Micro and Macro Level Graph Modeling for Graph Variational Auto-Encoders

Authors: Kiarash Zahirnia, Oliver Schulte, Parmis Naddaf, Ke Li

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experiments show that adding micro-macro modeling to the Graph VAE model improves graph quality scores up to 2 orders of magnitude on five benchmark datasets, while maintaining the Graph VAE generation speed advantage.
Researcher Affiliation Academia Kiarash Zahirnia, Oliver Schulte , Parmis Naddaf, Ke Li School of Computing Science, Simon Fraser University, Canada
Pseudocode No The paper includes diagrams of the model architecture (Figure 2) but does not contain structured pseudocode or algorithm blocks.
Open Source Code Yes The implementation and datasets are provided at https: //github.com/kiarashza/Graph VAE-MM.
Open Datasets Yes We utilize 3 synthetic, and 2 real-world datasets for the main results. The synthetic Grid and Lobster are from previous studies [47, 32], Triangle Grid is introduced in this paper. Protein and ogbg-molbbbp are real-world datasets from biology with information about proteins and molecules respectively.
Dataset Splits Yes Following previous experiments [47, 30, 11] we randomly split the graphs sets into train (70%), validation (10%) and test (20%) sets.
Hardware Specification No The paper does not provide specific hardware details in the main body. It refers to Section 8.11 (Appendix) for machine details, which is not provided in this text snippet.
Software Dependencies No The paper does not provide specific ancillary software details with version numbers in the main body. It mentions that library requirements can be found at the provided GitHub repository.
Experiment Setup No The paper does not contain specific experimental setup details (concrete hyperparameter values, training configurations, or system-level settings) in the main text. It states that the Appendix contains further details on experiment settings and hyperparameters.