Abstract: We introduce a distributed optimization framework for directed graph networks that addresses composite objective functions with smooth local components and a shared convex regulariser. Our ...