Algorithm 1 RR-Net

Input: samples $[{\mathbf{x}}_1,{\mathbf{x}}_2,\dots ,{\mathbf{x}}_n]$, C. Output: the trained network and the reconstruction losses ${ϵ}_1,{ϵ}_2,\dots {ϵ}_n$. 1:Initialize $\mathbf{W}$ and $\mathbf{b}$. 2:Update $\mathbf{\mu }$ by calculating Equation (3). 3:Calculate reconstruction errors with ${ϵ}_i={∥r{\left({\mathbf{x}}_{\mathbf{i}}\right)}^T·\mathbf{\mu }-{\mathbf{x}}_{\mathbf{i}}∥}_2^2$.