geometric-intelligence
Category: A2; Team name: DLLB; Dataset: HIC
Co-authored-by: luka-benic [email protected] Co-authored-by: dleko11 [email protected]
Checklist
- [x] My pull request has a clear and explanatory title.
- [x] My pull request passes the Linting test.
- [x] I added appropriate unit tests and I made sure the code passes all unit tests. (refer to comment below)
- [x] My PR follows PEP8 guidelines. (refer to comment below)
- [x] My code is properly documented, using numpy docs conventions, and I made sure the documentation renders properly.
- [x] I linked to issues and PRs that are relevant to this PR.
Description
This pull request adds support for the HIC (Hypergraph Isomorphism Computation) dataset collection introduced in papers Hypergraph Isomorphism Computation and Expressive Hypergraph Neural Networks [1] [2]. This contribution is part of the TAG-DS Topological Deep Learning Challenge 2025, under Mission A, Category A.2: Curating Natively Higher-Order Datasets [3].
The HIC benchmark consists of synthetic and real-world hypergraph datasets designed to evaluate the expressive power of hypergraph neural networks. The dataset collection includes:
- RHG family (RHG_3, RHG_10, RHG_table, RHG_pyramid): synthetic hypergraphs with controllable structural properties.
- IMDB-based hypergraphs (IMDB_dir_form, IMDB_dir_genre, IMDB_wri_form, IMDB_wri_genre, IMDB_dir_genre_m, IMDB_wri_genre_m): constructed from the original IMDB data.
- Real-world networks (steam_player, twitter_friend): interaction-based hypergraphs.
Each hypergraph dataset defines a graph classification task.
References:
[1] Feng, Y., Han, J., Ying, S., & Gao, Y. (2024). Hypergraph Isomorphism Computation. IEEE Transactions on Pattern Analysis and Machine Intelligence.
[2] Deng, Y., Li, H., Du, L., & Zhang, P. (2023). Expressive Hypergraph Neural Networks. NeurIPS 2023. [arXiv]
[3] TAG-DS Topological Deep Learning Challenge 2025
[4] Original HIC Repository
