Search papers, labs, and topics across Lattice.
Peter L. Reichertz Institute for Medical Informatics, Hannover Medical School, Hannover, Germany, Lower Saxony Center for Artificial Intelligence and Causal Methods in Medicine (CAIMed), Hannover, Germany
2
0
4
Ditch the global alignment: ST-Sheaf GNN learns dynamic, localized information flow for spatio-temporal forecasting, outperforming state-of-the-art methods by explicitly modeling latent local structure.
By explicitly disentangling redundant, unique, and synergistic information, COrAL achieves more stable and comprehensive multimodal embeddings compared to existing contrastive learning methods.