A 0.26% BER, 1028 Challenge-Response Machine-Learning Resistant Strong-PUF in 14nm CMOS Featuring Stability-Aware Adversarial Challenge Selection
A 0.26% BER, 1028 Challenge-Response Machine-Learning Resistant Strong-PUF in 14nm CMOS Featuring Stability-Aware Adversarial Challenge Selection
复制标题
A%200.26%%20BER,%201028%20挑战-响应%20机器学习%20抵抗%20强PUF%20in%2014nm%20CMOS%20特色%20稳定性意识%20对抗性%20挑战%20选择
DOI：
--
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
S. Mathew
中科院分区：
文献类型：
--
作者：
Vikram B. Suresh;Raghavan Kumar;M. Anders;Himanshu Kaul;V. De;S. Mathew
一种采用14纳米CMOS工艺的1028位挑战 - 响应强物理不可克隆函数（PUF），在600万个训练样本中展现出对机器学习（ML）攻击的抵抗力。具有对抗性挑战选择的两级非线性级联PUF阵列将机器学习攻击的准确率限制在约50%。基于可配置交叉耦合反相器的熵源，通过稳定性感知的挑战修剪，在650 - 850毫伏和0 - 100°C的范围内实现了9.8倍的更高阵列密度以及0.26%的峰值误码率。
A 1028 challenge-response strong-PUF in 14nm CMOS, demonstrates machine learning (ML) attack resistance across 6-million training samples. The 2-stage non-linear cascaded PUF array with adversarial challenge selection limits ML attack accuracy to ∼50%. The configurable cross-coupled inverter-based entropy source with stability-aware challenge pruning enables 9.8× higher array density and 0.26% peak BER across 650–850mV and 0–100°C.