Representation Learning on Spatial Networks

Authors: Zheng Zhang, Liang Zhao

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the strength of our proposed framework through extensive experiments on both synthetic and real-world datasets.
Researcher Affiliation Academia Zheng Zhang Department of Computer Science Emory University Atlanta, GA 30322, USA zheng.zhang@emory.edu Liang Zhao Department of Computer Science Emory University Atlanta, GA 30322, USA liang.zhao@emory.edu
Pseudocode No The paper describes operations and model blocks but does not include a clearly labeled pseudocode or algorithm block.
Open Source Code Yes The code for the proposed model is available at https://github.com/rollingstonezz/SGMP_code.
Open Datasets Yes We experiment on 5 chemical molecule benchmark datasets from [94], including both classification (BACE, BBBP) and regression (ESOL, LIPO, QM9). Particularly, QM9 is a multi-task regression benchmark with 12 quantum mechanics properties. The data is obtained from the pytorch-geometric library [35]. and We also conducted an experiment using the structural connectivity (SC) of the brain network to predict the age of the subjects, which is a significant task in understanding the aging process of the human brain [50]. In specific, SC is processed from the Magnetic Resonance Imaging (MRI) data obtained from the human connectome project (HCP) [82].
Dataset Splits No The paper states that training details including data splits were specified (in the checklist), but the main text does not explicitly provide specific percentages or sample counts for train/validation/test splits for the datasets used.
Hardware Specification Yes All experiments are conducted on a 64-bit machine with an NVIDIA GPU (GTX 1080 Ti, 11016 MHz, 11 GB GDDR5).
Software Dependencies No The proposed method is implemented with Pytorch deep learning framework [70].
Experiment Setup Yes The experimental settings are introduced first, then the performance of the proposed method is presented through a set of comprehensive experiments. All experiments are conducted on a 64-bit machine with an NVIDIA GPU (GTX 1080 Ti, 11016 MHz, 11 GB GDDR5). The proposed method is implemented with Pytorch deep learning framework [70]. and We vary the size and other parameters (according to Appendix. C for details) of spatial networks to collect 3, 200 samples in our synthetic dataset.