arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2007.14433cs.CVcs.LG
收藏

Cassandra:从对抗性扰动中检测特洛伊木马网络

Cassandra: Detecting Trojaned Networks from Adversarial Perturbations

Center for Research in Computer Vision, University of Central Florida(中佛罗里达大学计算机视觉研究中心);Computer Science and Software Engineering, The University of Western Australia(西澳大学计算机科学与软件工程学院);Department of Electrical and Computer Engineering, University of Central Florida(中佛罗里达大学电气与计算机工程系)

Xiaoyu Zhang, Ajmal Mian, Rohit Gupta, Nazanin Rahnavard, Mubarak Shah

摘要:

深度神经网络由于其高分类精度而被广泛部署用于许多关键任务。在许多情况下,预先训练的模型来自供应商,他们可能破坏了训练管道,将特洛伊木马行为插入模型中。这些恶意行为可以在对手的意愿下触发,因此对深度模型的广泛部署造成严重威胁。我们提出了一种方法来验证预训练的模型是木马还是良性的。我们的方法捕获神经网络的指纹,其形式是从网络梯度中学习到的对抗性扰动。在网络中插入后门会改变其决策边界,这些边界被有效地编码在其对抗性扰动中。我们训练一个双流网络木马检测从其全球($L_\infty$和$L_2$有界)的扰动和高能量在每个扰动的局部区域。前者编码网络的决策边界,后者编码未知的触发形状。我们还提出了一种异常检测方法来识别木马网络中的目标类。我们的方法是不变的触发类型,触发大小,训练数据和网络架构。我们在MNIST,NIST-Round 0和NIST-Round 1数据集上评估了我们的方法,其中多达1,000个预训练模型使其成为迄今为止关于特洛伊木马网络检测的最大研究,并实现了超过92%的检测准确率,以设置新的最先进的技术。

英文摘要:

Deep neural networks are being widely deployed for many critical tasks due to their high classification accuracy. In many cases, pre-trained models are sourced from vendors who may have disrupted the training pipeline to insert Trojan behaviors into the models. These malicious behaviors can be triggered at the adversary's will and hence, cause a serious threat to the widespread deployment of deep models. We propose a method to verify if a pre-trained model is Trojaned or benign. Our method captures fingerprints of neural networks in the form of adversarial perturbations learned from the network gradients. Inserting backdoors into a network alters its decision boundaries which are effectively encoded in their adversarial perturbations. We train a two stream network for Trojan detection from its global ($L_\infty$ and $L_2$ bounded) perturbations and the localized region of high energy within each perturbation. The former encodes decision boundaries of the network and latter encodes the unknown trigger shape. We also propose an anomaly detection method to identify the target class in a Trojaned network. Our methods are invariant to the trigger type, trigger size, training data and network architecture. We evaluate our methods on MNIST, NIST-Round0 and NIST-Round1 datasets, with up to 1,000 pre-trained models making this the largest study to date on Trojaned network detection, and achieve over 92\% detection accuracy to set the new state-of-the-art.

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。

↑