A Smart Privacy-Preserving Learning Method by Fake Gradients to Protect Users Items in Recommender Systems

Table 1

Notations.

Notation	Meaning
	Number of users
	Number of items
	The user profile vector
	The item profile vector
	The loss function
	The set of the ratings
	Bound of the norm of the gradient in privacy gradient descent
	The learning rate
	The regularization parameters
	-Smooth parpameters for the loss function