Dong Chen

LLMs can now directly generate relevant Point-of-Interest (POI) candidates for map search by encoding both semantic and geographic context, outperforming traditional retrieval methods.

Topological priors aren't just for features anymore: injecting persistent homology at multiple stages of point cloud networks—from sampling to regularization—boosts accuracy and robustness.

Spectral GNNs' purported spectral advantages for node classification are illusory; their performance actually hinges on their underlying MPNN structure, debunking the "graph Fourier transform" narrative.