When GNNs meet symmetry in ILPs: an orbit-based feature augmentation approach

Open in new window