High-Efficiency Min-Entropy Estimation Based on Neural Network for Random Number Generators

Table 6

Entropy estimates for real-world sources.

Real-world sources	90B’s predictors				Our work
	MCW	Lag	MultiMMC	LZ78Y	FNN	RNN
RANDOM.ORG	0.9951	0.9963	0.9966	0.9976	0.9802	0.9954
Ubld.it TrueRNGpro	0.9979	0.9955	0.9973	0.9966	0.9934	0.9728
Linux kernel entropy source	0.6173	0.1232	0.1269	0.6164	0.1230	0.3068
Linux/dev/urandom	0.9952	0.9935	0.9990	0.9964	0.9983	0.9911
Windows RNG	0.9953	0.9986	0.9975	0.9984	0.9833	0.9853