November 2019
Intermediate to advanced
346 pages
9h 36m
English
We start by reading a CRP dataset into a dataframe (step 1). In step 2, we create x and y NumPy arrays to hold the features and labels. Next, we train-test split our data (step 3) and then train and test a classifier for CRPs (steps 4 and 5). Based on performance, we can see that ML can accurately predict responses to PUF challenges. The implications are that, while using our trained model, we can build a software clone of the PUF and use it to (falsely) authenticate.