Research Article

Multiple Differential Distinguisher of SIMECK32/64 Based on Deep Learning

Table 3

SIMECK32/64 distinguisher of 6–10 rounds MLP compared to CNN.

r (round)Number of parametersTime of trainingAccuracy
MLPCNNMLP (s)CNN (s)MLPCNN

61558544001336257230.99990.9999
71558544001348157270.94320.9998
81558544001340157200.88010.9648
91558544001340857430.61280.7248
101558544001336157610.50160.5501