L10.5.1 The Main Concept Behind Dropout

[KountayDwivedi]

​ @Sebastian Raschka Dear Sir, thank you for the implementation-oriented explanation of Dropout. Could you please explain, for the example shown, how is it confirmed that the random sampling from a uniform distribution would give exactly 50% of numbers greater than 0.5 and 50% of them less than 0.5? Is it not possible that each generated number (or more than 50%, in this case) comes out to be greater than 0.5?

Thank you Sir.
:-}

[virgileperin]

09:13 : is it the activation or the net input that needs to be multiplied by (1-p)?