Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].

HGEN: Heterogeneous Graph Ensemble Networks

Authors: Jiajun Shen, Yufei Jin, Kaibu Feng, Yi He, Xingquan Zhu

IJCAI 2025 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on five heterogeneous networks validate that HGEN consistently outperforms its state-of-the-art competitors by substantial margin. Empirical study on five real-world heterogeneous networks validates the effectiveness of HGEN over the recent arts
Researcher Affiliation Academia 1Dept. of Electrical Engineering and Computer Science, Florida Atlantic University, USA 2Department of Data Science, William & Mary, USA EMAIL; EMAIL
Pseudocode Yes Main steps of HGEN are outlined in Algorithm 1 in Appendix.
Open Source Code Yes Codes are available at https://github.com/Chrisshen12/HGEN.
Open Datasets Yes Experiments on five heterogeneous networks validate that HGEN consistently outperforms its state-of-the-art competitors by substantial margin. Five heterogeneous graphs from real applications are used as benchmark datasets. Their statistics and detailed descriptions are deferred to Supplement B of Appendix due to page limits.
Dataset Splits No The paper mentions using five benchmark datasets but does not explicitly state the train/test/validation splits or their percentages, nor does it refer to standard splits with citations. It mentions training but lacks details on data partitioning for reproducibility.
Hardware Specification Yes All experiments are run on desktop workstations equipped with Nvidia Ge Force RTX 2080 Ti.
Software Dependencies No We choose Adam [Kingma and Ba, 2014] as our optimizer. - This only mentions an optimizer by name, not with a specific version number, nor does it list any other software dependencies with versions.
Experiment Setup No We perform a grid search with selected range of hyperparameters including hidden dimension, layer size, dropping rate, number of individual GNN, and control rate for regularizer. We choose Adam [Kingma and Ba, 2014] as our optimizer. We fix the learning rate, weight decay, the number of epochs and apply early stopping mechanism. - While it mentions types of hyperparameters and mechanisms like early stopping, it does not provide specific values or ranges for these hyperparameters (e.g., the actual learning rate value, batch size, number of epochs, or the specific range for hidden dimension).