arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2409.19638cs.CVcs.AI
收藏

BadHMP:针对人体运动预测的后门攻击

BadHMP: Backdoor Attack against Human Motion Prediction

School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore(电子工程学院,南洋理工大学,新加坡)

Chaohui Xu, Si Wang, Chip-Hong Chang

AI总结:

提出 BadHMP 后门攻击,通过局部肢体触发器与全局目标轨迹污染人体运动预测模型,并在两个数据集和两种架构上验证其有效性、隐蔽性及抗微调能力。

摘要:

根据过去观测在亚秒级时间范围内精确预测未来人体运动,对于多种安全关键应用至关重要。迄今为止,仅有少量研究考察了基于骨骼的神经网络在规避攻击和后门攻击下的脆弱性。本文提出 BadHMP,这是一种专门针对人体运动预测任务的新型后门攻击。该方法通过在骨骼的一个肢体中嵌入局部后门触发器来生成中毒训练样本,使所选关节在历史时间步中遵循预定义运动。随后,对未来序列进行全局修改,使所有关节都按照目标轨迹运动。精心设计的后门触发器和目标保证了中毒样本的平滑性与自然性,使其足够隐蔽,能够逃避模型训练者的检测,同时使中毒模型在未污染序列上的预测保真度不引人注目。即使在较低的中毒样本注入比例下,目标序列也能被所设计的输入序列成功激活。在 Human3.6M 和 CMU-Mocap 两个数据集以及 LTD 和 HRI 两种网络架构上的实验结果证明了 BadHMP 的高保真度、有效性和隐蔽性。该攻击针对微调防御的鲁棒性也得到了验证。

英文摘要:

Precise future human motion prediction over sub-second horizons from past observations is crucial for various safety-critical applications. To date, only a few studies have examined the vulnerability of skeleton-based neural networks to evasion and backdoor attacks. In this paper, we propose BadHMP, a novel backdoor attack that targets specifically human motion prediction tasks. Our approach involves generating poisoned training samples by embedding a localized backdoor trigger in one limb of the skeleton, causing selected joints to follow predefined motion in historical time steps. Subsequently, the future sequences are globally modified that all the joints move following the target trajectories. Our carefully designed backdoor triggers and targets guarantee the smoothness and naturalness of the poisoned samples, making them stealthy enough to evade detection by the model trainer while keeping the poisoned model unobtrusive in terms of prediction fidelity to untainted sequences. The target sequences can be successfully activated by the designed input sequences even with a low poisoned sample injection ratio. Experimental results on two datasets (Human3.6M and CMU-Mocap) and two network architectures (LTD and HRI) demonstrate the high-fidelity, effectiveness, and stealthiness of BadHMP. Robustness of our attack against fine-tuning defense is also verified.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑