Hierarchical Topology Isomorphism Expertise Embedded Graph Contrastive Learning
Authors: Jiangmeng Li, Yifan Jin, Hang Gao, Wenwen Qiang, Changwen Zheng, Fuchun Sun
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We conduct extensive experiments on real-world benchmarks to exhibit the performance superiority of our method over candidate GCL methods |
| Researcher Affiliation | Academia | 1Science & Technology on Integrated Information System Laboratory, Institute of Software Chinese Academy of Sciences 2State Key Laboratory of Intelligent Game 3University of Chinese Academy of Sciences 4Tsinghua University |
| Pseudocode | Yes | Algorithm 1: HTML |
| Open Source Code | Yes | Our code is available at https://github.com/jyf123/HTML. |
| Open Datasets | Yes | For the unsupervised learning task, we conduct experiments on eight regular-sized graphs, I.e., TUDataset (Morris et al. 2020), and one large-scale graph, i.e., Amazon Photo dataset, which has 7,650 nodes and 119,081 edges. For the transfer learning task, we pre-train our model on ZINC-2M (Sterling and Irwin 2015) and finetune it on eight biochemical molecule datasets. |
| Dataset Splits | No | The paper lists datasets used but does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) needed to reproduce the data partitioning for its experiments. |
| Hardware Specification | No | The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details, such as library or solver names with version numbers, needed to replicate the experiment. |
| Experiment Setup | Yes | Hyper-parameter experiment. Appropriate adjustment of the introduction of hierarchical topology isomorphism expertise can better improve the model performance, so we conduct hyperparametric experiments on the PROTEINS dataset for α and β, and the results are shown in Figure 6. α and β are range from {1, 1 × 10−1, 1 × 10−2, 1 × 10−3, 1 × 10−4}. |