Deep FeedForward Neural Network based Model Building Attack Tool on PUF
What things you need to install the software and how to install them
python >= 3.6
tensorflow >= 2.0
kerastuner
Other important packages required in python:
numpy
pandas
sklearn
For running the tests, go to the respective PUF folder and execute
python dfnn_attack.py --challenge=<path>XOR_APUF_Binary_chal_64_500000.csv --response=<path>/4-xorpuf.csv \
--features=parity --level=1 --verbose=0
For more options:
python dfnn_attack.py -h
usage: dfnn_attack.py [-h] [-c path] [-r path] [-f FEATURES] [-l LEVEL]
[-v VERBOSE]
Deep learning based model building attack tool on PUF
optional arguments:
-h, --help show this help message and exit
-c path, --challenge path
the path to challenges CSV file
-r path, --response path
the path to response CSV file
-f FEATURES, --features FEATURES
To convert the challenges to feature vectors
-l LEVEL, --level LEVEL
level specifies the first-level test (0.5 million
CRPs) or second level test ( <= 1.5 million)
-v VERBOSE, --verbose VERBOSE
verbose=10 enables the epoch prints and verbose=0
disables it