Residual Flows

• inspiration from Res Net
• use Banach fixed point theorem

Inverting Residual Network Layers

• form $h^{\prime }=h+f[h,\varphi ]$ , if we ensure that $f[h,\varphi ]$ is a contraction mapping (Banach fixed point theorem)
• lipschitz constant must be less than 1
• Assuming that the slope of the activation functions is not greater than one, this is equivalent to ensuring the largest eigenvalue of each weight matrix $\Omega$ must be less than 1
•