Do We Really Need Graph Convolution During Training? Light Post-Training Graph-ODE for Efficient Recommendation

GNN
Re-assign community