Critical Node-aware Augmentation for Hypergraph Contrastive Learning

Authors: Zhuo Li, Yuena Lin, Yipeng Wang, Wenmao Liu, Mingliang Yu, Zhen Yang, Gengyu Lyu

IJCAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on various tasks demonstrate that our method is significantly superior to state-of-the-art methods. ... Extensive experiments on various datasets validate that our model guides better hypergraph representations for downstream tasks and achieves significant superiority against state-of-the-art methods.
Researcher Affiliation Collaboration 1Beijing University of Technology 2NSFOCUS Technologies Group Company Ltd. 3Idealism Beijing Technology Co., Ltd. 4Travelsky
Pseudocode Yes Algorithm 1 The update process of M(i) key Input: E: the set of hyperedges; U(i), U(i+1): the set of low-score hyperedges; M(i 1), M(i): candidate critical nodes; M(i) key: identified critical nodes. Output: M(i+1) key 1: for ek E do 2: if ek U(i) and ek / U(i+1) then 3: Update M(i) key:k by solving (12); 4: else if ek / U(i+1) and M(i 1) k = then 5: Update M(i) key:k by solving (13); 6: end if 7: end for 8: return M(i+1) key
Open Source Code No The paper does not contain an explicit statement about the release of source code or a link to a code repository.
Open Datasets Yes We employ six popular datasets from three categories to evaluate the performance of our proposed CNHCL method. These categories include: co-citation datasets Cora C (C-C), Citeseer (CS), and Pubmed (PB); co-authorship datasets Cora-A (C-A) and DBLP (DB); computer vision and graphics dataset Model Net40 (MN).
Dataset Splits Yes For the task of node classification, we first randomly split up each data set into 10% for training, 10% for validation, and 80% for testing. ... For the task of hyperedge prediction, we first randomly split up each data set into 80% for training, 10% for validation, and 10% for testing.
Hardware Specification Yes All experiments are conducted on 2 NVIDIA Ge Force RTX 3090 GPUs with 24GB of memory.
Software Dependencies No The paper mentions various methods (e.g., Node2vec, GCN, DGI, HGNN, GRACE, Uni GCN, Hyper Conv, All Set, Tri CL, Vil Lain) but does not specify software or library names with version numbers required for replication.
Experiment Setup No The paper discusses general components like encoder, projection head, contrastive loss (including temperature parameter τ and weights we, wm), discriminator, and generator. It also includes a parameter analysis for 'm' (batch size), indicating specific values tested for 'm' ({2^3, 2^4, ..., 2^8}). However, it does not provide concrete numerical values for other key hyperparameters like learning rate, optimizer settings, number of training epochs, or specific chosen values for τ, we, wm used in the main experimental results.