Bias Busters：基于DL的光刻热点检测器对抗后门攻击
Bias Busters: Robustifying DL-based Lithographic Hotspot Detectors Against Backdooring Attacks
摘要：
深度学习（DL）在整个CAD工具流程中提供了潜在的改进，其中一个有前途的应用是光刻热点检测。然而，DL技术已经被证明特别容易受到推理和训练时间的对抗攻击的影响。最近的工作表明，一小部分恶意的物理设计者可以在其训练阶段悄悄地“后门”基于DL的热点检测器，使其准确地分类常规布局剪辑，但预测包含特制触发形状的热点为非热点。我们提出了一种新的训练数据增强策略，作为对这种后门攻击的强大防御。防御的工作原理是消除训练数据中引入的故意偏见，但不需要知道哪些训练样本中毒或后门触发器的性质。我们的研究结果表明，防御可以大大降低攻击成功率从84%到~ 0%。
英文摘要：
Deep learning (DL) offers potential improvements throughout the CAD tool-flow, one promising application being lithographic hotspot detection. However, DL techniques have been shown to be especially vulnerable to inference and training time adversarial attacks. Recent work has demonstrated that a small fraction of malicious physical designers can stealthily "backdoor" a DL-based hotspot detector during its training phase such that it accurately classifies regular layout clips but predicts hotspots containing a specially crafted trigger shape as non-hotspots. We propose a novel training data augmentation strategy as a powerful defense against such backdooring attacks. The defense works by eliminating the intentional biases introduced in the training data but does not require knowledge of which training samples are poisoned or the nature of the backdoor trigger. Our results show that the defense can drastically reduce the attack success rate from 84% to ~0%.
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。
