基于Fisher信息引导的后门攻击净化方法
Fisher Information guided Purification against Backdoor Attacks
AI总结:
本文提出Fisher信息引导净化(FIP)框架,通过正则化器将后门模型重新优化到更平滑的极小值,在抑制后门效应的同时保留干净数据知识,并在多种任务、数据集和攻击上取得最先进性能。
摘要:
近年来关于后门攻击的研究表明,攻击者可以通过操纵少量训练样本来破坏深度神经网络(DNN)的完整性。我们的分析表明,这种操纵会使后门模型收敛到较差的局部极小值,即与良性模型相比更尖锐的极小值。直观上,可以通过将模型重新优化到更平滑的极小值来净化后门。然而,简单地采用任何以更平滑极小值为目标的优化方法都可能导致次优的净化技术,从而损害干净测试精度。因此,为了有效地实现这种重新优化,受我们建立后门移除与损失平滑性之间联系的新视角启发,我们提出了Fisher信息引导净化(FIP),一种新颖的后门净化框架。所提出的FIP包含几个新颖的正则化器,通过利用Fisher信息矩阵(FIM)的知识,帮助模型在后门移除过程中抑制后门效应并保留已获得的干净数据分布知识。此外,我们引入了FIP的一种高效变体,称为Fast FIP,它显著减少了可调参数的数量,并获得了接近5倍的显著运行时间提升。大量实验表明,所提出的方法在广泛的后门防御基准上达到了最先进(SOTA)性能:5个不同任务——图像识别、目标检测、视频动作识别、3D点云、语言生成;11个不同数据集,包括ImageNet、PASCAL VOC、UCF101;涵盖CNN和视觉Transformer的多种模型架构;14种不同的后门攻击,例如Dynamic、WaNet、LIRA、ISSBA等。
英文摘要:
Studies on backdoor attacks in recent years suggest that an adversary can compromise the integrity of a deep neural network (DNN) by manipulating a small set of training samples. Our analysis shows that such manipulation can make the backdoor model converge to a bad local minima, i.e., sharper minima as compared to a benign model. Intuitively, the backdoor can be purified by re-optimizing the model to smoother minima. However, a naïve adoption of any optimization targeting smoother minima can lead to sub-optimal purification techniques hampering the clean test accuracy. Hence, to effectively obtain such re-optimization, inspired by our novel perspective establishing the connection between backdoor removal and loss smoothness, we propose Fisher Information guided Purification (FIP), a novel backdoor purification framework. Proposed FIP consists of a couple of novel regularizers that aid the model in suppressing the backdoor effects and retaining the acquired knowledge of clean data distribution throughout the backdoor removal procedure through exploiting the knowledge of Fisher Information Matrix (FIM). In addition, we introduce an efficient variant of FIP, dubbed as Fast FIP, which reduces the number of tunable parameters significantly and obtains an impressive runtime gain of almost $5\times$. Extensive experiments show that the proposed method achieves state-of-the-art (SOTA) performance on a wide range of backdoor defense benchmarks: 5 different tasks -- Image Recognition, Object Detection, Video Action Recognition, 3D point Cloud, Language Generation; 11 different datasets including ImageNet, PASCAL VOC, UCF101; diverse model architectures spanning both CNN and vision transformer; 14 different backdoor attacks, e.g., Dynamic, WaNet, LIRA, ISSBA, etc.
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。