix
    Li, Ke
    Min, Geyong
    [J]. PROCEEDINGS OF THE 2022 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE COMPANION, GECCO 2022, 2022, : 467 - 468