On the Initialisation of Wide Low-Rank Feedforward Neural Networks

01/31/2023
by   Thiziri Nait Saada, et al.
0

The edge-of-chaos dynamics of wide randomly initialized low-rank feedforward networks are analyzed. Formulae for the optimal weight and bias variances are extended from the full-rank to low-rank setting and are shown to follow from multiplicative scaling. The principle second order effect, the variance of the input-output Jacobian, is derived and shown to increase as the rank to width ratio decreases. These results inform practitioners how to randomly initialize feedforward networks with a reduced number of learnable parameters while in the same ambient dimension, allowing reductions in the computational cost and memory constraints of the associated network.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset