GraphMoRE: Mitigating Topological Heterogeneity via Mixture of Riemannian Experts
Authors: Zihao Guo, Qingyun Sun, Haonan Yuan, Xingcheng Fu, Min Zhou, Yisen Gao, Jianxin Li
AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | To evaluate Graph Mo RE1 proposed in this paper, we conduct comprehensive experiments and further analyze the effectiveness of Graph Mo RE on topological heterogeneity. [...] 5.2 Performance Evaluation Performance on Real-world Graphs. We evaluate our method on link prediction and node classification tasks, and the results are summarized in Table 2 and Table 3. [...] Ablation Study. We conduct ablation studies with four variants on two downstream tasks to verify the effectiveness of each component in Graph Mo RE. |
| Researcher Affiliation | Collaboration | 1SKLCCSE, School of Computer Science and Engineering, Beihang University, China 2Key Lab of Education Blockchain and Intelligent Technology, Ministry of Education, Guangxi Normal University, China 3Huawei Technologies Co., Ltd, China 4Institute of Artificial Intelligence, Beihang University, China |
| Pseudocode | Yes | Algorithm 1: The overall training process of Graph Mo RE |
| Open Source Code | Yes | 1https://github.com/RingBDStack/GraphMoRE. |
| Open Datasets | Yes | We conduct experiments on a variety of realworld datasets, including citation networks (Cora (Sen et al. 2008), Citeseer (Kipf and Welling 2017), Pub Med (Namata et al. 2012)), airline networks (Airport (Chami et al. 2019)) and co-purchase networks (Photo (Shchur et al. 2018)). |
| Dataset Splits | No | The paper references datasets but does not explicitly provide details about training/test/validation splits (percentages or counts) in the main text. It mentions "The statistics and topological heterogeneity analysis are detailed in the Appendix.", but not specifically for dataset splits. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory, or specific computing platforms) used for running the experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details, such as library names with version numbers, needed to replicate the experiment. |
| Experiment Setup | Yes | Settings. We set the number of layers to 2 for all methods. For other model settings, we adopt the default values in the corresponding papers. |