Recently, the selection outcomes for the 42nd International Conference on Machine Learning (ICML 2024) were unveiled, revealing a competitive acceptance rate of 27.5%. Notably, a paper authored by Professor Liu Jing's research team at the School of Artificial Intelligence, Xidian University, has been successfully chosen for presentation. This groundbreaking paper introduces an innovative and efficient automated framework for searching loss functions, specifically designed to tackle the challenge of class-imbalanced node classification. The loss functions unearthed by this framework exhibit a marked improvement in node classification accuracy. Furthermore, these loss functions, initially identified under conditions of high class imbalance, demonstrate remarkable generalizability and adaptability, effectively extending their utility to scenarios characterized by low class imbalance.