GRBM learning rate of bottom layer in Fusion 1 is set as [epsilon] = 0.001 and training cycle is set as J = 200.
The same as Fusion 1, GRBM learning rate of bottom layer in Fusion 2 is set as [epsilon] = 0.001 and training cycle is set as J = 200; the upper layer RBM is set as [epsilon] = 0.01 and J = 70.
Therefore, following the work [20, 24], Gaussian Restricted Boltzmann Machines (
GRBMs) are adopted to initialize weight parameters by performing pretraining, which has been already applied widely.