“…Compared with heuristic ones, model-based methods for network clustering may perform the task by optimizing the formulated objective functions, which quantify the holistic quality of vertex segmentation in terms of edge density, and other additional information, e.g., vertex features. Many machine learning techniques, e.g., spectral analysis [35], [45], matrix factorization [15], [43], [44], Bayesian modeling [7], [30], and deep learning [6], [18], [25], [48], have been successfully adapted to perform the task of uncovering network clusters, taking into consideration the edge structure, vertex features, or both.…”