Metropolis-Hastings Data Augmentation for Graph Neural Networks

Authors: Hyeonjin Park, Seunghun Lee, Sihyeon Kim, Jinyoung Park, Jisu Jeong, Kyung-Min Kim, Jung-Woo Ha, Hyunwoo J. Kim

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

Reproducibility Variable Result LLM Response
Research Type Experimental Our extensive experiments demonstrate that MH-Aug can generate a sequence of samples according to the target distribution to significantly improve the performance of GNNs.
Researcher Affiliation Collaboration Korea University1, NAVER CLOVA2, NAVER AI LAB3
Pseudocode Yes Algorithm 1 Metropolis-Hastings Data Augmentation (MH-Aug) Framework
Open Source Code No The paper does not provide any statement or link indicating that the source code for their methodology is publicly available.
Open Datasets Yes Datasets. We evaluate our method on five benchmark datasets in three categories: (1) Citation networks: CORA and CITESEER [31], (2) Amazon product networks: Computers and Photo [32], and (3) Coauthor Networks: CS [32].
Dataset Splits Yes We follow the standard data split protocol in the transductive settings for node classification, e.g., [4] for CORA and CITESEER and [32] for the rest.
Hardware Specification No The paper does not provide specific hardware details such as GPU models, CPU types, or memory specifications used for running experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., library names with their exact versions).
Experiment Setup No While hyperparameters are mentioned (e.g., 'λs are hyperparamters'), the paper does not provide concrete values for these or other training configurations (like learning rate, batch size, epochs) in the provided text, nor a dedicated 'Experimental Setup' section detailing these settings.