Evolution processes of complex networked systems in biology and social sciences, and their underlying mechanisms, still need better understanding. The authors propose a machine learning approach to reconstruct the evolution history of complex networks.