By contrast, our work is geared to investigate the scalable sensor selection problem for early detection of dynamic harmful cascades in networks from a model-driven prospective.