Balancing detectability and performance of attacks on the control channel of Markov Decision Processes

被引:0
|
作者
Russo, Alessio [1 ]
Proutiere, Alexandre [1 ]
机构
[1] KTH Royal Inst Technol, EECS Sch, Div Decis & Control Syst, Stockholm, Sweden
关键词
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We investigate the problem of designing optimal stealthy poisoning attacks on the control channel of Markov decision processes (MDPs). This research is motivated by the recent interest of the research community for adversarial and poisoning attacks applied to MDPs, and reinforcement learning (RL) methods. The policies resulting from these methods have been shown to be vulnerable to attacks perturbing the observations of the decision-maker. In such an attack, drawing inspiration from adversarial examples used in supervised learning, the amplitude of the adversarial perturbation is limited according to some norm, with the hope that this constraint will make the attack imperceptible. However, such constraints do not grant any level of undetectability and do not take into account the dynamic nature of the underlying Markov process. In this paper, we propose a new attack formulation, based on information-theoretical quantities, that considers the objective of minimizing the detectability of the attack as well as the performance of the controlled process. We analyze the trade-off between the efficiency of the attack and its detectability. We conclude with examples and numerical simulations illustrating this trade-off.
引用
收藏
页码:2843 / 2850
页数:8
相关论文
共 50 条
  • [1] Trading Performance for Stability in Markov Decision Processes
    Brazdil, Tomas
    Chatterjee, Krishnendu
    Forejt, Vojtech
    Kucera, Antonin
    2013 28TH ANNUAL IEEE/ACM SYMPOSIUM ON LOGIC IN COMPUTER SCIENCE (LICS), 2013, : 331 - 340
  • [2] Trading performance for stability in Markov decision processes
    Brazdil, Tomas
    Chatterjee, Krishnendu
    Forejt, Vojtech
    Kucera, Antonin
    JOURNAL OF COMPUTER AND SYSTEM SCIENCES, 2017, 84 : 144 - 170
  • [3] The complexity of decentralized control of Markov decision processes
    Bernstein, DS
    Givan, R
    Immerman, N
    Zilberstein, S
    MATHEMATICS OF OPERATIONS RESEARCH, 2002, 27 (04) : 819 - 840
  • [4] Performance Guarantees for Homomorphisms beyond Markov Decision Processes
    Majeed, Sultan Javed
    Hutter, Marcus
    THIRTY-THIRD AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE / THIRTY-FIRST INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE / NINTH AAAI SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2019, : 7659 - 7666
  • [5] Likelihood Analysis of Cyber Data Attacks to Power Systems With Markov Decision Processes
    Hao, Yingshuai
    Wang, Meng
    Chow, Joe H.
    IEEE TRANSACTIONS ON SMART GRID, 2018, 9 (04) : 3191 - 3202
  • [6] Monotone optimal control for a class of Markov decision processes
    Zhuang, Weifen
    Li, Michael Z. F.
    EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2012, 217 (02) : 342 - 350
  • [7] Server Frequency Control Using Markov Decision Processes
    Chen, Lydia Y.
    Gautam, Natarajan
    IEEE INFOCOM 2009 - IEEE CONFERENCE ON COMPUTER COMMUNICATIONS, VOLS 1-5, 2009, : 2951 - +
  • [8] Optimal control in light traffic Markov Decision Processes
    INRIA, Sophia Antipolis, France
    ZOR, 1 (63-79):
  • [9] Optimal control in light traffic Markov decision processes
    Ger Koole
    Olaf Passchier
    Mathematical Methods of Operations Research, 1997, 45 : 63 - 79
  • [10] Decentralized Control of Partially Observable Markov Decision Processes
    Amato, Christopher
    Chowdhary, Girish
    Geramifard, Alborz
    Uere, N. Kemal
    Kochenderfer, Mykel J.
    2013 IEEE 52ND ANNUAL CONFERENCE ON DECISION AND CONTROL (CDC), 2013, : 2398 - 2405