Graph neural networks (GNNs) have shown remarkable success in molecular property prediction as key models in geometric deep learning. Meanwhile, Kolmogorov–Arnold networks (KANs) have emerged as ...