Heterogeneous Graph Transformer with Poly-Tokenization
Authors: Zhiyuan Lu, Yuan Fang, Cheng Yang, Chuan Shi
IJCAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | 5 Experiments In this section, we first evaluate PHGT on the node classification task, and then perform ablation studies, parameter analysis, and efficiency studies. |
| Researcher Affiliation | Academia | 1Beijing University of Posts and Telecommunications, Beijing, China 2Singapore Management University, Singapore {luzy, yangcheng, shichuan}@bupt.edu.cn, yfang@smu.edu.sg |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide concrete access to source code (specific repository link, explicit code release statement, or code in supplementary materials) for the methodology described in this paper. |
| Open Datasets | Yes | Our experiments encompass four widely-used heterogeneous graph datasets: DBLP and ACM (academic networks), IMDB (movie network), and Freebase (knowledge graph). Among these datasets, DBLP, ACM, and IMDB are sourced from a standardized benchmark called the Heterogeneous Graph Benchmark (HGB) [Lv et al., 2021]. |
| Dataset Splits | Yes | For DBLP, IMDB and ACM, we used the same data split from HGB for training, validation, and testing on each dataset. For the Freebase dataset, we use the split provided by HINormer [Mao et al., 2023]. |
| 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. It lacks any mention of the hardware used for training or inference. |
| Software Dependencies | No | The paper mentions the use of 'Open HGNN library' but does not provide specific version numbers for this or any other software components necessary to replicate the experiments. |
| Experiment Setup | No | For detailed hyperparameter settings, please refer to Appendix C.3. |