[论文解读] Deep Partition Aggregation: Provable Defenses against General Poisoning Attacks
本文提出了深度划分聚合(DPA)和半监督DPA(SS-DPA),分别针对一般性投毒攻击和标签翻转攻击的新型可证明防御方法。DPA利用基于哈希的数据划分和集成学习提供可证明的鲁棒性,实现最先进的认证性能:在MNIST上可抵御超过500次投毒插入,在CIFAR-10上可抵御9次投毒插入;SS-DPA在MNIST上可容忍超过600次标签翻转,在CIFAR-10上可容忍300次标签翻转。
Adversarial poisoning attacks distort training data in order to corrupt the test-time behavior of a classifier. A provable defense provides a certificate for each test sample, which is a lower bound on the magnitude of any adversarial distortion of the training set that can corrupt the test sample's classification. We propose two novel provable defenses against poisoning attacks: (i) Deep Partition Aggregation (DPA), a certified defense against a general poisoning threat model, defined as the insertion or deletion of a bounded number of samples to the training set --- by implication, this threat model also includes arbitrary distortions to a bounded number of images and/or labels; and (ii) Semi-Supervised DPA (SS-DPA), a certified defense against label-flipping poisoning attacks. DPA is an ensemble method where base models are trained on partitions of the training set determined by a hash function. DPA is related to both subset aggregation, a well-studied ensemble method in classical machine learning, as well as to randomized smoothing, a popular provable defense against evasion (inference) attacks. Our defense against label-flipping poison attacks, SS-DPA, uses a semi-supervised learning algorithm as its base classifier model: each base classifier is trained using the entire unlabeled training set in addition to the labels for a partition. SS-DPA significantly outperforms the existing certified defense for label-flipping attacks (Rosenfeld et al., 2020) on both MNIST and CIFAR-10: provably tolerating, for at least half of test images, over 600 label flips (vs. < 200 label flips) on MNIST and over 300 label flips (vs. 175 label flips) on CIFAR-10. Against general poisoning attacks where no prior certified defenses exists, DPA can certify ≥ 50% of test images against over 500 poison image insertions on MNIST, and nine insertions on CIFAR-10. These results establish new state-of-the-art provable defenses against general and label-flipping poison attacks.
研究动机与目标
- 开发一种可证明的防御机制,以抵御涉及插入、删除或图像/标签失真等有限数量训练样本的通用投毒攻击。
- 设计一种专门针对标签翻转投毒攻击的认证防御方法,其鲁棒性优于先前工作。
- 在通用与标签翻转投毒威胁下,建立认证鲁棒性的新SOTA(最先进)水平。
- 利用集成学习与半监督训练提升认证保证,同时保持高准确率。
- 为每个测试样本提供理论认证,量化使预测结果被破坏所需的最小投毒强度。
提出的方法
- DPA使用随机哈希函数对训练集进行划分,基于不相交的子集训练基础模型,形成集成模型。
- 最终预测通过基础模型的多数投票决定,其鲁棒性通过改变预测所需最少样本数的下界进行认证。
- SS-DPA通过引入半监督学习扩展DPA:每个基础模型在完整未标记数据及自身划分中的标签上进行训练,从而提升对标签翻转攻击的鲁棒性。
- 认证依赖于组合论证,以限制改变集成预测所需的标签或数据扰动数量。
- 该方法继承了子集聚合与随机平滑的特性,结合两者优势以实现对投毒攻击的防御。
- 通过划分结构与投票机制的特性,高效计算鲁棒性认证,支持实际部署。
实验结果
研究问题
- RQ1能否构建一种认证防御机制,可可证明地抵抗涉及有限数量训练样本插入、删除或失真的通用投毒攻击?
- RQ2该防御的半监督变体是否能显著提升对标签翻转攻击的鲁棒性,相比现有认证防御方法?
- RQ3在给定测试样本下,该框架最多可认证多少次投毒扰动(插入或标签翻转)?
- RQ4所提出的防御方法在标准基准数据集(如MNIST与CIFAR-10)上的性能与先前认证防御相比如何?
- RQ5DPA的集成结构是否能在相同威胁模型下,提供比现有方法更强的理论保证?
主要发现
- DPA在MNIST上对至少50%的测试图像实现鲁棒性认证,可抵御超过500次投毒图像插入,创下新SOTA纪录。
- 在CIFAR-10上,DPA对至少一半测试图像提供对九次投毒插入的认证,展现出强大的通用投毒防御能力。
- SS-DPA在标签翻转攻击中优于先前认证防御方法,在MNIST上可容忍超过600次标签翻转(相较之前工作<200次)。
- 在CIFAR-10上,SS-DPA对超过300次标签翻转实现认证(相较之前工作175次),确立了新基准。
- 所提出的防御方法在通用与标签翻转投毒威胁模型下,均显著优于现有方法的认证鲁棒性。
- 划分、集成投票与半监督学习的结合,实现了强大的理论保证与标准数据集上的实用性能。
更好的研究,从现在开始
从阅读论文到最终审阅,大幅缩短您的研究时间。
无需绑定信用卡
本解读由 AI 生成,并经人工编辑审核。