Structural Information Enhanced Graph Representation for Link Prediction

Authors: Lei Shi, Bin Hu, Deng Zhao, Jianshan He, Zhiqiang Zhang, Jun Zhou

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our approach achieves remarkable results on multiple popular benchmarks, including ranking first on ogbl-ppa, ogbl-citation2 and Pubmed. We conduct link prediction tasks on six popular datasets: three Open Graph Benchmark (OGB) datasets (Hu et al. 2020) including ogbl-ppa, ogbl-citation2, ogbl-vessel, and three classic attributed graph datasets (categorized as Classic in this paper) including Pubmed (Namata et al. 2012), Cora (Mccallum et al. 2000), Citeseer (Giles, Bollacker, and Lawrence 1998). Tables 3 and 4 report the test performance.
Researcher Affiliation Industry 1Machine Intelligence Department, Ant Group, 2Consumer Finance Technology Department, Ant Group
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code Yes Code1 is available. 1https://github.com/anonymous20221001/SIEG OGB
Open Datasets Yes we conduct link prediction tasks on six popular datasets: three Open Graph Benchmark (OGB) datasets (Hu et al. 2020) including ogbl-ppa, ogbl-citation2, ogbl-vessel, and three classic attributed graph datasets (categorized as Classic in this paper) including Pubmed (Namata et al. 2012), Cora (Mccallum et al. 2000), Citeseer (Giles, Bollacker, and Lawrence 1998).
Dataset Splits No The paper states: 'The training and test settings, hyperparameters about architecture and data processing, and other implementation details are presented in the Appendix.' However, the specific dataset split information (percentages or counts) is not provided in the main text.
Hardware Specification No The paper does not specify any hardware used for running its experiments.
Software Dependencies No The paper does not provide specific software names with version numbers for reproducibility.
Experiment Setup No The paper states: 'The training and test settings, hyperparameters about architecture and data processing, and other implementation details are presented in the Appendix.' While it mentions these details exist, it does not provide the concrete values or configurations in the main text.