Below are the repository of codes that I participated. You may find the accompany papers in the corresponding repository. Please also check my homepage for a brief introduction.
- [KDD 2022] HyperAid: Leveraging hyperbolic geometry to improve the performance of tree metric and ultrametric learning methods.
- [ICLR 2022] GIANT-XRT: Method to extract high quality node features from raw data with self-supervised learning. Achieving rank-1 on ogbn-arxiv, ogbn-products and ogbn-papers100M datasets with huge improvement.
- [ICLR 2022] AllSet: A multiset function framework for hypergraph neural networks. We also collected and curated many benchmark datasets for hypergraph nerual networks.
- [ICDM 2021 regular] Poincare linear classification: Linear classification algorithms in Poincare ball (hyperbolic space) with theoretical guarantees. Our code includes perceptron and SVM.
- [ICLR 2021] GPRGNN: A GNN provably not only works for heterophilic datasets, but also prevents over-smoothing problem. Furthermore, GPRGNN has the benefit on interpretability.