No Change, No Gain: Empowering Graph Neural Networks with Expected Model Change Maximization for Active Learning | Read Paper on Bytez