“…31 The method proposed by Majzoobi, Koushanfar, and Potkonjak safeguards the PUFs by using an input network that avoids individual controlling of the challenges, and an XOR output network that compresses the responses so that they cannot be reverseengineered. 32 Even though attempts have been made to break this type of safeguard, no polynomial method for learning the PUF with an XOR output network has been reported thus far. To ensure attack resiliency, Majzoobi, Koushanfar, and Potkonjak proposed five tests: predictability, collision, sensitivity, reverse engineering, and emulation.…”