Announcement_2025 05 27_2

Excited to share we have two works accepted at ICML 2025 on rethinking historical embeddings in GNN scaling and hyperbolic geometry for backwards-compatible learning.