.png){kind=icon}
返回
State-Aware Perturbation Optimization for Robust Deep Reinforcement Learning
.png){kind=icon}
PRE
AI
DOI：10.1109/TMC.2025.3601531.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Recently, deep reinforcement learning (DRL) has emerged as a promising approach for robotic control. However, the deployment of DRL in real-world robots is hindered by its sensitivity to environmental perturbations. While existing white-box adversarial attacks rely on local gradient information and apply uniform perturbations across all states to evaluate DRL robustness, they fail to account for temporal dynamics and state-specific vulnerabilities. To combat the above challenge, we first conduct a theoretical analysis of white-box attacks in DRL by establishing the Adversarial Victim Dynamics Markov Decision Process (AVD-MDP), to derive the necessary and sufficient conditions for a successful attack. Based on this, we propose the Selective State-Aware Reinforcement adversarial attack (STAR), to optimize perturbation stealthiness and state visitation dispersion. STAR first employs a soft mask-based state-targeting mechanism to minimize redundant perturbations, enhancing stealthiness and attack effectiveness. Then, it incorporates an information-theoretic optimization objective to maximize mutual information between perturbations, environmental states, and victim actions, ensuring a dispersed state-visitation distribution that steers the victim agent into vulnerable states for maximum return reduction. Extensive experiments demonstrate that STAR outperforms state-of-the-art benchmarks.
Keyword:
Perturbation methods
Robots
Glass box
Stars
Robustness
Deep reinforcement learning
Navigation
Trajectory
Training
Robot sensing systems
Markov decision process
deep reinforcement learning
adversarial attack
robotic manipulation
期刊
IF:
9.2
论文数:
5.9K
被引数:
1.8W
机构
引用论文
Rethinking Adversarial Attacks in Reinforcement Learning from Policy Distribution Perspective.png)
ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)IF0
PREAI
分享收藏Advances in Adversarial Attacks and Defenses in Computer Vision: A Survey计算机视觉中对抗性攻击和防御的进展: 综述IEEE ACCESS
OAAI
IEEE ACCESSIF3.6
OAAI分享收藏Challenges and Countermeasures for Adversarial Attacks on Deep Reinforcement Learning深度强化学习对抗性攻击的挑战与对策OAAI
OAAI分享收藏Exploring misclassifications of robust neural networks to enhance adversarial attacks探索鲁棒神经网络的错误分类以增强对抗攻击APPLIED INTELLIGENCEOAAI
APPLIED INTELLIGENCEIF3.5
OAAI分享收藏Evading Defenses to Transferable Adversarial Examples by Translation-Invariant Attacks通过翻译不变的攻击规避可转移的对抗示例的防御
OAAI
OAAI分享收藏
查看更多内容