arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2203.04713cs.CV
收藏

基于黑箱推理的人类活动分类器的防御

Defending Black-box Skeleton-based Human Activity Classifiers

University of Leeds(利兹大学);Hefei University of Technology(合肥工业大学);Institute of Deep Learning, Baidu Research(百度研究院深度学习研究所)

He Wang, Yunfeng Diao, Zichang Tan, Guodong Guo

摘要:

人体活动识别(HAR)中的人体动作一直是研究的重点。最近,一个普遍的漏洞,基于HAR已被确定在各种分类器和数据,呼吁缓解。为此,我们提出了第一个黑盒防御方法,为我们最好的知识,基于HAR。我们的方法的特点是对干净的数据,对手和分类器进行完全的贝叶斯处理,导致(1)一种新的基于贝叶斯能量的鲁棒判别分类器的制定,(2)一种新的基于自然运动流形的对手采样方案,以及(3)一种新的用于黑箱防御的训练后贝叶斯策略。我们将我们的框架命名为贝叶斯能量对抗训练或BEAT。BEAT是简单而优雅的,它将脆弱的黑盒分类器转变为强大的分类器,而不会牺牲准确性。在各种攻击下,它在广泛的骨架HAR分类器和数据集上表现出令人惊讶的通用有效性。代码可在https://github.com/realcrane/RobustActionRecogniser上获得。

英文摘要:

Skeletal motions have been heavily replied upon for human activity recognition (HAR). Recently, a universal vulnerability of skeleton-based HAR has been identified across a variety of classifiers and data, calling for mitigation. To this end, we propose the first black-box defense method for skeleton-based HAR to our best knowledge. Our method is featured by full Bayesian treatments of the clean data, the adversaries and the classifier, leading to (1) a new Bayesian Energy-based formulation of robust discriminative classifiers, (2) a new adversary sampling scheme based on natural motion manifolds, and (3) a new post-train Bayesian strategy for black-box defense. We name our framework Bayesian Energy-based Adversarial Training or BEAT. BEAT is straightforward but elegant, which turns vulnerable black-box classifiers into robust ones without sacrificing accuracy. It demonstrates surprising and universal effectiveness across a wide range of skeletal HAR classifiers and datasets, under various attacks. Code is available at https://github.com/realcrane/RobustActionRecogniser.

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。

↑