Long-tail data distributions are prevalent in many real-world networks,
including financial transaction networks, e-commerce networks, and
collaboration networks. Despite the success of recent developments, the
existing works mainly focus on debiasing the machine learning models via graph
augmentation or objective reweighting. However, there is limited liter