一句话：提出弹性异构 DNN 加速器架构 DNNGuard，为对抗样本防御提供低开销、高能效且安全的硬件支持。
深度神经网络易受对抗样本攻击：对正确分类输入施加扰动即可导致误分类，在无人驾驶、金融和医疗等安全敏感场景中后果严重。已有对抗样本防御方法依赖多种计算单元，将其部署到现有 DNN 加速器会带来成本、计算效率和信息安全等问题；同时，现有加速器也难以支持防御方法所需的特殊计算。为此，论文提出 DNNGuard——一种弹性异构 DNN 加速器架构，旨在以弹性、异构的硬件组织为对抗样本防御提供有效支持，并缓解上述成本、效率与安全问题。摘要未报告具体实现细节与量化实验结果。
Recent studies show that Deep Neural Networks (DNN) are vulnerable to adversarial samples that are generated by perturbing correctly classified inputs to cause the misclassification of DNN models. This can potentially lead to disastrous consequences, especially in security-sensitive applications such as unmanned vehicles, finance and healthcare. Existing adversarial defense methods require a variety of computing units to effectively detect the adversarial samples. However, deploying adversary sample defense methods in existing DNN accelerators leads to many key issues in terms of cost, computational efficiency and information security. Moreover, existing DNN accelerators cannot provide effective support for special computation required in the defense methods.
@inproceedings{paperbot5474,
title = {DNNGuard: An Elastic Heterogeneous DNN Accelerator Architecture against Adversarial Attacks},
author = {Xingbin Wang and Rui Hou and Boyan Zhao and Fengkai Yuan and Jun Zhang and Dan Meng and Xuehai Qian},
booktitle = {ASPLOS},
year = {2020},
doi = {10.1145/3373376.3378532}
}