Fig. 2From: Distributed Gram-Schmidt orthogonalization with simultaneous elements refinementConvergence for networks with different topology and initial data distribution: either all nodes store the same amount of data (“uniform”) or most of the data is stored in one node (with minimum or maximum degree) (a - Regular topology with \(\overline {d}=5\); b - Geometric topology with \(\overline {d}=5\)). In case of the regular topology (a), the nodes i,j are picked randomlyBack to article page